Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Erosion in the middle Himalaya, Nepal with a case study of the Phewa Valley Ramsay, William James Hope 1985

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1985_A6_7 R34_7.pdf [ 11.37MB ]
JSON: 831-1.0096274.json
JSON-LD: 831-1.0096274-ld.json
RDF/XML (Pretty): 831-1.0096274-rdf.xml
RDF/JSON: 831-1.0096274-rdf.json
Turtle: 831-1.0096274-turtle.txt
N-Triples: 831-1.0096274-rdf-ntriples.txt
Original Record: 831-1.0096274-source.json
Full Text

Full Text

EROSION WITH  IN T H E M I D D L E H I M A L A Y A ,  A CASE STUDY  OF  THE  NEPAL  PHEW A V A L L E Y  by WILLIAM B.Sc.  JAMES  HOPE  (Hons.), University  D i p . Agric. Eng.,  Cranfield  A THESIS S U B M I T T E D  RAMSAY  of Sussex,  1974  Institute of Technology,  IN PARTIAL F U L F I L M E N T  THE REQUIREMENTS  FOR  MASTER  OF  THE DEGREE  OF  SCIENCE  in THE FACULTY OF  GRADUATE  STUDIES  Department of Forest Resources Management  We  accept this thesis as to the  T H E UNIVERSITY  required standard  OF  AUGUST ®  conforming  BRITISH  COLUMBIA  1985  W i l l i a m James Hope Ramsay,  1985  1976 OF  In  presenting  degree  copying  in  partial  fulfilment of  of  my or  copying of Department  publication of  this  thesis  or  by  his  this  thesis  for  scholarly  or for  her  requirements  Department of Forest Resources  Management  T H E U N I V E R S I T Y O F BRITISH 2075 Wesbrook Place Vancouver, Canada V 6 T 1W5  AUGUST  1985  COLUMBIA  for  an  advanced  I agree that the Library  further agree that permission  purposes  may  representatives.  financial  my written permission.  Date:  the  it freely available for reference and study. I  extensive  Head  thesis  at the T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A ,  shall make for  this  be  granted  by  the  It  is  understood  that  gain shall not  be  allowed without  Abstract Data on erosion processes and other aspects of environmental change Himalaya are scarce and unreliable, and consequently  i n the  policy decisions have been  taken  in a quantitative vacuum. Published estimates of denudation for large catchments Nepal vary from 0.51  to 5.14  in  m m / y r , and indicate a dynamic geomorphological  environment A review of the literature on erosion in Nepal revealed a consensus (1)  mass wasting is the dominant hillslope process;  (2)  virtually all failures occurring during the monsoon; (3) important determinants  activity is seasonal,  with  geological factors are the  of slope stability; (4) sediment delivery to channels  most  is high;  little quantitative evidence exists to link landsliding to deforestation. Although few exist, loss of forest cover does appear to be related to surface  that:  (5) data  erosion and gullying,  and a hypothesis linking the expansion of unmanaged, eroding areas to reduced nutrient subsidies from the forest is proposed. A reconnaissance 122  km  2  survey of sediment production and transfer  mechanisms  i n the  Phewa Valley in the Middle Mountains of Nepal identified a variety of mass  movement processes. The commonest  events were shallow translational failures on  of, typically, 36° to 45°, with volumes < 1  x 10  3  m \ and with recovery taking less  than ten years. Larger slides occurred on slopes oversteepened  by fluvial action. Flows  developed i n areas of weak rock and unfavourable structure, and were associated groundwater discharge. Flow velocities accelerated fractured and deeply weathered catchments",  movement  for approximately 90% of all sediment  production by mass wasting in the watershed. A first estimate of surface mass movement processes in the Phewa Valley is 2 - 3 on overgrazed pasture may be 5-6  with  during the monsoon. The highly  zones around faults were the sites of "mass  complex failures responsible  slopes  lowering by  m m / y r . Locally, surface  erosion  mm/yr. N o data were available on soil losses from  cultivated areas, and, similarly, losses due to shallow creep, gullying and solution  ii  remain unknown. The  fluvial  transport system in the valley bottom is unable to transport all the  material with which it is supplied. Sediment yield to the lake was not calculated owing to insufficient data. Discharge estimates and intensity-duration-frequency of rainfall records indicate that in Pokhara storms of 275 period of approximately 10  analysis  mm/day have a return  years.  The primary controls on mass movement processes in the M i d d l e Himalaya of Nepal are geological and climatic, and therefore man. However, surface therefore  are not amenable  to modification by  erosion is a consequence of poor land management,  can be controlled, given the right institutional environment  iii  and  Acknowledgements  I am indebted to my supervisor, Professor  D . L . Golding, for his faith in the  eventual completion of this thesis, and to the other members their tolerance of erratic progress.  of my committee  for  Brian Carson and Sandy Burton have provided  advice, friendship and hospitality both in Nepal and Vancouver. Gordon and Shirley Fisher kindly granted the leave of absence necessary for writing up.  The  field  work was carried out following a year spent with the United Nations  F o o d and Agriculture Organization, i n Pokhara, where Yadav Khatiwada and his staff at the office of the Department  of Soil Conservation and Watershed Management  were  unfailingly helpful and good-humoured.  My  most important debt is to my wife, Jane, for her assistance i n the field and  for her encouragement  at home. Despite the leeches, she  iv  understands.  Dedication  To Sophie who was never given a  v  chance  Table of Contents Abstract  ii  Acknowledgements  iv  Dedication  v  List of Tables  ix  List of Figures  x  1.  Introduction And Literature Review  1  1.1  The Himalayan Environment: Problems and Assumptions  1  1.2  Objectives  3  1.3  Erosion Processes i n Mountain Environments  4  1.4  The Nepalese  6  1.5  Literature Review: Erosion i n Nepal  8  1.5.1 Denudation  8  Himalaya  1.5.2 Mass wasting  ,  12 Previous work Anthropic  influences:  12 deforestation,  terracing,  construction  16  1.5.3 Surface erosion and gullying Measured  20  data  20 Estimates Surface 2.  and  23 erosion, gullying  and forest clearance  24  The Study Area: The Phewa Valley  28  2.1  Location  28  2.2  The Geological Context  28  2.2.1 The Pokhara Basin  28  2.2.2 2.3 Quaternary deposits  31  The Phewa Valley  32  Topography  33 vi  3.  2.4  Climate  -.  35  2.5  Weathering and Soils  44  2.6  Vegetation  48  2.7  Land Use  52  2.8  The Agroecosystem  53  Mass Movement In The Phewa Valley  59  3.1  Introduction  59  3.2  Methods  60  3.2.1 Aerial photography  60  3.2.2 Field data collection  62  Results and Discussion  65  3.3.1 Description of slope movement types and processes  65  3.3 Landslide Hillslope  classification  65  processes  67  3.3.2 Mass wasting rates i n the Phewa Valley Landslide  density Landslide  growth and recovery Volume  77  and area affected  77 79  and area Failure frequency  81 and surface lowering  by landsliding  3.3.3 Morphometry analysis Volume/area  82 84  relationships  84 Shape indices  86  3.3.4 Material properties and slope stability  93  3.3.5 Landslide inception  98 Geological:  undercutting, faulting, Anthropic:  effects of forest clearance Extrinsic:  pore water pressure vii  oversteepening  98 99 100 Extrinsic: 4.  seismicity  103  Fluvial Processes in the Phewa Valley  105  4.1 Introduction  105  4.2 Methods  107  4.2.1 Fluvial morphometry  107  4.2.2 Sediment transport  107  4.2.3 Sediment yield  110  4.3 Results and Discussion  5.  -112  4.3.1 Channel geometry  112  4.3.2 Sediment transport  119  4.3.3 Sediment yield  122  4.3.4 Sediment systems  126  Summary  and Conclusions  128  '.  5.1 The Problem  128  5.2 Erosion in the Phewa Valley  128  5.3 The Erosion System in the Middle Himalaya  131  5.4 Conclusions  133  and Recommendations  Bibliography Appendix 1.  138 Landslide Data from Prasad (1975)  157  Appendix 2. Erosion Plot Data, 1979, Phewa Valley  158  Appendix 3. Sediment Load: Dominant  159  Particle Size  Appendix 4. Particle Dimensions  160  viii  List of Tables  Table  Page  1.  Selected denudation rates for the Himalayan region  9  2.  Surface erosion rates reported from runoff plot studies in Nepal  22  3.  Slope categories of the Phewa Tal catchment  38  4.  Mean monthly rainfall for 7 stations  42  5.  Land use categories in the Phewa Tal catchment  53  6.  Pokhara area:  64  7.  Phewa Valley: age and angle of failure surface for 11 shallow translational  8.  Average values for the D / L ratio or process index for planar slides and flows  9.  Andheri K h o l a and Harpon K h o l a : hydraulic geometry, particle size class  i n the Pokhara area  landslide morphometry  ix  failures 66  discharge, and dominant  89 109  List of Figures  Figure cross-section  Page  1.  Schematic  of the Nepal Himalaya showing physiographic regions  2.  Location of the study area  3.  Phewa T a l : view from the west towards Pokhara  29  4.  The study area:  30  5.  The Phewa Valley: sketch  6.  Phewa Valley: Harpon K h o l a near Pame  36  7.  Schematic  36  8.  Stereogram  9.  Temperature  10.  Regression of mean annual precipitation on elevation for seven stations near the Pokhara Valley.  11.  Preliminary intensity-duration-frequency  12.  Land systems of the Phewa Valley  46  13.  Forest cover and land use in the Phewa Valley  50  14.  Simplified schematic  55  15.  Hillside near Tamagi, Phewa Valley  16.  Phewa Valley: air-photo coverage in 1972  17.  Phewa Valley: mass movement activity  63  18.  Varnes abbreviated  68  19.  Rockslide at Phedi, north of Naudanda, near Pokhara  68  20.  Debris slide, Phewa Valley.  71  21.  Slow debris flow at Pamdur, Phewa Valley  71  22.  Retrogressive  .29  sketch map of the local environment  cross-section  7  map showing places mentioned in the text  34  of Phewa Valley near Pame  o f central part of the Phewa Valley and rainfall regimes  37  at Pokhara Airport and Lumle  39 i n and  curves for Pokhara A i r p o r t  model of a Nepalese  hill farming system  and 1978.  slumping supplying mass movement  74  i  24.  Schematic diagram of part of the movement ;  61  catchment, Bijaipur, near Pokhara.  i Mass movement  43  57  classification of slope movements  23.  40  catchment at Pame, Phewa Valley Kaski ridge, showing four types of mass  x  76 78  25.  Plot of area of failure scar on age for 12 debris slides i n the Pokhara area.  26.  Regression of estimated volume on area for shallow slope failures  85  27.  Terms i n landslide morphology  87  28.  Regression of process index on age catchments  29.  Regression of tenuity index on process  30.  Diagram to illustrate the infinite slope approach to shallow slide stability analysis.  for debris slides and mass movement index for all slope failures  :  ...81  90 92  95  31.  Phewa Valley: location of stations  for channel variables survey  32.  Growth of delta in Phewa T a l , 1958  33.  Phewa Valley: long profiles of the Andheri Khola, Sidane ephemeral stream near Pame  34.  Station 5, Andheri Khola, looking downstream  114  35.  View upstream towards Station 9  114  36.  G r a p h of channel shape against distance.  116  37.  Regression of Mannings n on distance.  118  38.  G r a p h of dominant particle size class against distance  120  39.  Growth of delta i n Phewa T a l , enlarged  123  to 1978  xi  108 HI  K h o l a and an  113  Chapter 1  INTRODUCTION AND  LITERATURE REVIEW  "The concept of man as a modifier of his environment is an extremely ancient one which can be traced back to Sumerian times" (Glacken 1967).  1.1 T H E H I M A L A Y A N  ENVIRONMENT:  PROBLEMS  A N D ASSUMPTIONS  In Nepal 53% of the population (Goldstein et al. 1983) relief-energy environment of the hills and mountains. These largely dependent on subsistence  live in the high  eight million people are  agriculture for their livelihood. However, gross domestic  product per capita has been declining over the last several decades (see Bank 1979;  Blaikie et al.  1980), and intense pressure on the resource base has given  rise to fears of impending eco-catastrophe.  The first warnings of this were given by  expatriate foresters in the 1950's and 60's (e.g. Robbe 1954, not until the mid-1970's  e.g. World  Willan 1967), but it was  that the problems entered world consciousness, largely through  the writings of Eckholm (1975a, 1975b, 1976). The position has been described succinctly by Ives and Messerli: "The Himalaya-Ganges-Brahmaputra system can be categorized as one of the world's largest highland-lowland interactive systems. ;A number of assumptions have been reiterated so often that they are now largely accepted as fact and, i n turn, have had major impacts on the decision-making process and development-project design. Enumerated briefly these are: (1) population growth i n the Himalaya generates deforestation as demands for fuelwood and croplands increase; (2) deforestation leads to soil erosion and landsliding, and disrupts the normal hydrological cycle; (3) this leads to more disastrous floods and massive siltation in the wet season and lower water levels in the dry season; (4) the increased 'sediment load of the Ganges and Brahmaputra is causing an island to form in the Bay of Bengal. "This so-called vicious circle is an intellectually satisfying concept; as a working hypothesis it seems so reasonable that it is not surprising that it is accepted as fact and many consequences for development flow from that acceptance. One interesting component of this "fact"! is that it is the growing demand for fuelwood that is the cause of deforestation and, as the labour of walking greater distances to collect increasingly scarce fuel crosses a critical threshold, animal dung is increasingly used for fuel. This sets up another vicious circle: terrace soils, deprived of natural fertilizer, produce poorer crop yields, and weakened soil structure augments landslide 1  2  incidence so that more trees are cut to make room for terraces, and agriculture spreads to increasingly steeper slopes and thus to more marginal land." (Ives and Messerli 1984, p. 67). The development of these assumptions has been largely due to the  uncertainties  surrounding interdependent cause and effect relationships i n a heterogeneous  and  extremely dynamic environment The issue of uncertainty and its impact on both science and policy development in the region has been investigated by Thompson and Warburton (1985), in a provocative paper which should be required reading for all observers of the Himalayan scene. Their central thesis is that the traditional scientific approach of analysis in terms of deduced physical facts -  cis-science  -  is  inappropriate i n a region where the quantitative data are so variable that any particular policy can be justified. Instead, investigators entering the area should be aware that they will have to operate  in an environment where uncertainty is so high  that, owing to the generation of expectations processes,  of what the answers should be by social  institutional forces, rather than objectivity, can confer and withdraw credibility.  Consequently the methods of trans-science "Trans-science  (Weinberg 1972)  should be invoked.  is the science of messes" (Thompson and Warburton 1985, p. 116),  and  involves, inter alia, the study of a particular perceived physical problem as a point of entry to the complex physical, social and cultural system responsible for generating  the  uncertainty. Thompson and Warburton support their argument with the example of per capita fuelwood consumption, a variable which is intrinsically measurable.  Nevertheless,  excluding unreliable extremes, published data on per capita consumption in the Himalaya varies by a factor of 26 (Thompson and Warburton 1985, p. 117, citing Donovan 1981) . 1  Fuelwood consumption is a field which has been relatively well studied i n the region. Other aspects of resource  use and degradation have received even less attention,  A recent comprehensive review and case study of wood utilization and biomass production i n Nepal is available i n Wiart (1983). 1  3  as Ives and Messerli lament: "... there is almost no data available on rates of soil loss under different cover types and land-use practices; the sources of origin (and amounts) of the flood waters and silt, whether the Siwaliks, Mahabarat Lek, M i d d l e Mountains, Greater Himalaya, or Trans-Himalaya, are virtually unknown;"...."The catalogue could be extended almost indefinitely." (Ives and Messerli 1984, pp. 67-68). What data there  are should be treated with caution, and original sources  referred  to i n order to clarify the context and methodology of the original study. Only i n this way can the hazard o f generalizing locally-derived data over wider physical and social environments be avoided.  1.2  OBJECTIVES The objective  erosion processes  of this thesis is to reduce the overall uncertainty concerning  and rates of sediment production in the M i d d l e Mountains of Nepal.  T o this end: (i)  the literature on current rates of surface erosion and mass wasting in Nepal is reviewed, and a comprehensive bibliography prepared;  (ii) the results o f a reconnaissance survey of geomorphological processes in a small watershed i n the Middle Mountains near Pokhara are presented; (iii) the results of the survey are summarized, and the implications for land management in Nepal are discussed. The objectives (i)  of the case study near Pokhara were:  to make a preliminary description of mass movement processes the study area;  (ii) to make a preliminary description of fluvial aspects of the sediment transport system.  in  4  1.3  EROSION PROCESSES IN M O U N T A I N  ENVIRONMENTS  Mountains are essentially geological ephemera.  In comparison with geologic  time  there is no such thing as an " o l d " mountain range. This is due to the tendency erosion to increase relief (see  in effectiveness with both altitude (Price  e.g. Schumm 1963;  Ruxton and M c D o u g a l l 1967;  was summed up as long ago as 1876  1981, p. 166),  of  and local  Young 1969). The position  by the geologist J.W. Powell working in the  western United States: "We may now conclude that the higher the mountain, the more rapid its degradation; that high mountains cannot live much longer than low mountains, and that mountains cannot remain long as mountains: they are ephemeral topographic forms. Geologically all existing mountains are recent; the ancients are gone." (Powell 1876, p. 193). Geomorphological processes active in mountain environments include weathering, frost action, glaciation, nivation (essentially  a transitional process  periglacial systems (Price 1981, p. 209)), mass movement, surface  between  erosion, creep, solution  processes, fluvial action, and the effect of wind. These phenomena in a number of texts (e.g. Fairbridge 1968; McPherson 1972;  Ives and Barry 1974;  In the Himalayan context  glacial and  are well  Carson and K i r k b y 1972;  described  Slaymaker and  Price 1981), and are not reviewed again here.  it is as well to emphasize the distinctions  between  surface  erosion, fluvial action, and mass movement, since they are frequently  Surface  erosion includes rainsplash, sheetwash, rilling, and the effect of wind, and is  primarily a product of sparse or reduced reduced vegetation cover. Fluvial  confused . 2  action  involves erosion, sediment transport, and deposition by flowing water confined to channels, or during floods. Mass movement (synonymous with mass wasting) includes all gravity-induced movements  except those in which material is carried by a transporting  medium such as water, air or ice. In practice, processes sometimes  merge into each  The Nepali term for landslides, pahiro, includes both mass movement and the high-angle fluvial gullies with which landslides are often associated. It is a logical descriptive term, but is less useful for analysis of process. 2  5  other and distinctions between them become arbitrary (Fairbridge 1968). Nevertheless, it is important that policy-makers appreciate  the differences  between these three general  categories of erosion. They have fundamentally different causes and consequences, and whereas surface  erosion is controllable, mass wasting often is not (at an  price). Gullying can sometimes  be checked  by structural measures,  economic  but is best prevented  before initiation by minimising uncontrolled runoff. In the discussion which follows it is also useful to bear i n mind a number geomorphological concepts which can place the processes described i n perspective. simplest of these is the difference  between  no different at the fundamental process erosion has been man, whereas (U.N.F.A.O.  defined as the  "accelerated"  geological and accelerated  of The  erosion. Although  level (Novak and van Vliet 1983), geological  "normal" process  operating without the influence  of  erosion results from man's activities on the land surface  1965). Another relevant concept  is that of process domains or sets of  environmental conditions which, through the operation of a constant produce over time a set of characteristic  set of processes,  landforms, such as convex or  hillslopes, debris fans, or river terraces (see  e.g. Skempton 1953;  to the concept of process domains is that of relaxation  concave  K i r k b y 1978). Central  time, the period following an  event or impulse that carries out geomorphic work, and during which the responses the initial impulse eventually produce a characteristic  landform. F o r example,  recovery of a landslide scar after failure, becoming a stable, vegetated and Thornes have reviewed this and related concepts sensitivity and change  (Brunsden and Thornes  decay relaxation paths toward a characteristic  the  hollow. Brunsden  in a seminal paper on  landscape  1979), and suggest that, inter alia,  slope failure sites and the impact of deforestation  both  may follow first-order exponential  form, i.e., the responses  to these events  slow down as time increases after the initial incident Alternative system responses impulses include either change  to  to a new level of geomorphological activity (having  to  6 crossed some critical stability threshold), or entrance to an area of reinforcement through positive feedback.  A n example  of the former, given by Brunsden and Thornes,  is soil compaction leading to reduced infiltration capacity, increased  relative runoff,  exceedence o f critical thresholds, and gully initiation (Brunsden and Thornes 475)\  Feedback  1979, p.  may occur where, for example, slope failure exposes bare ground  causing increased runoff, which i n turn causes the area of the failure scar to expand. A  fourth useful concept is that of the spatial propagation of an impulse through  the landscape. Changes i n some  variables such as climate are ubiquitous,  occurring  effectively simultaneously over the whole landscape. Others, such as changes i n base level , are transmitted linearly along erosional axes, most importantly river channels, and 4  diffuse out from these axes to slopes. A s in other areas of high relative relief, this theme of slope-channel  coupling (Brunsden and Thornes  1979) is central to an  understanding o f the efficiency o f erosional processes i n the Himalaya.  1.4 T H F , N E P A L E S F  HIMALAYA  Regional physiography has been succinctly described by Brunsden et al. (1981) and  is illustrated in Figure 1: "The Nepal Himalaya, an 800 k m long unit o f the 2400 k m long Himalayan mountain system, may be divided into five strike-oriented, longitudinal relief units. T h e northern three units form an intimate association consisting of two mountain ranges, the Tibetan Marginal Range, which forms a 6000 to 7000 m high southern rampart to the Tibetan Plateau, and the Great or H i g h Himalaya, a line of mountain masses with an average elevation of 6000 m and containing numerous peaks rising to above 8000 m . Between these occurs the Inner Himalaya, a number of "natural compartments" (Hagen 1965) standing at 2000 to 6000 m and surrounded by high peaks. South of the High Himalaya lies the broad belt of the Middle Himalaya (generally 3700- 4500 m) and the Outer, Fore or 5  6  F o r a discussion of the concept o f thresholds i n geomorphology see Schumm (1979). In this context, base level means the local level to which, i n the absence of uplift, the land surface will eventually be reduced by erosion, i.e. the Terai. Sometimes termed the "Midlands". Probably a misprint for 700-2500 m .  3 4  5 6  Physiographic Regions: Nepal Himalaya  TERAI  SIWAL1KS  Quaternary  Tertiary  MIDDLE HIMALAYA (Middle Mountains)  TRANSITION HIGH HIMALAYA (High Mountains) Precambrian  to  Eocene  INNER HIMALAYA / TIBETAN MARGINAL RANGE / TIBETAN PLATEAU Cambrian to Cretaceous  Figure 1. Schematic cross-section of the Nepal Himalaya showing physiographic regions. Modified from: Nelson et al. (1980), and L . R . M . P . (1983). Geological symbols after Gardiner and Dackombe (1983).  8  Low Himalaya (including the Siwalik Hills), the latter forming a foothill zone of variable width with ridges that usually rise to 900-1200 m and which overlook the low lying (<300 m) Ganges Plain, known locally as the Terai." (Brunsden et al. 1981). 7  The geology of the region has been comprehensively reviewed by Stocklin (1980) , and the Himalaya are acknowledged to be the world's youngest  major  8  mountain chain. Estimates of current uplift include circa 1 m m / y r for the Himalaya Range (Zeitler et al. Pleistocene  (Brunsden et al.  1982), about  1981,  the Nepalese Himalaya (Iwata et  1-4  Great  m m / y r since the end of the Lower  based on L o w 1968), of the order of 1 m m / y r for al.  1984), and circa 9 m m / y r for the last  h a l f - m i l l i o n years i n the Nanga Parbat area of Pakistan (Zeitler et  al.  1982).  The climate is monsoonal, with precipitation decreasing from east to west along the range. Climax vegetation is forest up to the tree line, but over the centuries  the  vegetation cover has been extensively modified by man. For an introduction to the history of deforestation i n northern India and the Himalaya readers should consult Tucker (1982, 1983). A useful introduction to both physical and cultural aspects of the current changes  1.5  taking place in the Himalaya is Lall and Moddie (1981).  LITERATURE REVIEW:  1.5.1 An  EROSION IN  NEPAL  DENUDATION impression of the intensity of geomorphological processes  operating i n the  Himalaya can be gained from denudation rates (vertical lowering of the land surface) estimated for the region, and for catchments  of some of the main rivers (Table  The rates reported are high, up to five times or more higher than general  1).  estimates  of erosion even i n areas of steep relief (0.1-1.0 m m / y r : Saunders and Young 1983). The M i d d l e and Outer Himalaya together Himalaya". See also L e Fort (1975). 7  8  are sometimes termed the  "Lesser  9  Table 1. Selected  denudation  Location  rates for the Himalayan  Denudation Rate (mm/yr)  Himalaya Ganges/Brahmaputra catchment  1.0 0.7  R. Hunza catchment R . Tamur catchment R . Tamur catchment  1.8 5.14 4.7  R . Tamur catchment R . A r u n catchment R. A r u n catchment R . Sun Kosi catchment R. Sun {sic) catchment R. Kosi catchment  2.56 1.9 0.51  region f  Comments  Regional From present rate o f influx to Bay of Bengal F a n From sediment yield From sediment yield From sediment yield 1948-1950 1947-1960  2.5 1.43  2  3  R . Sapta Kosi catchment R. Karnali catchment Darjeeling area Darjeeling area  0.98 1.00 1.5 0.5-5.0 10.0-20.0  From suspended sediment  Forested/deforested In catastrophic storms  Author  Menard 1961 Curray amd Moore 1971 Ferguson 1984 Seshadri I960 Ahuja and R a o 1958 Williams 1977 Pal and Bagchi 1974 Williams 1977, after Das 1968 Pal and Bagchi 1974 1  1  1  1  Williams 1977, after Das 1968 Schumm 1963, based on Khosla 1953 Williams 1977, after Das 1968 U . N . D . P . 1966 Starkel 1972a Starkel 1972a  In: Brunsden et al. (1981) Sun Kosi Sapta Kosi f It should be noted that there is considerable uncertainty i n these figures since: (1) measurements o f suspended sediment are subject to major inaccuracies without rigorous sampling techniques and accurate data on discharge (see e.g. Walling 1977; Walling and Webb 1981; Dickinson 1981; Ward 1984); the characteristics o f the rivers listed are such that measurements o f both sediment transport and discharge are difficult, and therefore this constitutes a source o f gross error; (2) the figures are generally based o n measurements o f suspended sediment alone; the contributions o f the solution and bed loads are not included (see e.g. Branski 1981); the estimates are, therefore, subject to systematic error; (3) no allowance is made for catastrophic events outside the period o f measurement, yet these may well be responsible for a high proportion o f total sediment movement; (4) different authors have used different conversion factors and sediment delivery ratios to calculate denudation from sediment load. In Nepal, few authors except Zollinger (1979a, p. 28), have drawn any attention to these problems. See also Roehl (1962); Meade (1969); Warhaftig (1970); Meybeck (1976); and Trimble (1977). 1  2 3  10  Despite their inherent errors the environment  Sharma  Hooker 1854;  1973;  Krishnan 1960;  Remy 1975)  Hormann 1974;  Fort and Freytet  1979,  Heuberger  et  1982;  al.  effort has concentrated  Bordet 1961;  Gansser  1964;  on the  Hagen  1965,  and late Quaternary history of the Himalaya (e.g.  D e Terra and Patterson  Usselmann 1971;  1982;  suggest a very dynamic geomorphological  However, until recent years research  geology (e.g. Wadia 1957; 1969;  figures  1939;  Bordet et al.  M u k e r j i 1975;  1971;  Thouret 1975;  Freytet and Fort 1980;  Dollfus and  Fort in press,  Yamanaka 1982;  1979 ; 9  Yamanaka et  al.  1984), with a view to determining the geological history of the  area and to establishing a Pleistocene  chronology for the existing depositional and  erosional landforms. W i t h the increase  in concern over the region, more attention has been paid to  describing and understanding current processes of landform evolution. The contribution to the geomorphological debate is a paper by Brunsden et  key  al.  (1981) who  collated much of the data available to date on the Low Himalaya of eastern added to it their own geomorphological observations  Nepal,  from two field surveys for a road  alignment (Dharan to Dhankhuta), and produced the first comprehensive  description of  current hillslope and fluvial processes and sediment systems. They found that  the  relative relief of their study area had been increasing throughout the Pleistocene owing to stream  incision being greater than ridge crest lowering. The rapidly incising drainage  network was transmitting the effects of tectonic this largely accounted  and isostatic uplift to hillslopes, and  for the lengthening and parallel retreat of basal slopes by  undercutting and landsliding. Debris mobilized on the slopes  flowed  straight into river  channels (giving a very high sediment delivery ratio ) for subsequent 10  fluvial action. This form of process integration has been termed  removal by  synchronised  Contains excellent bibliography for introduction to the French literature. Ratio o f sediment yield in a river to gross sediment production i n the upstream. 9  10  catchment  11  degradation  by Starkel (1972b).  One of the consequences of this efficient slope-channel  sediment system is that  debris delivery to the valley bottom is not continuous. Material tends to arrive i n waves during storm events, causing pulses of heavily-silted water to move downstream. Carson (in press) reports rapid fluctuations in the sediment load of the Narayani  river,  with values of up to 25,000 ppm being recorded regularly during the monsoon. H e attributes these very high sediment loads to point sources, i.e. slope failures, and emphasizes  the implications for the design of engineering structures  intakes. Brunsden et Nepal over the 1974  al. (1981) observed the Leoti K h o l a (khola monsoon, and found that the sediments  such as irrigation  N e p . river) in  delivered to the  tended to be deposited quickly as flood stages receded. The transfer downstream is probably largely dependent on catastrophic  eastern channel  of this material  events, which accomplish a  very large amount of geomorphological work i n a short time. For example, i n August 1968  a landslide-dam at Labubensi on the Buri Gandaki river broke and caused  disastrous flooding downstream (Sharma  1974), and more recently a dam-burst  resulted  i n the River Tamur scouring its gorge in the Middle Mountains to a height of 20 above the bed (Carson, in press). The  frequency of such events i n the Himalaya  remains uncertain, but Starkel (1972b) has proposed a return period of 20-25 years catastrophic  m  rainfall i n the Darjeeling Hills, and Brunsden et al.  formative events for slope and valley landforms i n eastern  (1981) suggested  for  that  Nepal may occur on the  order of once every ten years. Both authors emphasized the importance of heavy precipitation i n achieving pronounced erosion, but did not omit seismicity as a possible trigger for major landslides . 11  Despite their finding that the L o w Himalaya of eastern most rapidly denuding- areas of the world (Table  Nepal is one of  1, R. Tamur catchment),  F o r a further discussion of the role of extreme meteorological evolution see Starkel (1976). 11  the  Brunsden  events i n slope  et  12 al.  concluded that present landforms may be regarded as characteristic  forms in  equilibrium with current processes and the controlling tectonic, climatic and base level conditions (Brunsden et al.  1981,  environment, Brunsden et al.'s  pp. 66 and 69).  In this dynamic geomorphological  paper successfully identified both transport and storage  processes, and recognized the linkages between them. However, quantification of these elements  and the development of even a preliminary sediment budget for a  watershed  is at a very early stage in Nepal. Original studies are extremely scarce. Papers i n the literature supplying original data on mass wasting, surface Nepal are reviewed below.  1.5.2  MASS  erosion, and gullying in  12  WASTING Previous work Even casual visitors to the Siwaliks and Middle Mountains of Nepal are by the prevalence  of landslides. Recent research  has  struck  "concluded that mass wasting is  the dominant process in the evolution of natural slopes throughout much of  the  Himalaya" (Carson, in press), although this opinion is due more to their spatial frequency rather than to any knowledge of quantities of material moved or data on failure frequency  and slope retreat The high incidence of landsliding is well illustrated  by geomorphological maps such as those of the L o w Himalaya of eastern  Nepal in  Brunsden et al. (1981), the valley of the Ankhu K h o l a in central Nepal (Thouret 1981a, 1981b), the 1983,  K a t h m a n d u - K a k a n i area north-west of Kathmandu (Kienholz et  al.  1984), and Goorkha, Mustang and Myagdi Districts by, respectively, White, Fort  and Shrestha  (White et al.  1983;  Fort et al.  1984). The largest documented landslide  in Nepal occurred some 30,000 years ago in the Langtang valley, and involved some According to Rieger (1978/79), an analysis of "about 500" titles pertaining to erosion, sedimentation, and environmental deterioration i n the Himalaya is available i n Rieger (1975). See also Rieger (1976). 12  13  15 k m  3  of material (Heuberger  et al.  1985).  For quantitative data on current rates of mass wasting most authors refer Starkel's work carried out in the Darjeeling Hills east of Nepal (Starkel 1972b). Starkel studied the response  of slopes to a catastrophic  700-1100 m m over 3 days in 1968,  1970,  rainfall event  and noted few signs of surface  1972a, of  erosion. However,  suffosional forms of mass movement (piping) and landslides were common. In to rainfall intensities of 40-60 m m / h r  "the considerable  to  response  seepage pressure of water led  to the formation of thousands of mudflows" and other forms of mass wasting (Starkel 1972b, p. 125). failures to be  Starkel considered the role played by vegetation i n inhibiting shallow "most important", with destruction i n the forest being "10  less than on the tea slopes"  (Starkel  on "deep landslides" (Starkel  1972b, p. 131).  estimated to be  times  However, forests had no effect  Mean degradation in the region was  5 m m / y r , including the contribution from catastrophic  return period of 20-25 years (Starkel An  1972a, p. 142).  to 20  events with a  1972a).  earlier paper concerning landslides and soil erosion in the same area is that  of Dutt (1966), who surveyed a number of watersheds  in the Kalimpong S u b - D i v i s i o n  of Darjeeling District following disastrous floods in North Bengal i n 1954.  H e observed  that the primary cause of the landslides surveyed appeared to be unfavourable  geology,  with subsidiary causes including the removal of forest, bamboo, and grass, overgrazing, concentrated  runoff from livestock trails, and cultivation on slopes greater than 20° with  no terracing. A major contributor of sediment to rivers was the collapse of terrace gravels due to undercutting. Since the publication of these papers the only quantitative material to become available on current mass wasting processes i n N e p a l has been by Bansode  13  13  of which the author is aware  and Pradhan (1975), Prasad (1975), Williams (1977), Laban  Apart from internal project  reports with limited circulation.  14  (1978/79, 1979), Wagner (1981, 1983), Brunsden et al. (1982).  14  Numerous authors (e.g. Singh et al.  1983)  (1981), and Caine and M o o l  have postulated pathways  accounting  for soil loss in the Himalaya, but without supporting numerical data. Bansode  and Pradhan (1975) carried out a reconnaissance survey of landslides  along part of the channels  of the Sun Kosi and Tamur rivers above  Their impression was of high levels of mass movement  Tribeni in  1963.  activity contributing to high  sediment loads in the rivers, " mostly due to heavy precipitation, deep weathering, steep dip-slopes of the valley walls, under-cutting of the banks due to high velocity of these rivers, unstable nature of the rocks due to their structural disposition, failure of the shear resistance of the accumulated debris, high seismicity of the area, unplanned deforestation, etc." (Bansode and Pradhan 1975, p. 253). Failure surfaces were between  30° and 70°  (n=19).  Prasad (1975) reported on 10 years' observations landslide occurrence  i n the Durbasha watershed  1), and his paper constitutes  of seismicity, rainfall and  near Chatra i n eastern Nepal (Appendix  the first published material relating these factors i n the  country. Overall, landslide incidence corresponded  with high levels of both .precipitation  and seismic activity i n July and August . However, slides also occurred i n years of 15  low earthquake  activity, and so Prasad concluded that "seismic shocks by  are not the main cause of occurrence epicentre"  (Prasad  1975,  p. 79).  of landslides in regions away from the  Hydrological conditions, i.e. high groundwater levels and  intense precipitation, were considered to be more  important.  Williams (1977), investigating the east to west shift of the ERTS MSS5  satellite  themselves  Kosi river , used 16  imagery to identify all slides larger than approximately 20 ha i n  the catchment of the Sapta  __________________  Kosi i n eastern Nepal, correlated  these with data from  \  S. Matsuura is registered to give a paper entitled "The characteristics of landslides in the M i d l a n d area of Nepal Himalayas" at the forthcoming IVth International Conference and F i e l d Workshop on Landslides, to be held i n Tokyo from 23-31 August 1985, under the auspices of the Japanese Landslide Society. The author does not comment on the apparent seasonality of earthquakes. See also G o l e et al. (1966). 14  15  16  15  1:63,360 topographic maps available for a small part of the basin, assumed a failure scar recovery period of 50 years, and used Simonett's (1967) empirical volume/area relationship derived for slides in New G u i n e a Sapta Kosi basin over 50 years of 0.91  17  x 10'  to give a "total slide volume" for the m \ He estimated that these  "large  landslides" contributed 31% of the sediment load of the Sapta K o s i , with a further 64% coming from small slides, surface  erosion and gullying. The assumptions concerning  failure age, area and volume, and basin homogeneity required for this procedure leave the accuracy of William's figures open to question. Laban (1979) carried out a reconnaissance  slide intensity survey of the whole of  Nepal, but expressed his data i n terms of number of failures per linear km when viewing from one side of a light aircraft, by ecological region . Perhaps his most 18  useful observation was the severity of the impact caused by road and trail construction, to which he attributed 5% of all slides observed. Wagner (1981, 1983) characteristics  used a statistical analysis of geological and other  of 100 landslides, mainly along roads in the Middle Mountains, to  develop a site-specific landslide hazard assessment and mapping methodology based on the use of "equatorial Schmidt projections" '. 1  H e concluded that geological factors  were  of overriding importance in determining debris and rock-slide hazard. Brunsden et al.  (1981) made the crucial observations that (in their study area in  eastern Nepal) mass movement phenomena were concentrated  in two locations: low  level undercut situations such as ravines and the outside of meander bends, and areas of structural discontinuity, suggesting an important role for "intensely shattered rock and preferred water movements". M o d a l angles for debris slides were 35°-43°, and they tentatively identified a slope angle of 30° as a lower limit for first-time shallow See See Nelson et al. (1980). A geometrical tool for studying the intersection of geological planes (see 1968). 17  18  19  e.g. John  16  debris slides. " M u d s l i d e s " movement catchments",  20  occurred on slopes of 25°-39°. They also described  "mass  steep, rapidly eroding channels with active, expanding heads  supplying material to the channels by debris slides, debris flows, rock-debris chutes, and gullies. Gullying i n areas underlain by gneiss, which weathered to depths of m , was severe, and was attributed to man-accelerated cultivation following clearance  (Brunsden et al.  20  soil erosion as a result of  1981).  Caine and M o o l (1982) investigated mass movements i n the  JCathmandu- Kakani  area as part of the United Nations University Hazard Mapping Project (see  Ives and  Messerli 1981). Their analysis of slide morphometry and slope material properties led them to emphasize the importance of material controls on the landslides i n their study area, particularly the brittle behaviour of the weathered, untransported bedrock.  The  high incidence of catastrophic landsliding was further explained by relief, seasonally high water tables, and recent  deforestation. Rainfall was thought to be of comparatively  minor importance. They gave an estimated rate o f surface lowering by landsliding of 12 m m / y r .  21  0 Anthropic  influences:  deforestation,  terracing,  and  construction  There is considerable evidence from n o n - H i m a l a y a n environments that anthropic disturbance of the land surface can increase  rates of mass wasting. Such  may be divided into two categories:  i n land use, principally the removal of  changes  disturbances  forest cover, terracing and irrigation; and construction activities, principally roads and canals. These are discussed below with respect to the Middle Hills of Nepal.  Deforestation.  Although one author has implicated reforestation in increasing slide  "Elongate or lobate masses of weathered debris which move i n well defined tracks bounded by steeply inclined lateral and basal shear surfaces." (Brunsden et al. 1981, p. 45). See also 3.3.1. Recalculation of their data gives 11.3 m m / y r (author). 20  21  17  incidence through encouraging soil formation and so reducing between-failure  periods  (Shimokawa 1984), increases i n rates of mass wasting following deforestation have  been  recorded from many parts of the world, including for example British Columbia (Schwab  1983), Oregon (Swanson and Dyrness 1975), and N e w Zealand (Trustrum et  al. 1984). The reasons for the increase root reinforcement 1979,  are generally stated as being due to decreased  of regoliths as roots die and decay (O'Loughlin 1972, W u et al.  Ziemer 1981), and/or increases i n pore water pressure  due to reduced  evapotranspiration and so higher groundwater levels. The ways i n which forest can influence slope stability and erosion have been  cover  summarized by O'Loughlin and  Ziemer (1982). Positive influences depend upon: * modification o f soil moisture distribution and soil pore water caused by forest evapotranspiration;  pressures  * accumulation o f an organic forest floor layer; * mechanical reinforcement  of the soil by tree roots.  Negative influences result from: * root wedging and windthrow; * surcharge  due to weight of the tree crop.  The accumulation of the litter layer is beneficial both i n preventing rainsplash, and  i n improving infiltration and ground-water recharge  through keeping open the  entrances to macropores i n the mineral soil. It could be argued that the resulting higher water tables will decrease slope stability, but this must be balanced against the increased moisture storage available due to lower water tables caused by forest transpiration. O n balance  the net influence of forests on slope stability is positive, with  the major factor being root reinforcement  ( O ' L o u g h l i n 1984a). There is thus a  presumption that removal of the forest cover will affect slope stability to some However, i n Nepal, observations  degree.  have not invariably supported this presumption.  18  F o r example, Upadhyay (1977) quotes Hunting Technical Services  (1975) : 22  "Areas with the most numerous and most spectacular land slides (apart from those associated with roads) were observed to be under forest cover and appeared to be more closely associated with structure and landform rather than land use" (Hunting Technical Services 1975, cited i n Upadhyay 1977, p. 31). Furthermore, although all the papers o n mass wasting i n Nepal reviewed i n the preceding section ( concur on the importance of geological factors i n precipitating landslides, none of the authors except forest clearance  Starkel found any clear relationship between  and mass movement activity.  One of the difficulties i n obtaining numerical data with which to examine this question is that i n the Middle Mountains of Nepal remaining areas of forest tend to be on slopes too steep for terracing and cultivation, which by definition have a greater  tendency towards mass wasting than areas which have been cleared. Thus, any  study of the relative frequency of landslides on forested and n o n - forested  slopes  immediately runs into the problem of locating undisturbed control sites. O n the basis of visual observations alone, it is reasonable  to suggest that geophysical considerations  such as relief, faulting, undercutting, and heavy precipitation, are likely to have more effect on slope stability than forest influences such as root reinforcement and lowered water tables. A n exception to this general rule may be the small, shallow, translational failures ubiquitous i n the M i d d l e H i l l s of Nepal. These are usually located i n m i d or upper slope positions, often recently cleared of forest, and involve material within the rooting depth of trees and shrubs. The persistence  of the public association of rapid mass failures with  i n Nepal can be attributed to four factors;  firstly, the existence  deforestation  of predisposing  evidence such as that of Wilson (1973), who found that on an international scale land use outweighed both precipitation and relief as the single most important factor  22  N o t listed i n Upadhyay's bibliography.  19 affecting sediment yield; secondly, the transposition of erosion models developed elsewhere  which demonstrate  such a link; thirdly, the undoubted effect of  deforestation  on surface erosion and gullying in the hills, as opposed to mass wasting (see below); fourthly, the credibility afforded to the link by institutions, "experts",  23  the media, and  e.g. Kollmansperger, who regards mass wasting in the Midlands,  Mahabharat and Siwaliks "as being a man-made (Kollmansperger 1978/79, p.  Terracing.  Terracing per  1.5.3  the  erosion process on an exclusive  basis"  19).  se is not a direct cause of mass wasting, and indeed in some  areas may make a positive contribution to slope stability through facilitating the rapid reclamation of failed sites which might otherwise expand (see Schnieder  Kienholz, Hafner and  1984). Although the addition of irrigation water to level terraces can cause  slope failure by increasing regolith mass and decreasing  cohesion and shear  the length and intensity of human occupancy of the mountains means  resistance,  that areas liable  to slide due to irrigation have probably already done so (B. Carson, i n press). Existing irrigated terraces (Nep. khet) are generally stable, and small slumps and collapsed terrace risers are quickly repaired . The perception of landslide hazard also 24  sometimes  leads farmers to shift to lower intensity land uses, such as from irrigated to rain-fed cropping (Johnson  Construction  et al.  activities.  1982).  The category  of construction activities includes dams, canals,  quarrying and road building. D a m building has been implicated i n mass wasting in Sri Lanka through saturation of materials at the toe of shallow angle deposits, causing slumping (Russell  1981), but i n Nepal is primarily associated  with problems along  See e.g. Asian Development Bank (1982); Brown et al. (1984). See Green (1978) and L . R . M . P . (1983) for a full discussion of terracing practices i n Nepal. 23  24  20  access roads and around borrow pits, as at Kulekhani south of Kathmandu. The interruption of fluvial sediment transport by dams and reservoirs may also cause new erosion downstream, since the rivers will have greater erosive power without their original sediment load, and will seek to regain equilibrium. Canals in the hills, both large and small, are frequently associated  with  slope  failure due to both the removal of toe support from slopes and to saturation of the regolith by seepage and overflow. The maintenance system is a major  of the irrigation water distribution  burden on rural inhabitants. Quarrying can also cause slope  by removing basal support, but these are usually small-scale A  failures  events.  much larger problem in Nepal is the impact on slopes of road construction.  Although much of the difficulty is due to poor or non-existent alignments, and poor construction techniques environment is extremely  (see  design studies, poor  e.g. Kojan 1978), the Himalayan  hostile to road building. Problems centre on road cuts, which  destabilize slopes through the removal of toe support and weight,  fills,  which overload  slopes already near critical angles, and drainage, principally the concentration  of water  which was previously dispersed or able to drain away. Even when extensive  engineering  geology investigations have been undertaken prior to construction formidable problems remain. The Dharan-Dhankhuta road (the a case i n p o i n t Road maintenance  1.5.3  subject of Brunsden et al.'s  constitutes  field  a severe drain on government  studies) is resources.  SURFACE EROSION A N D G U L L Y I N G Measured  data.  Measured rates of surface publications: Chatra Research  erosion i n Nepal are only available in four  Centre"  (1976), Laban (1978), Mulder (1978b), and Impat  A n Indian aid project associated with studies of the Sapta Kosi river. Variously referred to i n the literature as Chatra Research Centre, Chatra Forest Research Centre, Chatra Experimental Station, and Soil Conservation Research, Demonstration and 23  21  (1981). A l l four sources report on runoff plot experiments, and Laban's paper also includes a literature review and original data on erosion determined by silt accumulation behind gully check dams. Four runoff plot installations are known to have existed in N e p a l : at Chatra i n East Nepal (Chatra Research  Centre 1976); at Gagretal near Surkhet i n the West  (Laban 1978); and in the Phewa watershed near Pokhara i n Central Nepal (Mulder 1978b; Impat 1981). The fourth installation was observed on the Shivapuri watershed immediately north of Kathmandu but was in a state of disrepair (1983), and no results from it are known to have been published. Rates of surface erosion reported from the first three installations are shown i n Table 2. It should be noted that all the data are short-term, site-specific, and under no circumstances should be generalized over large areas. F o r example, Mulder's results were based on only 4 individual measurements made following rain events between June and 5 July 1978  (Mulder 1978b), and the small plot size (10  29  m ) is hardly 2  representative of a longer slope. However, the figures do serve to illustrate both the high absolute rates of surface erosion under some circumstances (Gagretal), and the relative differences between different land use types, e.g. overgrazed pasture (9.85 from 11 June to 15 O c t forest (0.43  1979), protected pasture (1.01  t/ha between 01 July and 07 O c t  1979)  t/ha  t/ha for the same period), and  at Banpale and Tamagi i n the  Phewa watershed near Pokhara (Impat 1981). O f particular interest is Impat's (1981) finding that, although precipitation peaked i n August, soil loss was greatest at the beginning of the measurement period, i n June (Appendix 2). This is most likely to be related to an increase i n the vigour of the grass cover during the monsoon, and to date is the only quantitative evidence of the importance of vegetation i n reducing surface erosion i n the M i d d l e Mountains of Nepal. "(cont'd) Training Centre, Chatra. F o r further information contact: Central Soil and Water Conservation Research and Training Institute, Dehra D u n 248 195, India.  Table 2.  Surface erosion rales reported from runoff plot studies In  Land  Location  and plot data  Siwaliks:  Chatra, east Nepal;  of measurement Siwaliks: number  average  Mountains:  central  Nepal;  clay loam,  each to  land  5  10  moderately  Middle north schist;  0.35;  1  Values Soil  '  Same  Nepal;  south  Severely  aspect,  grazed  grey well  phyllitic  drained;  to  grazing.  7.8-36.8  t/ha/yr  Chatra  degraded heavily  schist;  one  soils 40-70  10  200  t/ha/yr  forest on  intensively  use  type;  daily measurements Tamagi,  Research Centre,  Laban  29  from  composite measurements  given  erodibility  as  tons(j/c)/ha/yr  (VVischmeier  fenced pasture  as  Mulder  K-value 11  !  pers.  Laban  1978  1978;  two  Fenced/un fenced grazing  9.4/34.7  t/ha/yr  1  Mulder  1978b  land.  Protected  June -  15  Oct.  1979.  m; grey  one  10  m  !  with  pasture mixed  forest/overgrazed  0.43  forest.  plot; well  July to 07  Oct.  1979.  in original publications.  et al. 1971).  in Mulder's  study; trees were  t/ha/period of  seedlings at  t/ha/period of  measurement  schist and quartzite  01  1.01/9.85  Impat  1981  Impat  1981  measurement  land.' Dense  comm., in  gullied badlands.  June  of surface soil  1978  Sakya,  m ' plot on  measurements  Phewa watershed, near Pokhara;  elevation 1800  clay loam derived  1  forest  Various,  Source  rate  in  of measurement and  identical location to land  Mountains:  11  Erosion  1978.  east aspect,  drained  use  given.  given.  south aspect,  Mountains: as  sandstone; period  Banpale, Phewa watershed, near Pokhara,  m* plots each  given  period  use type; four individual  July  Middle  Surkhet, west  slope 60%;  of plots not  Middle cm  south aspect,  and number of plots not  Gagretal, near  sandstone,  Nepal  the  time  (author).  23  Laban's method for determining soil loss from silt accumulated behind check dams included the use of sediment delivery ratio and trap efficiency curves developed for use i n other environments such as the United States (probably based on the work of Brune (1948, 1953)), as well as assumptions on gully shape and sediment contribution from g u l l y - w a l l slumping. Consequently, his figures should be treated  with  caution. Other soil loss figures reported by Laban are based on a preliminary exercise in stream sediment sampling in the Kathmandu Valley undertaken by Kandel (1978). Sampling was carried out once or twice weekly, and the results extrapolated to give annual soil loss  figures.  Therefore  the same caution again applies.  Data on soil loss from cultivated land in Nepal are not available. Sastry and Narayana (1984) reported markedly reduced runoff and erosion from bunded areas compared to rates under forest i n the D o o n Valley near Dehra D u n in India, but this work has limited application away from similar environments such as the Chitawan valley i n Nepal.  Estimates.  A number of observers have tried to estimate surface erosion for specific areas in Nepal rather than to measure it, either by making "informed guesses" using the published figures noted above (e.g. Junor (1981), for the Bagmati catchment;  Carson, i n  press), or else by attempting to use the universal soil loss equation of Wischmeier and Smith (1978) (e.g. Fetzer and Jung 1978/79; Jahn et al.  1979;  Shakya 1982), despite  the paucity of input data for the equation" and the pitfalls involved in use of the " Rainfall erosivity (R) has been calculated by Fetzer and Jung (1978/79) for Kathmandu Airport, and by Impat (1981) for one season in the Phewa valley near Pokhara. A t Kathmandu Airport erosivity calculated by the K E >25 m m / h r method (Hudson 1971) for the period 1971-1976 had a mean annual value of 72, which is moderately low. However at Banpale i n the Phewa watershed erosivity in 1979, calculated by the same method, was 432. Erosivity has been calculated for other stations i n Nepal such as Pakhribas Agriculture Centre but the results have not been widely circulated. It is generally agreed that precipitation intensity decreases with altitude in Nepal (see e.g. Brunsden et al. 1981; L . R M . P . 1983).  24  model outside the area where it was developed (for a discussion of these Wischmeier 1976,  1984; Surface  Elwell  see  1984).  erosion, gullying  and  forest  clearance.  The effect of forest clearance on soil erodibility has been studied by Chakrabarti (1971), who investigated water-stable and established by deforestation.  aggregates in soils in eastern  Nepal,  that forest soils there were considerably less erodible than soils exposed Other effects of forest clearance on soils include compaction due to  trampling by livestock, reduced infiltration capacity, and consequently  higher  surface  runoff. Burning also exposes soils to erosion. In India, Dwivedi (1980), quoting Dalai et al.  (1961), reported increases i n splash erosion of up to 4000% in forests of  robusta subject to annual leaf litter burning. N o comparable  Shorea  data on the effect of fire  are available for Nepal, but Impat's (1981) figures for runoff from the erosion plots in the Phewa catchment  demonstrate  the influence of overgrazing and trampling on  runoff (Appendix 2). M e a n monthly runoff on protected pasture varied from 5.5% 17% of total precipitation, but runoff on overgrazed land ranged from 11% to (Impat  1981). This high rate of runoff causes sheetwash,  1981;  Caine and M o o l  53%  rilling, and eventually,  gullying, which has been noted to be severe in many parts of Nepal (Laban Brunsden et al.  to  1978;  1982), and which is intimately associated  with  communal grazing areas and marginal agricultural l a n d . 27  The effect of surface  erosion and gullying is to reduce the productive capacity  of the land affected, both by removing nutrients and degrading soil physical properties (most importantly texture and depth), and by altering the topography. Additionally, The variability of the Himalayan environment is well illustrated by comparing Impat's runoff figures with those found by Pandey et al. (1983) i n a study i n the Kumaun Himalaya near Naini Tal. Average seasonal overland flow and soil loss in very small catchments varied from 0.44% of total incident rainfall and 0.025 t/ha respectively under dense forest, to only 0.60% and 0.081 t/ha on a site with "soil deposition" (Pandey et al. 1983, p. 25). The authors concluded that they were dealing with subsurface flow systems. 27  25  downstream effects can be severe,  with sediment deposition affecting cultivated lands. In  India, Champion and Seth (1968, p. 397)  have described these consequences:  "The damage done to cultivation below the eroded slopes by the unchecked run-off, floods, deposition of sand, gravel and boulders is immense and notorious in the Siwaliks of the Punjab, Himachal Pradesh and Jammu." Although some control is necessary, probably have a negative dependence water (see  a total cessation  of all erosion in Nepal would  effect on agricultural output in the hills. This is due to  the  of lowland terrace production on the nutrients brought i n by irrigation 2.8  for a description of the  agroecosystem).  The marginal agricultural areas responsible  for much of the soil loss make a  minimal contribution to the rural system. In India, Whyte (1957) has described main functions of the grossly overgrazed public pastures  the  as being an "exercise place"  for livestock, and the same description has been applied in Nepal (Y.Khatiwada, pers. comm.). In central Nepal, demand for grazing on these areas peaks during the dry season, and again after the monsoon before crops are harvested (Fox  1982,  1983). The  grazing is most important to households with little private land. Equally, it is these same poorer households which are most dependent on public lands for fodder and firewood. The lack of local control or participation i n management  of these areas  results in a classic pattern of over-exploitation close to settlements and under-use  at  greater distances. Unable to withstand the pressure, the forests retreat Physically, the processes involved in forest repeated  lopping and consequent  layer which inhibits regeneration,  degradation and clearance include  reduction i n vigour, felling, the removal of the litter browsing and trampling by livestock, and fire, both  accidental and purposely set to stimulate a flush of grass. Most of these processes are, directly or indirectly, related to the maintenance of a Himalayan agroecosystem 1984a, 1984b; Singh et al.  of livestock. Studies of the energetics  immediately to the west of Nepal (Pandey and Singh  1984)  have emphasized the very high energy subsidy  26  supplied by the forest to cultivated areas, and a large part of this subsidy is in the form of fodder for livestock . Wyatt-Smith (1982) has analysed data on land use and 28  livestock populations i n the M i d d l e Mountains, largely from Tansen and Pokhara in central Nepal, and determined that, to sustain current activity, the average 5-6  persons, with an average  farm holding of 1.25  ha, requires 0.4  family of  ha of land for  timber, 0.3-0.6 ha for fuel, and 3.5  ha of land for fodder. Thus, rather than  search  of cultivation being required to explain  for fuel or the encroachment  deforestation  (see  Bajracharya  1981,  the  1983), it is possible to argue that the process  can  be ascribed largely to the demands caused by livestock. This view is supported by the presence,  i n many parts of the hills, of fallow or  abandoned rainfed terraces (Gilmour 1984). These can be explained as being due to a fertility limitation, which has acted to reduce the area under cultivation rather than expand it ( L . R . M . P . 1983). Insufficient compost is available from remaining woodland to maintain yields on the total cultivated acreage, and so resources  are concentrated  on  the most productive fields and marginal areas abandoned. O n the basis of this evidence it is possible to propose the following sequence : increasing population and 29  static agricultural technology cause food shortages; intensify agricultural production, with associated  attempts  are made to expand and  increases i n the livestock population.  The demand placed on the forest by the larger herd initiates a decline in forest  area.  The reduced forest area is unable provide sufficient nutrients to support an increase  in  the cultivated area, and this remains static. Further decreases in forest cover result in the abandonment  of marginal fields, and the only land use type to expand is the  overgrazed and barren chararu Thus, far from hunger resulting i n an increase  i n the  cultivated area, it may serve, indirectly, to reduce it. The unmanaged, communally  F o r an analysis of the energy budget of slash and burn agriculture, jhum, north-eastern India, see Toky and Ramakrishnan (1982). A similar sequence was first presented in L . R . M . P . (1983). 28  29  in  27 grazed areas become the focus for surface erosion.  The following three chapters describe a case study of erosion processes in the Phewa valley near Pokhara i n central Nepal.  Chapter 2 THE  STUDY AREA: T H E  PHEWA  VALLEY  2.1 LOCATION The study area is located in the M i d d l e Mountains of Nepal (Figure  2)  immediately to the south of the Annapurna massif, which rises to some 8078 m 35 km north of Pokhara. It comprises the catchment  of the Andheri Khola and Sidane  Khola (khola N e p . river or stream) which feed Phewa Tal (tal  Nep. lake), Nepal's  most prominent lake (Figure 3). Estimates of the area of the watershed above Pardi D a m at the outlet of the lake vary from 113  km  2  (Fleming 1978)  the  to 126 k m  2  (Anon. 1980). Planimetry of a topographic map of the watershed at a scale of 1:25,000 ( I . W . M . 1979)  gave a value of 121.8  k m . The Phewa valley is situated at 2  the western end of the Pokhara valley, an intra-montane basin drained by the K h o l a ("White R i v e r " )  30  and its tributaries, and famous in Nepal for its impressive  river terraces through which the Seti Khola has cut a spectacular  2.2 T H E G E O L O G I C A L  2.2.1  Seti  gorge.  CONTEXT  THE POKHARA  BASIN  The Pokhara Basin (Gansser  1964)  extends for nearly 42 k m from Barbhure i n  the north-west to Dobhan in the south-east (Fort and Freytet 1982), and is bounded to the south by the Barsami Ridge (Figure 4), a fault-scarp ridge which forms the northernmost expression of the Mahabharat L e k h (lekh  Nep. range). This boundary is  comparatively straight arid runs parallel to the M a i n Central Thrust zone to the north. The thrust zone separates the Himalayan Meta-sediments (mostly  meta-sandstone,  So-called due to its milky colour caused by a suspended load of fine particles transported from high on the Annapurna massif. 30  28  calcareous  Figure 3.  Phewa Tal: view from the west towards Pokhara and the Pokhara Basin.  30  Figure 4. The  study area: sketch  map of the  local  environment  31  quartzite, chlorite-sericite schist, and crystalline schist) which make up the hills surrounding the basin from the Himalayan Gneiss o f the Annapurna Range, and is one of the major tectonic lines i n the Himalaya (Yamanaka  et al. 1982). A possible  sequence for the genesis and tectonic development of the Pokhara Basin is described and illustrated i n Yamanaka et al. (1982, pp. 132-133), and the consensus Basin is a a depression of tectonic origin (Fort and Freytet northwest-southeast  anticline depressed along its axis (Sharma  is that the  1982), situated alorlg a 1975, Pecher 1978). It is  apparent that the Seti K h o l a is an antecedent river whose current course is controlled by the general W N W - E S E structural trend of the Himalayan Range. Quaternary  deposits  Because the Middle Hills of Nepal form a zone of subsidence relative to the Great Himalaya and the Mahabharat Range (Yamanaka 1982), Quaternary deposits are found preferentially i n this region (for example i n the Pokhara Basin; i n the Kathmandu Valley; i n the Kusma Basin of the K a l i Gandaki (Fort and Freytet  1982);  in the valley of the Marsyangdi Khola (Yamanaka 1982); along the Buri Gandaki and Ankhu  Khola i n the Arughat Basin (Thouret 1977)). The Pokhara Basin is  detrital material (pebbles,  filled  with  gravels, sands, and muds), mainly calcareous (marbles) and  with a minor percentage of sedimentary rocks (gneiss, micaschists) (Fort and Freytet 1982). This  filling  has been interpreted variously as the result of lacustrine  sedimentation (Akiba 1980), secondary deposition of morainal material (Hagen 1969), glacial, fluvioglacial and partly lacustrine deposition (Sharma  1975; Sharma et al. 1978),  fluvial and fluvioglacial deposition (Gurung 1970; Dollfus and Usselman 1971; Hormann 1974; Yamanaka et al. 1982), and colluvial, alluvial and fluvioglacial action (Fort and Freyet  1979, 1982). Relative stratigraphies for the deposits have been proposed by  Dollfus and Usselman (1971), Hormann (1974), Sharma (1975), Fort and Freytet (1979, 1982), Akiba (1980), and Yamanaka et al. (1982), and maps are presented i n Gurung  32  (1970), Dollfuss and Usselman (1971), Sharma et al. (1978), Fort and Freytet (1982), and Yamanaka et al. (1982). The calcareous gravelly conglomerates  which form the majority of the basin  deposits have a well defined geological origin (mainly Nilgiri and Larjung limestones (Fort and Freytet 1982) which presently crop out at the summit of Annapurna H i m a l (Bordet et al. 1971). Two deposits, the Pokhara and Ghachok Formations, have volumes estimated to be 5.5 k m and >8.7 3  km  3  respectively (Yamanaka et al. 1982).  It is generally agreed that they were transported to the basin by intense and rapid re-working of morainic materials, possibly associated with the failure of a glacial or landslide dam (B. Carson, pers. comm.). The rapidity of the aggradation  process  resulted i n the damming of valleys adjacent to the basin (Gurung 1970), causing the development of small lakes (Phewa T a l , Begnas T a l , Rupakot Tal). This may have occurred at least twice during the late to middle Quaternary (Yamanaka et al. 1982). The most recent aggradation, which created the Pokhara Formation, is thought to be Holocene, and radiocarbon dating indicates ages of 1100 to 500 years B.P. (Yamanaka 1982;  M . Fort pers. comm.). Such a recent event is within the reach of oral history,  and a local guidebook does indeed recall a "lost city" under Phewa Tal (B. Carson, pers. comm.). Currently the Seti K h o l a is i n an erosional phase and has cut a channel to bedrock i n some locations i n the Pokhara valley. There is evidence of continued tectonic movement, with the whole basin tilting to the south (Yamanaka  et  al. 1982).  2.2.2 T H E P H E W A  VALLEY  The geology o f the Phewa valley has been described briefly by Mulder (1978a) and Fleming (1978). Heming's account is largely based on the work of Remy (1975) who mapped the geology of western Nepal at a scale o f 1:506,880. The watershed  33  area has been mapped more recently by the Land Resource Mapping  Project  ( L . R . M . P . ) at a scale of 1:125,000 and appears on geology sheet N o . 62  P-D  (Topographical Survey Branch 1984a). The Phewa catchment (Figure 5) is underlain by two lithologic units with a general east-west  strike and a moderate (15° -  30°) dip to the southwest  The  northern part of the catchment is formed i n grey phyllitic schist or phyllite consisting mainly of micas and chlorites which form the high ground of the Kaski ridge. The phyllites are weakly bedded, low grade metamorphic rocks which dip to the south at approximately the same angle as the topography. The southern part of the consists of t a l c - r i c h , red phyllitic schist (Mulder 1978a) and quartzose or  catchment carbonaceous  schist of a higher metamorphic grade (biotite) than that to the north (Fleming 1978). Bedding structures are moderately strong, and dip is again generally to the southwest and into the slope. Quartzite schist crops out at the western end of the watershed, forming cliffs. Soft grey talc schist appears i n several locations including the saddle of the Kaski ridge at Naudanda and i n the valley of the Andheri Khola  between  Pamdur and Deorali. It consists of very small, weak platelets and is associated  with  active slope failures. A major fault (probably a continuation of the fault on the southern side of the Pokhara valley) runs through the centre of the watershed. In common with the rest of Nepal, numerous minor faults and local deformation are imposed on this macro-structure, causing zones of weakness which have been exploited by erosive forces such as the Andheri K h o l a (Figure 5).  2.3  TOPOGRAPHY  The elevation of the watershed varies from 790 m at Pardi D a m at the outlet to Phewa Tal lake to 2520 m at Panchase some 17 k m away at the western end of  Figure 5. The  Phewa Valley: sketch  map showing places mentioned  in the  text  35  the watershed. Slopes are generally steep to very steep, but with a virtually flat valley floor  caused by sedimentation above the lake (Figure 6). Above this level of debris  accumulation the tributary valleys are V - s h a p e d with gentler slopes above, indicating rejuvenation. The topography is illustrated in Figure 7, a schematic cross-section of the central part of the watershed, and is readily visible in Figure 8, a stereopair of the same area. A summary description of the major landforms is given below (after Impat 1980): (1) Scarp slopes: north aspect gradients 60-120%, mainly forested, generally stable, little mass wasting. (2) Toe slopes and spur ridges: mainly north aspect, gradients 30-100%, 60% forested, 20% cultivated, 20% grazing, some large landslides. (3) D i p slopes: south aspect generally plane surface conforming to bedrock dip but with marked erosion ampitheatres, gradients 30-60%, mostly cultivated, some grazing, forest on steeper slopes, high level or erosional activity including gullying, slumping and debris avalanches. (4) Toe slopes: south aspect, gradients 30-60%, dissected slopes above colluvial accumulations, 60% cultivated, 30% grazing, 10% forest (5) Bottom lands: decrease in width towards headwaters, coalescing fan component radiating from side drainages at gradients of 5-10%, riverplain gradient 1-3%, all cultivated and subject to damage by meandering and sediment deposition. A slope categories map is presented i n Fleming (1978), and indicates that  60%  of the watershed has slopes of between 11° and 31°, with an average of 22°. Slope categories are shown in Table 3.  2.4  CLIMAXE The climate of the study area is humid sub-tropical to humid temperate, with a  dry season i n the winter. Rainfall seasonality is due to the monsoonal conditions of the Indian sub-continent M e a n daily temperatures vary between 12.6° C in January and 25.1-25.2° C  C i n July and August i n Pokhara (elevation 827  in January and  m), and between  19.8° C in July and August at Lumle (elevation 1675  8.6°  m) (Figure  36  Figure 7.  Schematic cross-section of Phewa Valley near Pame.  37  Figure 8. Stereogram of centra] part of the Phewa Valley. North at top. Note: (1) Steepness of topography: gullying at top. (2) Mass movement catchments (arrowed). (3) Valley floor; braided channel at left and meandering channel near lake. (4) Delta growth in lake. (5) Remaining forests on steeper slopes.  38  Table 3. Slope  categories of the Phewa  Slope  (%)  < 10 10-20 20-40 40-60 60-80 80-100 > 100 Lake  (degrees)  Tal  catchment . 1  Area (% of total)  10.7 9.3 31.4 29.8 9.9 5.0 0.2 3.7  < 6 6-11 11-22 22-31 31-39 39-45 > 45  Modified from: Flerning 1978, F i g . 2.3  9).  31  Some 85% of annual precipitation falls during the four monsoon  months  (June-September, Figure 9). M e a n monthly rainfall for seven stations i n the locality is shown i n Table 4. M e a n annual precipitation is elevation-dependent  and varies from  3724 m m at Pokhara Airport to 5140 m m at Lumle (13 years record both stations). A regression  o f mean annual precipitation on elevation  for the 7 stations  gave the  following relationship, significant at the 95% level (Figure 10): precipitation (mm) =  2176 +  1.64 elevation (m)  (r=0.847)  Meteorological stations i n the area are Pokhara Airport ( N o . 0804) and Lumle (No. 0814), both operated by the Department of Irrigation, Hydrology and Meteorology. Additionally, rainfall data is available for Pokhara Hospital (the Shining Hospital, not the new Gandaki Zonal Hospital), from 1967 to 1976 (no data for 1969, 1975); from Pokhara Agricultural F a r m for 1978 and 1979; and from three stations i n the Phewa watershed operated by the Department o f Soil Conservation and Watershed Management: Toripani (elevation 1340 m) from 1977: daily gauge; Banpale (elevation 1405 m) from 1978: daily gauge and weekly recording gauge; and Tamagi (elevation 1615 m) from 1978: daily gauge and weekly recording gauge, moved to Sidane (elevation 1560 m) i n 1982. In general raw meteorological data i n the area is unreliable and requires interpretation and smoothing before analysis. 31  Pokhara Airport: temperature Lumle: temperature Pokhara Airport: rainfall Lumle: rainfall  MONTH  Figure 9. Temperature and rainfall regimes at Pokhara Airport and Lumle. Data from Department of Irrigation, Hydrology and Meteorology records.  40  Figure 10. Regression of mean annual precipitation on elevation for 7 in and near the Phewa Watershed. Data from Table 4.  stations  41  The mean annual precipitation for the whole watershed, calculated from this regression and a hypsometric curve , is 4202 m m . The area is one of the wettest i n 32  the country, with Lumle having the highest annual rainfall of any station in Nepal. This is probably due to the generally low elevation of the Middle Mountains lying between Pokhara, at the base of the Annapurna massif, and the Terai, thus allowing monsoonal air-masses  to approach the H i g h H i m a l unimpeded (Stainton  region does not appear to be affected  by the winter (west) monsoon. A l l precipitation  falls as rain, and frost is only recorded above The absolute  about 1500 m.  maximum 24 hr precipitation officially recorded in the area is  mm at Pokhara Airport in August 1979, mm i n July 1980  from a daily gauge at Tamagi (elevation  1340  279  m) operated by the  It is also reported  m m fell i n one and a half hours at Lumle i n July 1974  preliminary set of intensity-duration-frequency 7-10  278  although there is a less reliable report of  Department of Soil Conservation and Watershed Management 233  1972). The  that  (Wormald 1976a). A  curves for Pokhara Airport based on  years of weekly recording gauge data is presented  in Figure 11. The  curves  indicate 1 hour rainfall intensities of 64 m m , 78 m m , and 91 mm for, respectively, return periods of 2, 5, and 10 years. The and 276  24 hour intensities are 168  m m for the same return periods, with a tentative  312 m m suggested  25 year period. It is not known whether rainfall intensities increase the locality, but, i f they are related to the increase  m m , 234  mm,  for the  with elevation i n  in annual precipitation with  elevation, some very high intensity events can be expected in the upper parts of the catchment  The reported and predicted rainfall figures may be compared with  highest recorded 24 hour precipitation i n Nepal, 505 of Kathmandu, on 25 August 1968  the  m m at Gumthang, east-north-east  (Nayava 1974).  Plot of area against elevation. SeejStrahler (1952) for an application of hypsometric analysis to erosional topography. 32  Table 4. Mean monthly rainfall for 7 stations In the Pokhara area Station  Pokhara Farm  Elevation (m) Station  Agric.  Pokhara Airport  Pokhara Hospital'  Toripani'  Banpale  827  918  1340  1405  0804  0803  2  Tamagi  5  Lumle'  6  1  792  Number  1615  1675 0814  Month Jan.  6  20  22  36  36  26  29  Feb.  33  31  25  60  47  20  42  Mar.  33  55  60  76  145  54  52  Apr.  119  116  106  124  68  107  121  May  295  352  306  352  307  394  293  June  364  620  764  704  679  638  830  July  590  908  955  1190  953  1442  1383  Aug.  1233  797  772  1365  883  981  1325  Sep.  236  573  588  761  588  446  784  Oct.  195  217  212  258  213  173  223  Nov.  7  23  21  39  76  19  50  Dec.  32  12  3  44  21  7  8  TOTAL  3143  3724  3834  5009  4016  4307  5140  (3578,  1  Pokhara  Agricultural Farm:  2 years  !  Pokhara  Airport:  record:  '  Pokhara  Hospital (the Shining  • Toripani:  5  13 years  to 6 years  record:  8 years record:  Banpale:  1 to 5 years  record:  1978-1983.  3 to 5 years  record:  1978-1982.  ' *  Lumle:  13 years  record:  1970-1982  supplied  by the Dept. of Irrigation,  Pokhara  Hospital: mean  based  on 19 years  regression  analysis  annual  record  1967-1976  (Note:  Data supplied direcdy  Hydrology and Meteorology  rainfall given as 3578 mm based  from  of elevation  1978-1979. (no data  for 1969, 1975).  1977-1983.  • Tamagi:  5  !  1968-1980.  Hospital):  record:  3689)  1958-1976,  but with  and precipitation  1966,  (Figure  8).  from  Lumle Agriculture Centre  does  not correspond  with  data  for Lumle  in all months). on 18 years  record  (Water  1969 and 1975 missing (Impat,  and Energy  pers.  comm.).  Commission 1982), This  last  figure  and 3689 mm  was chosen  for the  43  10 . 15 Minutes  Hours DURATION  Figure 11. Preliminary intensity-duration-frequency curves for Pokhara Airport based on 7-10 years of data.  44  Relative humidity, sunshine duration, and Class A pan evaporation data are available for both Pokhara and Lumle, but many observations presented  figures  are missing. Shah  (1980)  on solar radiation and "equilibrium potential evaporation" for Pokhara.  Potential evaporation exceeds precipitation from November to March and under natural conditions very little moisture is available for plant growth (Shah  1980). W i n d run is  recorded at Lumle.  2.5 WEATHERING AND SOILS The climate of the Phewa Valley puts it well into the zone of "intense weathering"  described by the Land Resource  Mapping Project  chemical  ( L . R . M . P . 1983, after  Peltier 1950). The underlying phyllites tend to weather rapidly where water can penetrate joints and bedding planes, forming  fine-textured  saprolite with a high  proportion of illite clays (Brian Carson, pers. comm.). Soils developed on phyllites tend to be reddish, have strong clay accumulation i n the subsoil, are mildly acidic, moderately  fine-textured,  and n o n - stony ( L R . M . P . 1983). The quartzites and schists  tend to be more resistant and result i n coarser debris and shallower soils. The rapid removal of weathering products by solution and erosion allows weathering to proceed uninhibited (Oilier 1984)". The intensity of erosional processes means that soil development is largely an expression of stability and position i n the landscape with horizonation and structure developing best i n the most, stable locations. Untransported regolith thicknesses  are  generally less than 3m, but may be much more locally, especially around faults. Colluvial deposits can exceed 15 m i n depth, as at the entrance to the mass movement catchment bedrock  near Pame, and it is possible that  filled with colluvium may be present  "wedges"  or hollows i n the  Such wedges have been implicated i n  Strakhov's (1967) concept of erosion suppressing weathering does not seem appropriate i n this environment 33  45  slope failure i n mid-latitude environments such as the west coast of North America (see,  e.g., Dietrich et  al.  1982).  The soils i n the valley have been surveyed and described at 1:50,000 by Mulder (1978a) and the Land Resource  Mapping Project (Topographic Survey Branch 1984b),  and at 1:25,000 by Impat (1980) . The majority of the watershed falls into the 34  bedrock controlled unit (no. 12) of the L . R . M . P . : mountainous terrain with dominant slopes and <50  cm to bedrock (Figure 12).  steeply to very steeply sloping  > 3 0 ° , loamy skeletal texture , well drained, 35  Soil development is better expressed in unit 11,  mainly found on the more stable upper slopes on the north side of the valley: moderately to steeply sloping mountainous terrain, dominant slopes < 3 0 ° , loamy skeletal texture, moderately well to well drained, and 50 cm to 1 m to bedrock. The valley floor and fans with slopes of < 5 °  alluvial  (unit 9) tend to be extremely stony with  dominant textures varying from fragmental sandy to loamy/bouldery ( L . R . M . P . 1983). Soil p H is generally low, about 5.1-5.5 (Impat  1980 ), decreasing to 4 on the 36  valley floor where ammonium-sulphate fertilizers have been used (Shah  1980, citing  V a n de Putte 1979). This is sufficiently low to reduce nutrient availability, although under the anaerobic conditions of paddy rice production p H may rise to 6.5, obviating any availability problem ( L . R . M . P . 1983). Although the cation exchange capacities  of  soil minerals in the area can be appreciable (e.g. illite, which may have values of 10-40  meq/100 g), cation exchange capacities i n the soil are also low, approximately  10 meq/100 g. The low organic matter content of cultivated soils, typically 1.5-2% in M u l d e r (1978a) carried out a reconnaissance survey of the northern part of the catchment; Impat (1980) extended Mulder's survey to include the southern part of the watershed. His map shows soil units differentiated by parent material, depth class, slope class, and runoff class, rather than by soil series or subgroup. The recent L . R . M . P . survey covered most the country. Recognizing the extreme spatial variability of soils i n Nepal their legend (Figure 12) is based on repeating landscape units. Rock fragments make up 35% or more by volume; enough fine earth to fill interstices > 1 m m ; the fraction finer than 2 m m is loamy as defined for the loamy particle size class (Soil Survey Staff 1975). p H measurement method not stated. 3 4  3 5  3 6  Figure 12. Land systems of the Phewa Valley. Legend overleaf. After: Topographical Survey Branch (1984b).  Land  MIDDLE MOUNTAIN REGION  Land System 9  Systems  Legend  Precambrian t o Eocene p h y l H t e s , q u a r t z i t e s , s c h i s t s , l i m e s t o n e s , g e n e r a l 1y d e e p l y w e a t h e r e d . S u b t r o p i c a l t o Warm T e m p e r a t e .  Unit  Dom i n a n t  So i 1 s  Dominant Slopes (deg)  Domlnant Texture  and g n e i s s e s .  Seasonal Range o f Depth t o Watertable  Drainage  Landform  Land  Alluvial Plains and Fans (depos i 11onal)  9a  river channel  Psamments Ustorthents  <1  Fragmenta1 Sandy  0-2 m  var1able  9b  alluvial p i a 1ns  Ust1fluvents F1uvaquents Ustachrepts  <1  Loamy/ Bouldery  0- 2 m  wel 1  9c  al1uv1al fans  Ustachrepts Haplustalfs  1-5  Loamy/ Bouldery  1- 15 m  wel 1  0-5  Loamy  >2 m  wel 1  10  Ancient lake and R i v e r Terraces (Tars)  Typic & Rhodlc Haplusta1fs Ustochrepts  1 1  Moderately to Steeply Sloping Mounta1 nous Terrain  Typic, Rhodic Udic, Anthroplc Subgroups o f Ustochrepts Dystrochrepts Haplumbrepts  <30  Loamy/ Skeletal  >50cm moderately t o b e d r o c k wel1 t o wel1  12  S t e e p l y t o Very Steeply Sloping M o u n t a i nous T e r r a 1n  L i t h l c Subgroups >30 o f 11 a n d U s t o r t n e n t s  Loamy/ Skeletal  <50cm t o b e d r o c k wel 1  48  the plough layer, and the presence  of large amounts of coarse material, reduces the  contribution to overall cation exchange  capacity which these minerals can make (S.  Burton, pers. comm.).- Organic matter content reflects land use, decreasing rapidly following forest  clearance.  2.6 VEGETATION The vegetation associations of the country south of Annapurna and H i m a l  Chuli  have been described by Stainton (1972, p p . 32-34), and a climax or potential vegetation map of the area at a scale o f 1:250,000 has been produced by Dobremez and Jest (1974). Although now much modified by man, remnants o f the original  forest  cover of the Phewa valley remain, and indicate that from 850 to 1500 m elevation (most of the catchment) chilaune) (Shorea temperate  a subtropical wet forest composed of Schima  and Castanopsis  indica  wallichii (Nep.  (Nep. katus) would have predominated, with Sal  robustd) on dry southern exposures at low elevations; above 1500 m a lower mixed broadleaved association including Michelia-  have existed with  Lauraceae- Lithocarpus  Quercus associations above about 2000 m. Alder, (Alnus  would  nepalensis),  is common on moist, disturbed sites. A n annotated checklist of plants i n the area has been provided by Wormald (1976b). W i t h i n the watershed the most detailed observations of forest characteristics  have  been made by Levenson (1979), i n the course o f a study on fuelwood utilization. One of Levenson's most interesting comments is that many of today's dominant species were components of the understorey when Stainton made his observations i n the 1950's and 1960's.  37  Both Stainton (1972) and Dobremez and Jest (1974) emphasize the  wallichii/Castanopsis  indica  Schima  forest as being the climax i n the area, but Levenson (1979)  postulated a "disturbance secondary type species association" comprised mainly of  37  In this area, 1954, 1963, 1966, 1967 (Stainton 1972).  49  Daphnephyllum  himalayense  (Nep. chandan)  and Symplocos  ramosissima.  The gradual  degradation of the forest i n the valley has recently been confirmed by the Land 1  Resource Mapping Project:  although a large proportion of the watershed is shown as  being forested (Figure 13), the legend of the L . R . M . P . maps indicates that virtually all stands are composed of "immature or small size timber material", with crown densities of 40%-70% (Topographical Survey Branch 1984c). Most forest stands are surrounded by a leech-infested belt of lopped trees, shrubs, and scrub. It is possible that the forests on the Kaski ridge were burnt for military purposes several centuries ago (J.B. M c D o n a l d , pers. comm.). Forest biomass production i n the area has been estimated at 12.5 m / y r i n 3  Bhadaure panchayat  38  by Levenson, based on plot measurements and taking into account  species composition, differences i n wood specific gravity (averaged at 0.517 stocking rate, and village preference  g/cm ), 3  for fuelwood species, and assuming that all species  follow a growth pattern similar to that of Daphne  himalayense  (linear after 14.3 years  of age) (Levenson 1979). Wormald (1976c) estimated that the mean annual increment for presently unmanaged woodland around Lumle Agriculture Centre was 15-20 m / y r , 3  and could be raised to 25-30 m / y r by management  Yields of leaf material for use  3  as fodder are uncertain. Farm surveys indicate approximately 39-150 kg/tree/yr (fresh weight) according to species (Shah  1980) . Preferred fodder trees are  nemoralis  39  lakoocha  (Nep. badhar), Ficus  (Nep. dudhilo),  and Ficus  pakhuri),  although some 35 other species are also used (Shah 1980).  Indigenous grass species are mainly subtropical, and include articulatus, Chrysopogon  Dicanthium montanus,  spp., Bothriochloa  intermedia,  Andropogon pumilus,  Imperata  Artocarpus  glaberrima (Nep.  Chrysopogon  B. pertusa,  Themedia  orlindrica,  Sporobolus  anathera, spp., Setaria  Administrative area at western end o f the Phewa watershed with a population of between 4000 and 5000 and an area o f approximately 2000 ha. Fodder tree utilization and management are discussed extensively i n Shah (1980), and i n the annual reports o f Lumle Agriculture Centre. 38  3 9  I  I  I  I  Figure 13. Forest  I  I  I  cover and land use in the Phewa Valley. Legend overleaf.  After: Topographical Survey Branch (1984c).  u> o  Forestry  Legend  Land  VALLEY  COVER TYPE H  Hardwood  S  Shrub;  - 75% o r more o f t r e e s p e c t e s a r e hardwoods  shrub  v e g e t a t i o n which may  I n c l u d e hardwood r e g e n e r a t i o n  SPECIES TYPE Tropical Sal  Types  - Shorea  DMB  robusta  Tropical  TMH  Temperate Type - Deciduous mixed broad  leaved  - P r o t e c t i o n F o r e s t - f o r e s t s w i t h management s1te f r a g l 1 1 t y  problems due to  CROWN DENSITY Expressed 2. 3. 4.  as a p e r c e n t a g e  of t h e a r e a  covered  10-40°/. 40-70% >70%  MATURITY  CULTIVATION  V a l l e y F l o o r s I n c l u d i n g T a r s , F o o t s l o p e s and/or A l l u v i a l Fans which a r e t o o s m a l l t o map T a r s , A l l u v i a l Fans and/or Lower F o o t s l o p e s GRAZING LANDS  G  NON  S u b - T r o p i c a l Zone <1000m Warm Temperate Zone 1000m - 2000m  1 2  Lake Urban  Rice R1ce MaIze-RIce Ma 1ze-R tee Maize or M i l l e t Ma Ize Maize Mixed D o u b l e monsoon c r o p  T2ae(r)  M - M a t u r e t o o v e r m a t u r e t r e e s have r e a c h e d r o t a t i o n age o r saw timber s i z e timber  size material  at l e a s t  Fallow Cerea1 Fa 11ow Winter crop Fallow Mustard Cerea1 In b r a c k e t s  (.)  Level t e r r a c e s 50-75% c u l t i v a t e d Rice/Fallow, R1ce/Cereal, Maize/Rice double cropped before winter crop  estimated  HILLSLOPE  TYPE LEGEND SAMPLE  Level  HTMH3I  Intense: Medium: Light:  Hardwood T r o p i c a l mixed hardwoods 40-70% crown d e n s i t y Immature  V F  AGRICULTURAL LANDS  TYPE LEGEND SAMPLE  CLASS  I - Immature o r s m a l l  Legend  DOMINANT CROPPING PATTERNS MONSOON SEASON WINTER/DRY SEASON  mixed hardwoods  CONDITION TYPE PF  Use  CULTIVATION  Terraces 75-100% c u l t 1 v a t e d 50-75% c u l t i v a t e d 25-50% c u l t i v a t e d  a e u r J k 1 m  52  glauca,  Heteropogon  contortus, Digitaria  spp., Axonopus affinis, and Paspalum  1980). Legumes are rare, although some species of Desmodium  spp. (Shah  occur in moister areas.  The botanical composition and productivity of the sward are severely affected grazing pressure. In protected areas bunch type grasses such as Andropogon, and the Chrysopogon Artemisia  group survive, but in heavily grazed areas Imperata  spp. are spreading (Shah  1980), with bracken, Pteridium  by  Themeda  and  aquilinum,  increasing  in areas of scrub. The effect of enclosure on pasture productivity is marked. Several areas of overgrazed land in the watershed have been fenced off for trials by the Department of Soil Conservation and Watershed Management, and in these paddocks pasture yields are estimated to have increased from 1.2 weight) (Shah  2.7  LAND  t/ha/yr to 4.1 t / h a / y r  1980), simply due to the exclusion of livestock.  USE  In the Kathmandu area Nepalis recognize six major land use types: (irrigated terrace);  bari and pakho-bari  (rainfed terrace);  barren to be cultivated); charan (grazing land); ban (Johnson  (fresh  et al.  1982;  khet  bhir (land too steep, rocky or  (forest);  and gaun (settlement)  B. Carson, pers. comm.). Land utilization in the Phewa watershed  follows much the same pattern, with the inclusion of kharbari,  privately owned fenced  pasture used mainly for thatching grass. Land use i n the watershed has been mapped several  times , 40  and use categories  as interpreted by Fleming (1978) are shown in  Table 5. Land use interpreted from the same photos by the Land Resource  Mapping  Project is shown i n Figure 13. It should be noted that this drawing gives an inadequate impression of the extent of grazing land in the catchment,  which is largely  B y J. Kraayenhagen and co-workers, based on 1:15,000 scale air photos taken i n 1972 and on cadastral maps, and reported i n Section 5.5 of the Management Plan for the Integrated Development of Phewa Tal Watershed (Anon. 1980); by Impat (1980), based on 1:25,000 scale air photos taken i n 1978 for Lumle Agriculture centre; and by Fleming (1978) and the Land Resource Mapping Project (Topographical Survey Branch 1984c) from 1:50,000 scale air photos also taken i n 1978. 4 0  53  Table 5. Land use categories in the  Phewa  Land use  Total'  Source:  catchment  1  Area (ha)  Forest Scrub Cultivated Open grazing Enclosed pasture Water Gullies/landslides Townsite  1  Tal  % total  2935 1070 5410 1193 71 428 73 156  25.9 9.4 47.7 10.5 0.6 3.8 0.7 1.4  11,336  100  area  Fleming 1978  in parcels too small to be included i n the generalized L . R . M . P .  map. Fleming's  figures  indicate that just under half the watershed is cultivated, about a quarter is still i n forest , 41  and scrub and open grazing  predominates on the south-aspect  42  account  for most of the remainder. Agriculture  slopes on the north side of the valley. The  remaining forest is largely on the southern side of the watershed where slopes are too steep to allow successful  terracing. Most of the grazing land is in the northern and  western parts of the watershed on slopes averaging 22° (Fleming 1978).  2.8 T H E AGROFCOSYSTRM The Middle Mountains of Nepal support a mixed farming system with agriculture predominating between  settled  1000 and 2000 m. Livestock are crucial to this  system, providing manure, draught power (cattle), food for domestic consumption (buffalo milk, meat from sheep, goats, pigs, poultry), products for o f f - f a r m sales  41 42  Forest names are given i n Levenson (1979, F i g . 1). Names of grazing areas are given i n Shah (1980, Table  5.10).  54  (mainly ghee, clarified buffalo butter), and are also essential for some religious ceremonies. The principal crops are paddy rice (Oryza (Triticum  sativa var. indica),  vulgare), maize (Zea mays), and ringer millet (Eleusine  wheat  coracana). A t lower  elevations high temperatures and irrigation allow double cropping, but at higher elevations this is limited by lower temperatures, greater cloudiness, and water shortages. A simplified schematic model o f a Nepalese hill farming agroecosystem showing the main energy and nutrient pathways is presented i n Figure 14. Agriculture i n the Phewa valley has been the subject o f intensive scrutiny by the H i l l Agriculture Development Project (Ministry o f A g r i c u l t u r e / F A O / U N D P ) , and an introduction is available i n V a n de Putte (1979), Scoullar (1980), and Shah (1980). The total population o f the valley is about 35,000, increasing at approximately 2% p.a. Family size is typically 5 - 6 . Population density and landholding per household average 288 inhabitants/km  2  and < 1 ha, respectively (Shah  1980), but are not evenly  distributed. Poulation density i n Dhikur Pokhari panchayat on the Kaski ridge is given as 364/km  2  i n Fleming (1978), and only 13% o f the farmers own 48% of the khet  (Scoullar 1980). About 19,000 large ruminants (cattle and buffalo) subsist on some 2000 ha of scrub and grazing land, but their productivity is low, with a pronounced negative trend i n production parameters (Shah  1980). Forage cultivation is non-existent,  largely due to shortcomings i n the extension system forage cultivation possibilities (Shah  farmers do not know about  1980). They are generally aware of the practice of  double-cropping khet with winter wheat, but enumerate constraints such as lack of irrigation water, uncontrolled grazing, and (most importantly), a shortage o f compost for the crop and consequently a possible adverse effect o n subsequent rice yields. Despite the grazing problem, farmers i n the valley sometimes cite insufficient livestock as being a primary constraint preventing increases i n agricultural production, as in Pumdi Bhumdi Panchayat, Phewa watershed (Shah  1980). This is related to the  55  Figure 14.  Simplified schematic model of a Nepalese Modified from: L . R . M . P . (1983).  hill farming  system.  56  gross inadequacy of fodder resources  (crop residues, weeds, grasses from terrace risers,  fodder trees, leaves and litter from the forest, and least importantly, communal grazing areas), graphically described by Shah (1980, p. 45):  livestock "are  in a semi-starved  condition for 6 months ( N o v e m b e r - A p r i l ) " . In the Phewa valley crop residues now provide 73% of total feed (Shah  1980), but this does not lessen the pressure on the  forests caused by fodder collection and grazing. The watershed is unusual in some respects,  with the town of Pokhara nearby  providing a market for commodities such as milk (some 500 tourism injecting cash to hoteliers and tea-shop  t/yr (Shah  1980)), and  owners i n villages along the trekking  routes. Another form of cash subsidy is the remittances  of mercenaries  recruited from  amongst the Gurung communities at the western end of the valley. Although  often  spent on "luxury" consumption (weddings, furniture, travel etc.), such capital does allow its beneficiaries to experiment with new farming techniques and commercial ventures, risks which the majority of the population cannot afford to take i n case they The effect of population pressure and socio-economic  fail.  differentiation can be seen i n  Figure 15, which shows a hillside near Tamagi at the western end o f the watershed. Here, a low-caste community is forced to live on a slope acknowledged to be liable to catastrophic  failure at some unknown date; there is nowhere else for them to go i f  they are to maintain a settled existence  with some guarantee  of family unity and  economic survival. A penetrating study of rural life in the area is available in MacFarlane (1976), and a sobering analysis of the interaction of economic and political forces in Nepal, and their consequences et al.  (1980).  for the individual, has been made by Blaikie  43  For further anthropological studies of communities i n central Nepal see, e.g., Jones and Jones (1976); Messerschmidt (1976); Hitchcock (1980); Coburn (1982). A description of historical and political trends i n Nepal can be found i n Rose and Scholz (1980). 43  57  Figure 15. Hillside near Tamagi, Phewa Valley. Note: (1) Reclamation of old landslide scar to bari. (2) H u m a n occupancy of failure-prone slope in background by low-caste community. (3) Degradation of forest to scrub on hilltop.  58  Physically, this agroecosystem and  nutrients. As elsewhere  is totally dependent on indigenous sources of  energy  in the country, chemical fertilizers have yet to become a  major source of inputs to the farming system in the watershed, and so gains  are  limited to nitrogen fixation, weathering, inputs i n runoff and irrigation water from upslope, and inputs in precipitation and through the trapping of wind-borne dust, aerosols  and gases on plant surfaces  (the  "filtering effect") . Losses include soil 44  erosion, leaching, volatilization from manure and compost, and export in cash The role of the forest is largely that of a nutrient capture facility, with "cascading"  crops.  nutrients  from the forests on the upper slopes through farm households and out  onto bari terraces as compost, from where runoff carries them down to the  khet  below. This link was implicitly recognized as long ago as 1823-24 by Bishop Reginald Heber, who travelled through northern India and observed: "Great devastations are generally made i n these woods, partly by the increase of population, building, and agriculture, partly by the wasteful habits of travellers, who cut down multitudes of young trees to make temporary huts, and for fuel, while the cattle and goats which browse on the mountains prevent a great part of the seedlings from rising. Unless some precautions are taken the inhabited parts of Kemaoon will soon be wretchedly bare of wood, and the country, already too arid, will not only lose its beauty, but its small space of fertility." (Heber 1849). T o date no quantitative study of this nutrient pathway has appeared in the literature on Nepal, although fertility management  is discussed in L . R . M . P . (1983).  Recent studies from middle and northern latitudes have indicated that precipitation and airborne nutrient capture can together supply forests with their total nutrient requirement over a rotation (see e.g. Tamm 1979, M i l l e r 1984 for discussion). It is interesting to speculate on the relative importance of these sources i n the sub-tropics where weathering is faster. A n introduction to biogeochemical cycling in tropical forest ecosystems can be found i n Anon. (1978). 4 4  Chapter 3 MASS  3.1  MOVEMENT  IN T H E PHEWA  VALLEY  INTRODUCTION Owing to the importance of Phewa Tal as a tourist attraction, and to the  relative accessibility of the northern part of the watershed at the start of a major Himalayan trade and recreational route , the Phewa valley has been the subject of 45  considerable attention by national and international development agencies. The principal objective  of this attention has been to reduce the rate of environmental degradation i n  the catchment  through integrated development, and thereby to protect Phewa Tal from  siltation and eutrophication. A s is common in many development projects government and aid technicians involved have tended to become welter of detail of project  enmeshed in the  design and implementation, and so have not been able to  devote much time to strategic studies of the natural processes modify. Chief amongst these processes of many of the agencies  the  which they are trying to  is erosion. Despite being a major stated  involved, rates of erosion and sedimentation in the  have not been systematically surveyed. Current knowledge of erosion processes  concern  catchment and rates  in the watershed is based on: (i)  the erosion plot studies of Mulder (1978) and Impat (1981) discussed in Chapter 2;  (ii)  suspended sediment sampling of the Harpon K h o l a near Pame Phewa Tal i n 1979 (Impat 1981), discussed i n Chapter 4;  (iii) bathymetric surveys of the lake carried out i n 1976 discussed i n Chapter 4;  and 1979,  above  again  (iv) an assumed correspondence between erosion rates and a simplified mass balance analysis of the phosphorus dynamics of the lake and watershed made by Fleming (1978, republished as Fleming 1983); see Chapter 4. Trans-Himalayan trade along the K a l i Gandaki valley, and trekking routes round the Annapurna massif.  45  59  60  The presence of the lake, which acts as a sediment trap, makes  the Phewa  valley one o f the few sites in Nepal where estimates of erosional activity on the hillslopes can be verified by comparison with sediment deposition downstream. Despite this advantage, no work has been  done on either mass wasting or fluvial activity in  the watershed. A s well as being highly visible sources  of sediment input to the valley  bottom fluvial system (Figure 8), these processes are also hazards well known to the inhabitants o f the area. In M a y 1983 a reconnaissance  survey of mass movement activity i n the  watershed was carried out, with the objective processes and their controlling mechanisms.  of identifying the dominant hillslope  Several of the larger failures had been  investigated informally during the previous year i n connection with erosion control activities i n the watershed.  3.2 M E T H O D S  3.2.1 A E R I A L  PHOTOGRAPHY  Aerial photography is important i n allowing the rapid assessment o f erosional landforms and, i f sequential cover is available, facilitates the investigation o f landscape change over time. Vertical air-photo coverage  of the Phewa Valley exists from:  (i)  January 1958: full cover, 1:40,000, flown for the Survey of India as a base for the 1:63,360 maps series;  (ii)  post-monsoon 1972: partial cover (Figure 15), approx. 1:15,000, probably flown by the A i r M a p Company o f Italy for Pokhara Town Planning Department;  (iii) M a r c h 1978: full cover, 1:50,000, flown for the Land Resource Mapping Project and available from the Topographical Survey Department, Kathmandu; (iv) November/December 1978: partial cover (Figure 16), approx. 1:25,000, flown by Huntings Surveys for Lumle Agriculture Centre; negatives now with Topographical Survey Branch, Kathmandu.  Figure 16. Phewa Valley: air-photo coverage i n 1972  and  1978.  62  In addition, Mulder (1978a, p. 9) mentions full coverage  of the watershed by  air-photos at 1:64,000. These may have been flown for the Department of Forestry, but, together  with the 1958  Indian photography and the March  1978  photography, were  not available for this study. Mass movement sites i n the watershed were identified by interpretation of the 1972  and November/December  1978  air-photos, and marked onto a 1:25,000  topographic base map ( I . W . M . 1980)  produced from the March 1978  photos. Failure  locations are shown in Figure 17. A number of sites outside the watershed were also identified, and were marked onto 1:50,000 scale topographic maps (photographic enlargements  3.2.2  of the Indian O n e - I n c h  FIELD DATA  In M a y 1983,  series).  COLLECTION  50 suspected mass movement sites identified by air-photo  interpretation in the Phewa watershed and surrounding area were visited on foot Morphometric and other characteristics and 4 outside, were determined (Table and morphometric measurements  of 22 of these slides i n the Phewa watershed, 6). Slope angles were determined by clinometer,  by h a n d - h e l d tape for smaller slides, by pacing for  larger ones, and, for extremely large failures, by measurements  from air-photos. Slide  age was estimated from a combination of scar freshness and degree of recovery through vegetation establishment, local interviews, and existence 1958  topographic maps and the 1972 Other characteristics  and post-monsoon 1978  recorded in the  field  and relative size on the air-photos.  were the nature of the failure surface,  the nature of the failed material (soil, untransported regolith, colluvium, bedrock), soil texture, structure, horizonation, and drainage, root penetration of the regolith, bedrock geology (including lithology, bedding, dip, jointing, faults, and competence),  surface and  subsurface moisture regimes, topography, connections to the drainage net, debris delivery  Figure 17. Phewa Valley: mass movement activity identified from 1972 photography and 1983 field survey.  and 1978  aerial  ON  Table 6. Pokhara  area: landslide  SLIDE IDENTIFICATION  T Y P E  APPROX. AGE  morphometry  siore or FAILURE SURFACE  I years1 (degrees 1  30  Rockfal1  13 33  D. D.  s i Ide, slide,  29 Pane 11  D. s i I d s ,  1 5 1 f 15 20 28 31 39 40 Ohophare Kask i k o t  Oebris  23  s1 i d s  • •  •  D.  avalanche  Shallow  Oaorati  Complex.  (ml  OVERALL LENGIH L  (ml  (ml  WIDIH OF SCAR MAX . MEAN W (ml  DEPIH OF SCAR MAX . MEAN D (ml  2  90  10  10  20  IO  IO  2  10- 12 1  40 24  33 8 60  N/A N/4  33 8 60  58 60  58 40  1 2 7  5-6 2  27 29  24 45  35 80  59 125  17 20  15 14  5- 10 1-2 10- 15 5 1 15-20 1 5-7 5-7 1 1  39 44 45 39 42 45 42 36 36 38 37  10 7 17 6 30 14 9.5 40 14 2 15 10 45 30  20 120 90 30 25 35 50 80 60 50 80  31 138 130 34 35 75 64 95 70 95 110  7.6 11 .3 19 20 8 15 11.3 10 10 19 30  7.6 8 15 12 7 9 7 8 8  ml  VOLUME  Isq.  ml  (cu.  ml  200  IOO  19GO 3600  1666 18000  5 3  2 0.75  360 630  BBS 1750  7 20 473  10.0  8.5 1 .6  1 46 1 .78  25  1 . 1 1 4 2 1 1 . 2 4 5 2 5 1.5 15 3 2  1 1 1 . 1 1 5 1 0.9 2 1 .8 1 1 2 1 3  81 141 450 168 67 360 99 120 80 360 750  130 1 104 1B00 408 245 675 448 760 560 760 2750  89 155 675 168 60 720 179 130 80 7 30 975  6.91 8.07 9.5 20.0 6.67 3 . 33 4.52 6.67 6.67 6.33 15.0  3.5 1 .0 1 7 2 9 3 4 6.0 3.9 1.6 3. 1 3.3 1.8  1 .87 6 82 3.00 3. 14 2.63 0.88 3.52 5.33 6.00 1. 11 2.67  1 1 .67  1.3  0.80  39  Including  >26  18-30  1330  10- 15 26-30 >26  5-35 34 >27  4 20 5O0 500  2  40  15  45  60  12  9  2  135  540  163  >50  70  100  170  35  30  12  8  2 100  5 100  16B0O  <26  35  150  330  75  40  40  25  60CO  36000  15OOO0  <26  23  320  630  75  36  13.5  1 1520  226BO  806 5 0  Phulwar 1  Debris  Lamachaur  Rock s i I d e  Jumtetl  Complex, f 1 ow  siide  5- 10 Including  See text for definition of terms.  100  235  35  30  3  1.5  3750  6750  5625  230  350  14.5  10  2.3  1.5  198  2500  297  400  1630  134  too  10  6  122000  132000  730000  4 20 1400 1 125  840 1900 1625  125 150 550  60 90 325  < 10 60 60  5 20 25  252O0 450O0 162500  4 20O0 36BO0O 400OOO  126000 90O00O 406OOOO  180 ( f a n ) 3 10  TENUITY lin/lc  100  53  WATERSHED  PROCESS D/L 1X1  I960 3600  2-4  19 8  INDICES SHAPE W/D  0 85 >5  e  1  1sq.  10IAL AREA AFFECTED  10- 15  1 low  125  (ml  AREA OF FAILURE SCAR  rotational  OUTSIDE PHEWA  Bl Jaipur  micaceous in talus  •  Dama f d a n d a  Pamdur Pane OkhaIdhunga  undercut undercut  lENC.m OF LENGIH OF SCAR DEBRIS Lm Lc  1 2  7  S O  10 0  1 .00  48.33 8.57  3 6 11.7  N/» N/«  3.4  0.9  1 1.62  13.4  0 6  0 33  12.5 2.5 9. 17  1.1 3.2 3.7  1.00 2.80 2.25  6.0  3 3  3.00  2 .92  7 . 1  1 .43  6.30  3.0 5.56  12 1 2 . 1  1 . 20 0.97  65  to channels, land use, elevation, aspect, association with other failures, and possible causal mechanisms and triggers. Angle of failure surface and approximate age were obtained for a further 11 slides (Table 7). The relatively homogeneous bedrock types i n the Phewa valley precluded any assessment of failure variation between lithologies, and insufficient data were collected to examine the effect on failure of slope age.  3.3 R E S U L T S A N D D I S C U S S I O N  3.3.1 D E S C R I P T I O N O F S L O P E M O V E M E N T Landslide  TYPES A N D PROCESSES  classification  Caine and M o o l (1982) describe the Nepali word for "landslide" (pahiro)  as an  all-embracing term for all forms of catastrophic slope failure including the fluvial gullies which often develop from failure scars. Since many different types of process are involved i n slope failure, and "improvements i n technical communication require a deliberate and sustained effort to increase the precision associated with the use of words" (Varnes 1978), it is essential to break down the overall concept of pahiro process-related  into  categories.  The classification of landslides and other forms of mass movement is a topic which has exercised specialists ever since Penck (1894) first distinguished between mass movement and mass transport. Mass movement describes movement under the influence of gravity alone, whilst mass transport allows material to be carried i n a moving medium such as water, air or ice (Fairbridge 1968). Although helpful as an analytical concept, i n reality these two categories merge as the content of the transporting medium i n the failed material increases. A t every level o f classification similar problems of "where to draw the line" abound, and the ideal system has yet to be  66  Table 7. Phewa watershed: age translational failures  Slide identification  and  angle of failure  Type  surface for  eleven  Approx. age  shallow  Slope  of failure surface (degrees)  (years)  17 37 38 45 46 48 49 50 51 52 55  Debris slide " " " " " " " " " "  >10 4-8 " " " "  "  " " " "  47 40 40 48 60 45 0 40 40 40 52  4  proposed. Currently, landslide classifications  are predominantiy descriptive tools based on  three fundamental processes of planar sliding (translation),  rotation, and flow, which  well described in a number of texts including, for example,  weak soil layer or boundary between determine  weathered  materials  and have a  is present within the slope (e.g. and unweathered  a  material), it can  the surface of failure and prevent the development o f a simple  slide; an element of translation is introduced. In general, the smaller the heterogeneity,  are  Skempton and Hutchinson  (1969). Characteristically, rotational slides occur i n homogeneous curved surface of failure. Where a heterogeneity  the  rotational depth to the  the greater the translational element will be. Compound slides reflect the  combination of a variety of curved and planar failure surfaces, generally at  moderate  depth, and usually involve severe distortion and shearing with corresponding  disturbance  in the slide mass (Skempton  and Hutchinson 1969). Flows usually have a high water  content, involve viscous deformation with inter-granular movements  predominating over  67  shear, and straddle the boundary between  mass movement and mass transport  (Hansen  1984). Recent descriptive classification schemes include those of Hutchinson (1968), Coates (1977), and Varnes (1978), and have been reviewed by Hansen (1984). Descriptive systems have proved extremely  useful i n providing simple grouping methods,  and some authors are now attempting to refine them through the use of morphological indices (see  Section However, it is likely that future classification work will  concentrate on the use of geotechnical  data rather than on morphology, owing to its  greater propensity for use in slope failure prediction. For this study Varnes' slope movement classification scheme was used (Figure 18).  First proposed in 1958  (Varnes  1958), and widely recognized and favoured in  South Asia by, e.g., the Central Road Research  Institute of India (Mehra and  Natarajan 1966). Varnes' system is partially quantified and clearly illustrated. However, it still requires some modification for Himalayan conditions, particularly with regard to complex failure events, Hillslope  flows, and high-angle fluvial transport  systems.  processes  As noted by Brunsden et al.  (1981) i n eastern  Nepal, air-photo  interpretation  revealed failure scars to be so widespread that mass wasting should be regarded as the norm i n this landscape rather than the exception. Following Varnes' classification five main categories of mass movement were identified: rockfalls, rockslides, debris slides, rotational slides, and deep and shallow flows. Shallow translational debris slides were by far the most numerous. Mass movement features ranged in scale from these small isolated slides through flows to large areas of complex movements movement catchments",  Brunsden et al.  ("mass  1981). Removal of displaced material often  involved both viscous deformation and fluvial transport i n steep channels which transitional between  slide debris tracks and gullies. The  were  material involved was usually  68  TYPE OF MATERIAL TYPE OF M O V E M E N T  BEDROCK  ENGINEERING SOILS Predominantly coarse  Predominantly  FALLS  Rock fal1  Debris  fall  \ Earth  fall  TOPPLES  Rock  topple  Debris  topple  | Earth  topple  Rock  slump  Debris  slump  Rock  block  Debris  block  ROTATIONAL SLIDES  Few Units  TRANSLATIONAL Many Uni t s LATERAL FLOWS  COMPLEX  SPREADS  siide  Earth  slump  . Earth  block  (  slide  Rock  siide  Debris  slide  | Earth  slide  Rock  spread  Debris  spread  | Earth  spread  Debris  flow  | Earth creep)  flow  Rock flow (deep creep) Combination  o f two  ( s o i1 o r more  principal  types  of  fine  slide  movement  Figure 18. Varne's abbreviated classification of slope movements. Source: Varnes (1978). Note: this classification should be used i n conjunction with the original reference, which specifies and illustrates the categories in detail.  Figure 19. Rockslide at Phedi, north of Naudanda. Debris travelled a further approx. 1 k m downslope in a ravine to the Yangadi K h o l a (not shown).  69  untransported, weathered bedrock, but in some cases colluvial material was entrained. The various categories are described i n greater detail below.  a) Rockfalls.  These were infrequent and occurred almost exclusively on cliffs and bluffs  developed i n the quartzose schist in the western and southern part of the watershed. Elsewhere the weaker grey phyllitic schist and phyllites were unable to support slopes sufficiently steep to allow rockfall, except along active fault lines. The cliffs and bluffs were generally associated with the middle and upper reaches of stream  channels,  themselves often exploiting lines of structural weakness.  b) Rockslides.  Rockslides, involving displacement of fractured rock along joints or  bedding planes, occurred preferentially in two locations: on the outside of  meanders  where streams undermine rock outcrops, and on the steep faces of fault-induced mass-movement  catchments.  Occasional examples were noted on the quartzose  schist  outcrops on the south side of the watershed. Mention should be made here of two large rockslides outside the watershed but still in the Pokhara area. One of these at Phedi, north of Naudanda, involved approximately 30,000 m  3  of material and was responsible for the loss of 7 lives  (Figure 19). A rock outcrop some 100 m high failed without warning towards the end of the monsoon i n 1982.  Part of the debris spread out on the shallower slopes below  the cliff, forming a fan, and part fluidised and travelled for approximately 1 k m as a torrent down a pre-existing stream channel to the Yangadi Khola, some 300m lower. The site of the failure lay close to a fault identified by Remy (1975). The  second  large rockslide was a fault-induced failure at Jumleti, north of Pokhara (Figure  4),  where rockfalls and rockslides from the steep sides of the failure provided a continuous supply of debris, which was then transported by fluvial action to a fan  70  covering terraces on the low-angle alluvial slopes below. This site was active  during  the dry season, with small size material ravelling from the scarp, and seeps lubricating small flows i n the centre of the  c) Debris  slides. The most common forms of mass movement  were shallow translational  failures or  divided into three categories: (1)  and  river  common category: to allow the  (e.g.  29, Pame  banks due to undercutting (13,  watershed  These slides could be  i n unusually weak or  11, in Table 33, in Table  6); 6);  (2)  failures on  and (3),  the  most  failures on undisturbed regolith with sufficient runout to a channel  formation of flows in the displaced material.  Slides i n the first category movement  i n the Phewa  "debris slides" (Figure 20).  failures on slopes of < 3 6 °  disturbed material, often micaceous stream  scar.  were often associated  due to the general weakness  with other  forms of mass  of slope materials i n the vicinity. Slides due  to undercutting had a much larger width-length ratio than those i n n o n - fluvial locations. The where  factors such as jointing and local convergence  contributed 1,000  preferred location of the slides on undisturbed regolith was interfluves,  m  3  of subsurface  flow  nets  to instability. Characteristically these slides were small, involved less than o f material (mean  of 36°-45°,  volume was 331  m ) , and occurred i n regolith on 3  slopes  although one failure was noted at 60° (Tables 6, 7). Site features indicated  that at failure, a shallow (1-3  m deep) mass of soil and weathered  downslope with the material invariably disintegrating, and sometimes  bedrock  moved  becoming fluid and  travelling as a debris flow to cover the undisturbed surface below the failure site. Failure surfaces were usually the bedrock/regolith  interface.  N o seepage was visible at  any of these shallow sites at the time of the survey, in M a y (pre-monsoon).  A t some  locations roots of up to 2 cm diameter protruded from undisturbed material above primary scarp, and appeared to have broken i n tension. Root penetration  of bedrock  the at  71  Figure 21.  Slow debris flow at Pamdur, Phewa Valley. Rubble to front and side of debris lobe has been washed clean of fines by rain and fluvial action.  72  the failed sites was not observed. Although occurring i n both forested and  non-forested  locations, these shallow slides were particularly common throughout the watershed i n areas of scrub.  d) Slow debris  flow/earthflow.  In some locations in the area debris generated  failures accumulated and moved downslope as an elongate  by slope  mass of disintegrated  debris  of all sizes (Figure 21). Movement was by sliding over steep lateral and basal shear surfaces  with viscous deformation in the centre of the mass. Velocities were  measured, but comments  by local farmers indicated movement in Varne's  "rapid" categories (between  1.5  m / y r and 0.3  not  "slow" to  m / m i n ; Varnes 1978).  In the Phewa valley these flows occurred i n four locations (Deorali, Pamdur, Pame, Okhaldhunga) associated  with either very weak micaceous schist (Pamdur and  Deorali), or deeply weathered phyllitic schist (Pame, Okhaldhunga). Morphometric details are available i n Table 6. This category of Brunsden et al.  (1981), and the  large flow at Deorali (circa  7.3  of mass movement equates to the "mudslides"  "flowslides" of Caine and M o o l (1982). The  x 10  5  very  m ) also bears comparison to the earthflows in 3  the California Coast Range described by Kelsey (1978), with a hummocky surface, undrained depressions, and well-developed lateral margins defined by shear planes. These movements  are here called slow debris flows to continue the utilisation of  Varne's classification, although such a category  is implied rather than described i n his  system. In the Pokhara area the coarse nature of the material involved generally excluded such slides from the otherwise appropriate "earthflow" and "mudflow" categories used by Varnes, which require more than 50%  fines  (Varnes  1978,  Evidence for the location of these flows in groundwater discharge zones  p.  18).  was  considerable. A l l the flow sites were observed to be moist at the time of the survey, in M a y , with evidence of recent movement such as torn roots and dying  vegetation  73  (Figure 21). A t two sites seepage was sufficiently reliable for the local inhabitants to regard the springs as sources of dry season drinking water. A t Pamdur retrogressive slumping had eaten back to a rockshelf from the base of which a steady flow of water emerged to lubricate the head of the failure. A l l the flows were noted to be active during the monsoon season of 1982,  with activity decreasing during  the  subsequent dry season as, presumably, pore-water pressures declined. In some cases, during the monsoon rain generated sufficient runoff on these flows to allow the formation of fluvial channels cutting across the surface. Evidence such as scoured stretches of debris track suggested that, on some slopes, the flows may have  become  thixotropic on failure and fluidised, resulting in rapid flow, a phenomenon similar to the debris torrents of Japan and the Pacific Coast of North  America.  Although few in number the scale of these flows and their transport efficiency meant that they were responsible for a high proportion of overall sediment movement and input to valley bottom river systems. The four slow debris flows were all associated with  e) Rotational  "mass movement catchments"  and are discussed further below (f).  slides. Rotational slides were not common features in the study area  owing to the generally low cohesiveness of regolith materials and the intensity of slope processes, resulting i n shallow regoliths. They were confined to two locations: the upper end of mass movement catchments  where retrogressive slumping supplied material to  the debris transport system (Figure 22), and much larger, deeper slumps i n the regolith which occurs i n some re-entrants  deep  on the southward-facing slopes of the  Kaski  ridge. Here failure was due to erosion nickpoints moving slowly upslope i n "slope failure ampitheatres"  (Brunsden et al.  1981), thereby removing support from material  above. Both types of rotational slide rapidly developed a more translational form as the blocks disintegrated and slid on the planar surface of the underlying  74  Figure 22. Retrogressive slumping supplying mass movement catchment, Bijaipur, near Pokhara. Fault runs up centre of gully. Stepped appearance of main scarp is caused by terracing (ban).  75  outward-dipping phyllites.  f) Mass movement  catchments.  In four locations i n the watershed (see (d) above)  debris flows were fed by complex failures i n a rapidly eroding catchment  above  (Figures 8, 23). Failure types included rockslides, debris topples, debris slides, debris and earth slumps, and dry ravel. A t a fifth location near Pamdur, debris from a mass movement catchment  moved directly into a stream channel. Debris from the Pame and  Okhaldhunga failures had formed large fans spreading out onto the valley floor, with concave profiles gradually increasing from 1° on the valley floor to about 14° at the base of the failures. These fans had encroached  on the channel of the Harpon K h o l a .  During the monsoon intense activity was noted i n these catchments, expansion at the head by retrogressive  with rapid  slumping. Activity continued at a lower level  for the remainder of the year, with talus slides and ravel supplying coarse debris from the steeply sloping failure scarps. A t all these sites seepage was observed during the dry season, A t this time of year, waves of coarse sediment moved by flow and viscous deformation i n lobes some  1-3 m deep along the bed of the failures, but  these were largely washed away by runoff processes generated  within the catchment  during the monsoon. Distinction should be made between these discrete, extremely larger erosion ampitheatres ampitheatres,  actives sites and the  on the north side of the watershed (Figure 8). These  were some 800 m i n length, and although dissected by vigorous  ephemeral streams and gullies, were sufficiently stable to allow cultivation o f the majority of their area. The streams and gullies provided rapid removal of any material reaching their channels, depositing it i n fans where the streams debouch on the valley floor (see profile of ephemeral stream i n Figure 33).  76  Figure 23. Mass movement catchment at Pame, Phewa Valley. Fault runs up centre of failure. Note: (1) Fluvial action in bed of catchment. (2) Fresh failures at top left (3) Displacement at top indicating deep movement  77  g) Progressive ampifheatres  creep. This phenomenon  was detected i n a number of the erosion  on the north side o f the valley. T h e whole mantle was moving downslope  at a rate o f 1-5 m / y r by means of both shallow creep and deeper deformation, creating an undulating profile with occasional  viscous  tension cracks. The process  has been described i n the K a t h m a n d u - K a k a n i area o f Nepal by Kienholz et al. (1983), who termed it "slow mass movement", creep"  of Ter-Stepanian  and is synonymous with the "depth  (1965). Geomorphological interpretation  local oral history suggest that eventually catastrophic  failure o f these moving masses  occurs, and very large masses o f material are transported triggered by extreme precipitation events and/or  h)  Gullies.  ephemeral  to the valley floor, possibly  seismicity.  T h e slopes on the south side o f the Kaski ridge were dissected by streams, some o f which had eroded down to bedrock and were relatively  stable, but others o f which were extremely rills,  of the landscape and  active. Material was supplied to them by  dry ravel, shallow planar slides, topples from the banks, and slumping at the  headcut. These features were distinguished from the slow debris flows by having an obvious fluvial form, and a bedload consisting o f sorted material rather than the varied particle sizes o f the flows.  The various hillslope processes identified are illustrated i n Figure 24, a composite landscape  3.3.2  based  on the north side o f the Phewa  watershed.  MASS WASTING RATES I N THE PHEWA Landslide  density and area  VALLEY  affected  Overall landslide density i n the Phewa watershed was 1.6 slides/km . 9 5 ^ o f 2  these sites were small shallow failures. The figure includes all recent active mass  Figure 24.  Schematic process.  diagram of part of the  Kaski ridge, showing four types of  slope  79  movements  identified on the 1972 and 1978 air photos and by the 1983 field survey  (190 separate sites in all). The aerial photography covered only 74% (1972) and 90% (1978) of the watershed (Figure 15), and therefore regarded as conservative.'  the density figure should be  16  Excluding "slow earth flow" the total area of active and unhealed failures was 0.7  k m , or 0.5% of the basin area. The total area directly affected 2  by landslides  includes the area of debris as well as that of the failure itself and so was considerably larger, at 3.25 k m , or 2.7% of the watershed. O f this, a very large 2  proportion (90%) was due to debris fans resulting from activity i n mass movement catchments. Landslide  growth and  recovery  A s Caine and M o o l (1982) noted for their study area northwest of Kathmandu, true sliding failures were much smaller and shallower than those involving slow flow and gully development  Based on a plot of age against area, they hypothesized a  linear expansion of their slides at about 60 m / y r from an initial failure area of, 2  typically, about 600 m . However, this model cannot be supported i n the Pokhara area. 2  Although a regression of area o f failure scar on estimated age for all sites gave a positive relationship significant at the 99% level: Area (m ) = 2  2599 Age (yrs) -  10176  (r=0.626,  n=26)  many o f the shallower debris slides were observed to be healing, rather than expanding with time. This was confirmed by a plot of shallow debris slide area against age (Figure 25), which failed to show any obvious trend of size with time  In addition no allowance was made for slides not visible on the photos due to e.g. forest cover; equally no adjustment was made for false positive identification of slide sites on the photos beyond those excluded during fieldwork. O f the 46 sites visited i n the watershed, 12 (26%) were false identifications due to reflectance o f light tones from barren areas, and rocks. However, at the same time, four previously unidentified failures were located. 46  80  (r=0.150). The tendency of these shallow failures to cluster below an age of  eight  years suggests that, typically, they may heal and so no longer be identifiable within a decade. In the K a t h m a n d u - K a k a n i area Caine and M o o l (1982) also found their slides to be of very recent age, less than five years. These figures may be with the observations  of Zollinger (1979b), who reported revegetation  debris  compared  times for new  landslides on "very steep slopes" i n the Mahabharat Lekh and Midlands to be, sometimes, as little as 1-5  years;  Dutt, who found that slides in Siwalik formation  near Darjeeling caused by the 1934  earthquake  were revegetated  by 1955  (although  slides in phyllite continued to grow) (Dutt 1966), and Simonett, who estimated approximate revegetation  time of 30 years  for "large isolated debris avalanche New  Guinea (Simonett In contrast  for "non-contiguous landslides", and 40  years  or composite gully slides", all on granitic terrain in  1967).  to the shallow failures, the future of the mass movement  in the Pokhara area (which are largely responsible age and area in the regression)  for the high correlation  catchments between  may well involve further growth. Brunsden et  (1981) have described a mechanism whereby mass movement catchments  ground becomes so extensive  in relation to total catchment  dissipated by remaining vegetation, runoff will increase, erosion; an "autocatalytic"  al.  could be  maintained i n an active condition by runoff processes: where the area of bare  and  an  unstable  area that runoff cannot  be  resulting in further instability  condition exists. The authors suggest that this condition  can be maintained because the recovery time of the vegetation than the reaction time of the system to monsoon events.  and soils is longer  Stability will not return until  the failure is sufficiently large to permit the development of slopes at a stable angle. The major failures in the Pokhara area may well be i n an autocatalytic condition, and the time required for them to stabilize is not known. Phewa watershed.  Figure 25. Plot of area of failure scar on estimated age for 12 debris slides in the Pokhara area. Data from Table 6.  82 Volume  and  area  The" survey results revealed wide variability i n mass movement types i n the Pokhara area. Slide volume varied from 60 m figure represents  3  to over 4 x 10  m \ although this last  6  material moved by a succession of incidents in a single mass  movement catchment,  rather than a single failure event  In the Phewa watershed the  largest volume of material currently moving as a unit is probably the Deorali slow debris flow, circa 7.3  x 10  s  m , although progressive creep in the erosion 3  on the north side of the valley may involve up to 5 x 10 Failure  frequency  and  surface lowering  Estimates of the average rate of slope retreat  by  6  ampitheatres  m . 3  landsliding  by landsliding can be obtained by  combining the volumes of debris for individual mass movements with their spatial distribution (landslide density), and then finding some way of estimating their frequency of occurrence  (Saunders and Young 1983). For the Phewa watershed mean volumes of  material displaced by each type of slide were calculated from the morphometric measurements  (Table  6), and their spatial distribution determined for the year 1978,  date chosen owing to the high proportion of air photo coverage year (90%,  Figure  a  of the catchment  that  16).  Frequency of occurrence can be estimated i n two ways: firstly, from the  average  age of slides, and secondly, from sequential counts of actual slides on the ground. In this study the first method gave values of failure age of 24 years for mass catchments,  and 5.5 years for all other slides (in 1983)  ages by the landslide density figure of 1.6 slides/km of about 0.3  2  (Table  movement  6). Multiplying these  (section gave a frequency  slides/km /yr. These figures are very approximate owing to lack of 2  precise dating, and a small and unrepresentative  sample.  The second method of frequency estimation should be more reliable: excluding the mass movement catchments,  81 slope failure sites were identified on the 1978  air  83  photos i n the area o f overlap with the 1972 photos. O f these, 25 sites were common to both sets o f photographs; the six years between the catchment  ergo, in this area, 56 new mass movements occurred i n  1972 and 1978. The area o f overlap is 84.88 k m , or 70% o f 2  This gives a slide frequency o f 0.11 slides/km /yr, or, pro rata, 13.3 2  slides/yT for the whole  catchment  47  These failures were largely shallow translational slides, with a high proportion caused by undercutting. By adding an average volume for these slides (5000 m , an 3  approximate weighted mean volume o f debris slides i n both undercut and open slope situations), to that o f the annual output o f the mass movement catchments divided by age), it was possible to make a crude estimate volume of material mobilized by mass movements  48  (volume  o f the average annual  i n the catchment:  3.1 x 10  5  mVyr.  This is equivalent to a rate of surface lowering by landsliding of about 2.5 m m / y r . To determine a denudation rate from this estimate  it is necessary  to add the  contributions from other erosion processes (soil creep, sheetwash, rilling and gullying, solution), and to adjust the total by a sediment delivery ratio to account redeposited within the catchment  Values for surface  erosion can be estimated from  Impat's (1981) runoff data: assuming a bulk density o f 1.6 g/cm material, Impat's soil loss data from dense pasture i n 1979 (Table  forest  for material  3  for undisturbed  protected pasture, and overgrazed  2) convert to approximately 0.7 m m / y r , 1.6 m m / y r , and 5.6  m m / y r , respectively. These figures could, cautiously, be extrapolated to the remainder o f the watershed, except  that nearly half o f the area is under cultivation (see 2.7).  Erosion rates on the bari and khet have not been measured, and neither have the  Both frequency estimates suffer from the methodological problems common to all frequency assessments such as the assumptions o f climatic and slope strength constancy (see Crozier (1984) for discussion). Calculation: (1) 13.3 slides/yr x 5000 m = 66,500 m ; (2) 5.8 x 10 m (total volume o f complex failures i n Phewa watershed from Table 6) divided by age (24 years) = 2.4 x 10 m ; (3) Sum = 3.1 x 10 m ; (4) Sum divided by area o f watershed (122 k m ) = surface lowering (2.5 mm/yr). 47  48  3  s  2  3  3  5  3  6  3  84  contributions o f creep, gullying, and solution. In addition, the sediment delivery ratio remains unknown (see 4.3.3), and the overall uncertainty makes it unwise to propose a denudation rate for the catchment  3.3.3 M O R P H O M E T R I C Volume/area  at the present  time.  ANALYSIS  relationships  Empirical relationships between  landslide volume and area have proved useful in  mass movement and sediment studies through allowing the extrapolation of limited field data over a wider area by air-photo interpretation. For example, i n southern California Rice  et al. (1969) determined a power relationship between volume and area of: volume  =  0.234  area '  1 11  and i n N e w Guinea Simonett (1967) established a very high correlation between  slide  volume and slide area: log volume (ft ) = 3  1.368 log surface area (ft ) 2  0.6885  (r=0.98)  based on a sample of 201 failures containing a "wide variety of landslide types". In the Pokhara area linear regression of volume on area for all failures gave a relationship of: volume (m ) = 3  18.4 area of failure (m ) 2  34669  (r=0.888, n = 26)  significant at the 99% level. However, as with the area/age regression, this curve is strongly affected  by the very large mass movement catchments,  giving the unrealistic value for area of nearly 2000 m  2  with the intercept  at zero volume. It is more  appropriate to investigate the relationship within process groups, and analysis of the large failures; on their own gave a regression of: volume The regression  =  19.7 area of failure -  194270  (r=0.841,  n=6)  (r=0.993,  n=16)  for all shallow failures (Figure 26) was:  volume  =  1.5 area of failure -  23  85  1 0  1 100  1  1  1  200 300 400 500 600 700 800 AREA OF FAILURE SCAR (m2)  1  1  1  900  1000  Figure 26. Regression of estimated Data from Table 6.  1  1  1  r  1100  volume on area for shallow slope  failures.  86  significant at the 99% level. Volume and area are not, of course, independent but the correlation coefficients  variables,  do indicate some linearity in their relationship. In the  case of the Pokhara area debris slides the high correlation coefficient suggests that similar processes and material properties may be involved throughout this category  of  failures. It should be emphasized that these regression are based on very small samples and should not be taken out of Shape  context  indices  Although Blong (1973) found multivariate analysis of morphometric attributes be unsuccessful i n distinguishing between slide, avalanche workers (Skempton 1982)  1953,  have suggested  M c L e a n and Davidson 1968,  that morphometric measurements  and  flow-type  Crozier 1973,  to  failures, other  Caine and M o o l  of landslide scars can help to  illuminate conditions on the slope at the time of failure. The hypothesis  under  consideration when using morphometric analysis is that the morphology of a particular slide is closely related to its dominant genetic processes (Crozier 1973). For any study verification o f the process/morphometric a subjective  index relationship can only be based on  assessment of the the dominant process  event, and the usefulness  one  as diagnosed i n the  field  after the  of morphometric analysis as a diagnostic tool should not be  overemphasized. However, the use of quantitative shape indices does  facilitate  comparisons with other results in the literature, and a brief exercise in morphometric analysis is carried out here. Relevant terms i n landslide morphology are illustrated i n Figure 27 . 49  Three shape  indices were applied to the Pokhara landslide data: a  shape index, a process index (Skempton  cross-sectional  1953), and the tenuity index (Crozier 1973).  Detailed morphometric definitions for landslides are available i n Brunsden (1973) and Hansen (1984).  4 9  87  Figure 27. Terms  in landslide morphology.  88  The  cross-sectional  shape index (width o f failure scar over maximum depth o f  failure scar, W / D ) was highly variable. Values ranged from 3.33 to 20.0 for slides o f less than 1,000 m  3  (mean=8.03, st, dev.=4.15, n=18), with a modal value o f 6 to 7,  to 2.5 to 48.33 for larger failures (Table form is not constant  6). These ratios indicate that  and is controlled by site-specific  cross-sectional  factors rather than being purely  a function o f process. The process index is so-called because, i n his study of landslips i n boulder clay in north-east between  England Skempton found this index useful as a means o f distinguishing  "surface  slips", "deep rotational slips", and "slumps" (Skempton  1953). A s  defined by Crozier the process index, or D / L value, is "the ratio o f the maximum depth o f the displaced mass, prior to its displacement (true depth), to the maximum length measured up the slope, expressed as a percentage." (Crozier 1973). The  "maximum length measured up the slope", has also been described by  Skempton and Hutchinson (1969) as the "maximum initial downslope extent", and is taken to mean the distance from the primary scarp to the tip o f the displaced material (see F i g . 27). Since  Skempton's original paper i n 1953, several workers have found D / L ratios  to be extremely useful as an indicator o f failure mechanisms  and consequently i n  classification studies (Davidson 1965, Selby  1967, Davidson and M c L e a n 1968, Crozier  1973,  value o f the process  Caine and M o o l  1982). T h e average  index for the 12  debris slides surveyed i n the Pokhara area, expressed as a percentage,  was 2.87. This  value is very low compared to figures for planar slides reported by other (Table  workers  8), even though the debris from many failures debouched into stream  and longer run-outs (and therefore  channels,  even lower D / L values) would have resulted on  longer slopes. However, it is above values for flows, which range from 0.83 to 2.40 (Table  8). T h e workers quoted i n the table were working i n areas o f deeper regolith  89  Table 8. Average flows  values for  D/L  the  D/L  ratio or Process Index for planar slides and  Ratio  Author  (%) Planar slides  Flows  8.00 7.66  2.40  Skempton 1953, West Durham, U . K . Crozier 1973, Eastern Otago, nr. Dunedin, N.Z. Caine and M o o l 1982, Kathmandu- Kakani, Nepal Davidson 1965, nr. Gisborne, E. Coast N . Island, N . Z . Extracted from the literature by Davidson, 1965 Selby 1967, Waikato, N . Z . This study  6.63 6.00  1.50  5.00  0.83  5.00 2.87  and generally less intense rainfall, and their higher D / L values indicate more coherence in the failed materials. The value of D / L for the Phewa debris slides is consistent with failure with a high water content  and consequent  viscous deformation or flow o f  the displaced material to produce long downslope extensions appears reasonable  of debris. It  therefore  to support Crozier's supposition that areas with shallower regoliths  and higher and more intense rainfall will have slides with lower D / L values (Crozier 1973). A plot o f D / L for both open slope debris slides and mass movement  catchments  against age (Figure 28) was not significant (r=0.231, n=20). This suggests that  either  slides undergo no expansion whatsoever with time (i.e. they heal), or that any enlargement  is allometric , i.e. extension increases i n proportion to depth. This may be 50  an artefact of the regression, since it appears unlikely that shallow translational failures  50  See  Allometry: when two processes change e.g. Fairbridge (1968); Bull (1975).  rate but retain the same ratio to each  other.  90  • O — —  translational failure mass movement complex including flow limiting envelope ? y = 0.06x • 2.71 r = 0.231  131  o  12  Jumleti  11109 29  8 7  5  6  28  o  2  t  6  8  10  12 U 16 18 AGE (years)  20  22  24  26  28  30  Figure 28. Regression of process index on age for debris slides and mass movement catchments. Data from Table 6.  a  91  will deepen i f their original slip surface  was bedrock. The plot also suggests a limiting  envelope for D / L values of between  and 4.0  with D / L values > 4 streambanks  0.5  for the majority of failures. Events  were generally either rockslides and rockfalls, or planar slides on  with much reduced or non-existent  runout paths.  Crozier's tenuity index, the length of the displaced material over the failure scar (Lm/Lc:  Figure 27), reflects  the tenuity of the displaced mass in relation to its  original size, and so gives an indication of fluidity. However it is not a true of fluidity because it does not take into account  measure  lateral spreading or the effect of  slope inclination (Crozier 1973). The mean tenuity index value for all the debris slides in the Pokhara area (mean=3.25, SL dev. = 1.88, value for  n = 12), was similar to the tenuity  "fluid flow" failures found by Crozier (1973) in N e w Zealand (3.33). The  tenuity value for the mass movement catchments  (mean=1.43, st  dev.=0.92, n=6)  was  similar to Crozier's figure for "viscous flow" (1.71). As with the process index, these figures may indicate a high water content and/or low cohesion in displaced material immediately after failure. In Nepal, Caine and M o o l (1982) obtained a similar range  of  tenuity index values for failures in their study area near Kathmandu. However, in contrast  to the  when regressed  Kathmandu area, L m / L c  for the Phewa slides showed no reliable trend  against slope angle (r=0.252). If genuine, this may indicate the  importance of site specific factors other than slope angle in determining degree of runout O f greater interest is the relationship between  D / L and L m / L c (Figure  Although weak (r=0.313), the regression line of tenuity on process  29).  for all failures in  the Pokhara area (except for rockslides and falls and undercut debris slides) is approximately parallel to the regressions slides and flows i n the  determined by Caine and M o o l (1982) for 15  K o l p u K h o l a watershed near Kathmandu, and by Crozier  (1973) for flows i n N e w Zealand (Figure 29).  The plot for the Pokhara debris flows  12  •  •  Debris slide  •  Other failure  Figure 29. Regression on tenuity index on process index for all slope failures in the Pokhara area, except for rockfalls, rockslides, and undercut debris slides.  93  alone was steeper (y=-0.98x+6.07, r=0.706, n=12), and significant at the 95% level. The implication of these inverse relationships between tenuity and process shallower failures are more tenuous, i.e. fluid, than deeper ones.  is that  A s with the  process  index results, this may indicate a higher water content i n the shallower slides.  3.3.4  M A T E R I A L PROPERTIES A N D SLOPE STABILITY  Mass movement occurs when the stress applied to material on a slope equals or exceeds the strength of the material along a failure surface, with consequent  movement  by fracture or plastic deformation. The shear strength, s, of a soil was first defined, empirically, by Coulomb in 1776, s =  in terms of cohesion and intergranular friction: c  +  a  tan0  (1)  where s is shear strength mobilized at failure, c is cohesion, a to the shear plane, and 0  is the angle of internal friction  51  is the stress normal  (Coulomb 1776). s varies  with a  due to the effect of increasing normal stress which increases the friction  between  moving particles.  In engineering terms soils can be divided into two broad classes with regard to slope stability: coarse grained cohesionless  sands and gravels i n which, by definition,  the angle of repose when dry is equal to the angle of internal friction ( 0 = 0 ) , and i n which failure surfaces are essentially planar; and soils which possess cohesion i n which failure surfaces are generally curved i f the material is homogeneous  and  (Graham  1984). Failures in the Pokhara area were usually translational. Because of this it is possible to use the concept of infinite slope analysis as an approach to a discussion of slope stability". Natural slopes will rarely display the ideal conditions assumed by  the  infinite slope model, but it is a useful approximation when, as i n the Pokhara area,  5 1 52  Also termed the angle of shearing resistance. For a recent review of slope stability analysis see Graham (1984).  94  the thickness of the unstable mantle is small compared to the length of the slope in question (O'Loughlin 1972). The assumptions of an "infinite slope"  and isotropic soil  53  conditions permit the analysis of a typical slice of material above a hypothetical failure surface (Figure  30).  F r o m considerations of continuity the forces on both sides of the slice must be equal, opposite, and colinear. For a slice of unit width and vertical thickness z (Figure 30), the vertical force across the base of the slice must equal its weight W , and can be resolved into normal (N), and tangential (W), W where 7  =  N  7z;  =  components:  z cos0;  7  T  7  z sine?  is density. Since the length of the slide surface is l / c o s 0 ,  shear stresses (0 and T ) a  =  7  z cos 0;  T  2  =  7  z sin0  the normal and  sliding (F) can be defined by the ratio S/T, c  =  +  (7 7  cos0  (3)  must not exceed the mobilized shear  strength, s, of the material. Thus, for a cohesionless  F  (2)  are:  For stability, the downslope shear stress T  where c  =  mass, the factor of safety  against  or:  z cos f? 2  z sin0  u) tan<6'  (4)  cost?  and <j>' are the effective cohesion and effective angle of internal friction . 54  W i t h the free water surface at a height of m z above the failure surface  (Figure  30), and seepage parallel to the slope, the pore water pressure, u, is equal to 7 cos 0, 2  where 7  W  =  c'  +  (7  -  7  W  w  7  m) z cos 0 tanri'  z sin0  2  cos0  "... a constant slope of unlimited extent which has soil properties at any given distance below the surface "Effective" strength parameters are those determined than for the undrained state used i n short-term (total 54  mz  is the density of water. Substitution into (4) gives: F  53  W  (5)  constant conditions and of the slope." (Taylor for the drained state, stress) analysis, where  constant 1948). rather s = c.  Figure 30.  Diagram to illustrate the infinite slope approach to shallow slide stability analysis. See text for definition of terms and symbols.  96  The types of material involved in the Pokhara failures varied from untransported weathered bedrock to colluvial debris and talus to surface soils. The taluvial materials" included deposits of very recent origin, as within mass movement  catchments,  and much older deposits now being reworked. Generally these slope materials consisted of non-plastic sands and silts with a variable proportion of gravel, boulders and bedrock corestones. non-existent  It is reasonable  to assume that these materials had a low or  cohesion, and that any strength was imparted by frictional properties. The  assumption of non-cohesiveness  is supported by evidence that most natural materials in  slopes do not exhibit tensile strength properties at the. macroscopic level (Kenney 1975), and that over time the cohesive component of strength in materials near the ground surface  is lost due to unloading and weathering (Carson 1976). It has also been  validated empirically through successful application of the infinite slope approach to the analysis of slope failure problems. Under the assumption that c = 0 , or threshold  slope angle (Carson 1975)  the angle at which stability can be maintained becomes a function of the angle of internal  friction of the material comprising the failure zone (<£')> the maximum pore-water pressure (u) on the failure plane, and the unit weight of the mass (Carson 1976). Brunsden et al.  (1981) suggested that typical values of 0'  for taluvial materials  might be between 33° and 45°, based on a review of measurements In Nepal, Caine and M o o l (1982) found values of <pp  i6  i n the literature.  of between 20.4° and 47.0° in  untransported, in situ debris developed on phyllites in the K a t h m a n d u - Kakani  area.  Back calculation using the infinite slope model gave threshold slope angles for this material of 32.8° when drained, and 18.5° when saturated. In the Pokhara area mean failure surface angle of shallow translational failures was 38.5° (st  the  dev. = 5°,  n=15), which is very similar to the mean slope angle of undercut slopes in phyllite " 56  M i x e d talus and colluvium. Peak friction coefficient, equals angle of internal friction.  97  in eastern Nepal (39.2°) determined by Brunsden et Brunsden et al.  al.  (1982) from profile data.  found that their debris slides generally occurred in the range of 35°  to 43°, which prompted them to suggest they may be controlled by full frictional strength rather than being dependent on saturated regoliths and high pore water pressure conditions. Sufficient data to carry out a realistic assessment of slope material properties i n the Phewa watershed are not available. In other forested environments sensitivity analysis of the infinite slope model indicates that, whilst the factor of safety may be relatively insensitive to changes i n <j>' and density, it responds rapidly to changes i n 8,  z and c, and also to piezometric  pressure at high groundwater elevations (O'Loughlin 1972;  W u et al.  1979;  Gray and  Megahan 1981). In many of the cases analysed the contribution of roots was critical to stability, increasing cohesion both through root tensile strength and by soil arching or buttressing between trunks and root systems. For example, i n Alaska, W u et al.  (1979)  found a shear strength value provided by roots of only 5.9 kPa to be critical in preventing failure, and O'Loughlin (1984a) found values of artificial cohesion due to roots of 1.0 kPa to as much as 20 kPa reported i n the literature. '..  Most of these studies were carried out i n areas of coarse-textured  soils. In the  Pokhara area the underlying phyllitic rocks weather rapidly to clays, and this might supply some cohesion. Alternatively, shear strength could be increased by the frictional component supplied by interactions between coarse particles i n the debris and roughnesses  i n the underlying bedrock. The argument for cohesion can be supported by  the morphometric analysis which indicated saturation and high pore water pressures on failure, and so high resistances to movement i n pre-failure material. Two direct shear tests carried out by Caine and M o o l (1982) i n the Kathmandu area indicated an almost total loss of cohesive strength i n remoulded materials developed on granite and augen gneiss, and penetrometer  tests in the same area indicated that debris derived  98  from phyllitic rock was appreciably weaker than that from other lithologies. The significance of this loss of strength on remoulding is that  the  "entire Theological character of a mass may change from quasi-plastic viscous i f amounts of pore fluid are large enough" (Carson 1976)  to  with a resulting extension of downslope travel.  3.3.5  LANDSLIDE Geological:  INCEPTION undercutting,  faulting,  over steepening  The survey of mass movement characteristics  in the Pokhara area suggests three  categories of slope failure with fundamentally different causes. These are: (1)  failure due to the removal of basal support by undercutting, usually by fluvial action but occasionally by construction activities;  (2)  failure on or near structural discontinuities;  (3)  failure by shallow sliding and rotation on non-undercut slopes, or by deeper creep and eventual failure of large masses of material i n areas of deeper regolith;  The  first  category  of failures is important as a direct source of sediment  fluvial transport. In the Phewa watershed the second category, wasting, was responsible  for  fault-induced mass  for a very large proportion of all material displaced  (see  section 3.3.2). The prevalence of large, active failures close to structural discontinuities was probably due to movement along active faults, and to the presence of rock which allowed the penetration of water and consequently  shattered  deep weathering and loss  of cohesion. D i p and bedding also contributed to instability, and it is interesting  to  note that according to Karunakaran (1975), the joints along which stress release failure takes place in many Himalayan river valleys are completely independent of structural and lithological boundaries. These "valley joints" cracks"  (Carson 1971)  result of unloading,  (Karunakaran 1975)  or "deep  tension  generally run parallel to the surface of the slope. They are a  the reduction i n pressure caused by the removal of overlying  99  material by denudation, and the consequent 1976). Strength  volume expansion or deconsolidation  (Carson  reduction by deconsolidation affects both cohesion, by promoting  fractures, and friction, by reducing interlocking between  the resulting blocks.  In the area around Pokhara, it was possible to confirm the presence o f faults by drawing lines between mass movement clusters marked on 1:50,000 topographic maps and then visiting the sites i n the field. A n excellent  example of this was the  fault line running north west from Bijaipur near Pokhara to Dandagaon and Jumleti , 57  traceable on maps numbers 62 P/16 and 71 D / 4 . The third category  o f failures is a response  to oversteepening,  and can be  viewed as an attempt by the slope to retreat to a stable angle. The grouping o f shallow debris slides around 38.5° and the presence o f apparently stable slopes at lower angles is confirmation o f Brunsden et al.'s (1981, p. 49) suggestion eastern  that, as i n  Nepal,  "a significant control on the distribution o f mass movement is likely to be the intensity o f lateral and vertical erosion or stream incision which maintains steep slopes, promotes basal removal o f debris, and prevents the free degradation o f the slopes to a stable angle." Oversteepening, weathering and deconsolidation are all long term causes o f slope failure, either increasing shear stress or decreasing shear strength. Shorter-term include loss o f root reinforcement  following deforestation, high pore water  causes  pressures  during the monsoon, and vibration or shock caused by seismicity . 58 Anthropic:  effects of forest  In the Phewa catchment  clearance  a large number o f shallow translational slides were  observed i n the small, recently deforested  catchment  o f the Pokhreebyasee  Khola  immediately west o f Pokhara (Figure 5). R a p i d destruction o f the forest i n this catchment  57 58  commenced with the nationalisation o f forest resources  N o t shown o n the recent 1:125,000 L . R . M . P . geological maps. Also by explosions, thunder, sonic booms and traffic.  i n 1957, and was  100  complete by the early 1970's ( Y . Khatiwada, pers. comm.). The slides all occurred between  1972 and 1978, and it is interesting to speculate that they could be due to  deterioration of root tensile strength following final deforestation. O'Loughlin (1984b) states that on clear-felled areas the landslide-susceptible period usually occurs  between  2 and 8 years after harvesting, depending on tree species, slope, and climatic conditions. The Pokhreebyasee  Khola catchment  slides fall neatly into this period.  Elsewhere i n the Phewa Valley, shallow failures were particularly numerous on the southern side of the Kaski ridge (Figure 17), but also occurred i n forest The geographical association of the majority o f these shallow failures with cleared or scrub areas can be explained, i n part, by increased runoff causing fluvial incision and gullying, and consequently the oversteepening  of slopes. Thus, there may well have  been an increase i n the number o f small, shallow landslides in the watershed due, both directly and indirectly, to deforestation. However, they are only minor contributors to total sediment production (see 3.3.2). Elsewhere i n Nepal it has been observed that although shallow slides are frequent on deforested slopes, larger, deeper failures are more common on forested terrain (B. Carson, i n press). This is most likely to be due to the location of remaining forests on steep slopes which are intrinsically liable to major failures (see In the Phewa watershed extensive  planting of Alnus nepalensis i n and around  the mass movement catchment at Pamdur failed to prevent further sliding (Figure 21), but did provide wood and fodder from otherwise useless land. Extrinsic:  pore water pressure  According to Cedergren (1967), soil water lowers slope stability by: (i)  causing soil particles to migrate to escape exits, resulting i n piping and erosional failures;  (ii) reducing or eliminating cohesive  strength;  101  (iii) increasing neutral pore water pressures and thereby reducing effective stresses and shear strength; (iv) producing horizontally inclined seepage forces which increase downslope tangential forces on soil masses and the possibility of failure; (v)  lubricating failure planes after small initial movements occur;  (vi) supplying an excess of fluid that becomes trapped in soil pores during earthquake or other severe shocks and promotes liquefaction failures.  "  In the study area overland flow was a common sight during rainstorms. Although this may have been due in part to precipitation intensities exceeding soil infiltrability, with consequent  surface runoff over unsaturated soils, the length of the monsoon and  the intensity and duration of individual rainfall events make it reasonable  to assume  that at times the piezometric surface would have risen close to or even above ground surface. The effectiveness  the  of high piezometric levels i n affecting stability has  been demonstrated by O'Loughlin (1972), who reported the maximum stable slope of a saturated, cohesionless  soil mantle with downslope seepage to be approximately half that  of an unsaturated mantle. H e is supported by Terzaghi (1950, quoted i n Thornes 1980), who noted that shallow failures may usually be accounted for' by high pore water pressures causing a reduction i n the shear strength of slope materials. Transient high water tables are thus prime candidates for explaining the initiation of the small, shallow debris slides i n the Pokhara area. In discussing groundwater conditions, it is essential to differentiate between two basic groundwater regimes, and two categories  of failure. Groundwater can be  seasonally  high, as is indicated for the shallow failures which were often on interfluves and other mid or upper-slope locations, or perennially, caused by groundwater discharge. In the first case, monsoon rains will have a direct effect on stability. In the second, the constant moisture reduces any cohesive strength and lubricates failure planes, and the low-elevation position allows the build up of high piezometric pressures under confining layers of debris. Such mechanisms could be involved i n the continuous  102  activity of the slow debris flows i n the Phewa valley, and their acceleration during the wet summer months. Whilst new failures i n Nepal almost invariably occur during the monsoon period (see e.g. Prasad's data i n Appendix 1), and often can be related to specific storm events, as yet no study of the response of piezometric levels to rainfall has been undertaken i n the country. In the K a t h m a n d u - K a k a n i area, Caine and M o o l (1982) suggested that water tables had to rise to within 2 m of the ground surface (i.e. into the weathered mantle) before sliding was initiated, but they noted little new geomorphic activity during a heavy rainfall i n June 1980. Following Crozier and Eyles (1980), this led them to emphasize the importance of antecedent  moisture conditions to  slope failure. Barring extreme  events, the magnitude o f rainstorms which trigger slope failures  in Nepal is uncertain. Caine (1980) has proposed a limiting curve for precipitation intensity below which failure is seldom recorded (Figure 11), which suggests a threshold for slope stability of approximately 100 mm/24  hr. In the Pokhara area  rainfall events with intensities above this value are a frequent occurrence  (Figure 11).  A t Pokhara Airport 150 mm/24 hr will occur at least once a year. Caine and M o o l (1982) commented that, owing to the frequency with which precipitation exceeded Caine's general threshold, extrinsic conditions might not be the most important limiting factors affecting mass wasting i n the Kathmandu area. intrinsic characteristics,  Instead,  i.e. topography and material properties, were put forward as the  primary controls. O n the other hand, Brunsden et al. (1982), discussing the magnitude-frequency question, emphasized the role of heavy precipitation i n achieving pronounced erosion. This view is supported by oral history. F o r example, as reported by Carson (in press), i n one area southwest o f Banepa near Kathmandu the villagers acknowledged that all the landslides still visible on the surrounding slopes had occurred  103  during two heavy rainfall events, in 1934  and 1971.  nature in Nepal occurred i n late September  1984,  The most recent  event of this  when the last storm of the  season  deposited approximately 700 m m of rain on the Siwaliks in 36 hours, resulting i n widespread slope failure on all land-use types (B. Carson, pers. comm.). Extrinsic:  seismicity  Seismicity has been implicated in mass wasting in a number of studies, including that of Simonett (1967), with numbers of failures tending to decrease logarithmically with distance from the epicentre. The effects and relative intensity of seismic shaking are influenced by surface slope, near-surface  geology, regolith properties, groundwater  and soil moisture conditions, and vegetation cover (Hewitt 1983). Rapid mass movements are a central feature of many high magnitude earthquakes, and often have devastating results (see  e.g. Hansen 1965). The physical and cultural factors influencing  the extent of earthquake introduction to earthquakes  damage have been reviewed by Hewitt (1983), and an and their effects in the Himalayan arc is available i n  •Chaudhury (1981). Notable earthquakes which have affected Nepal include the Assam earthquake of 1897, 1934  the Dhubri earthquake of 1930,  the B i h a r - N e p a l earthquake  which caused extensive damage in the Kathmandu Valley, and the  of  1966  earthquake i n the west of the country. The combination of seismic shock both with and without monsoon rain events is undoubtedly a trigger for slope failure in the region.  The results of the survey of hillslope processes  indicate that, historically, mass  movement is probably the principal mechanism by which slopes have evolved in the study area. Currently, the most numerous failures are shallow, translational, and have extended runout paths. They occur on all forms of slope and under all land use types, and are not usually incorporated into the drainage net  Consequently recovery is  104  rapid, and most sites appear to be revegetated after ten years. Volumetrically, the largest producers of sediment are  "mass movement catchments"  associated with structural  or lithological weaknesses. These result from specific geological conditions and may have crossed some geomorphological threshold, encouraging further growth. The shallow failures are probably triggered by transient increases i n pore water pressure resulting from prolonged intense rain. The large failures may be triggered in the same way, but are invariably associated with groundwater discharge, which maintains sediment transfer activity throughout the year. Locally, surface erosion may be responsible for rates of sediment production which are twice that attributable to mass wasting. The contributions of gullying, shallow creep, and solution to surface lowering are  unknown.  Sediment produced by these hillslope processes is transferred by mass movement or mass transport to the valley floor, where, owing to the perched base level caused by the presence  of the lake, it accumulates. The interaction of this material with  fluvial system forms the subject of the next chapter.  the  Chapter 4 F L U V I A L P R O C E S S E S EN T H E P H E W A  4.1  VALLEY  INTRODUCTION In recent years increasing attention has been paid to sediment budgets and  routing, i.e. the production, transport and discharge of detritus from a drainage basin . 59  The construction of a sediment budget for a catchment presumes the recognition and quantification of transport processes between them (Dietrich et al. detailed monitoring to measure  and storage  elements, and identification of the links  1982). Quantification of the budget requires lengthy and the components of the sediment balance, and the design  of the monitoring network itself requires the construction of an approximate budget based on preliminary studies (Dietrich and Dunne 1978). However, where episodic events are a major contributor to landscape change measurement  may not be sufficient to include these formative events.  Sediment movement is a continuous process slopes to  even extended periods of  final  extending from source areas on  deposition i n sedimentary basins inland or offshore. Once delivered to  channels by slope processes,  sediment becomes part of the fluvial system for onward  transport In transit particles undergo changes  i n size and angularity which  interact  with hydraulic factors to change the morphology of the channel system. Fluvial geomorphology has been studied extensively i n mid-latitude areas (see al.  1964;  Gregory and Walling 1973;  e.g. Leopold et  Dunne and Leopold 1978), but infrequently i n  tropical mountains. In Nepal, there is only one published quantitative study of basin and channel geomorphology to complement Brunsden et a/.'s  (1981) descriptive work on  fluvial systems i n eastern Nepal, and that is Caine and Mool's (1981) paper on two small streams i n the Mountain Hazard Mapping Project study area northwest  59  F o r a recent discussion see Swanson et al. 105  (1982).  of  106  Kathmandu (Caine and M o o l  1981). The authors concluded that the channels studied  did not behave in an "abnormal way" compared to streams that  in mid-latitude areas, and  therefore "...models derived from mid-latitude areas can be applied with few modifications to the serious stream problems of the M i d d l e Hills of Nepal." (Caine and M o o l 1981, p. 243). The Phewa watershed is unusual in Nepal in that, currently, base level is  controlled by seasonal  water levels in the lake rather than by riverbed levels  downstream. Phewa Tal has formed and been drained several times during the Pleistocene (Yamanaka  et al. 1982), and was created most recently by the  catastrophic  fluvial deposition of the sediments of the Pokhara Formation in the Pokhara Valley some 500-600 years B.P. (see The existence  of laminated lacustrine deposits in  the vicinity of Pame, currently being eroded by the Harpon Khola, suggests that this event resulted i n the formation of a stable lake surface some 2-3  m higher than at  present, unless the deposits are the result of some unrecorded dam . The lake acts as 60  a sediment trap, and is useful in permitting an estimation of the sediment yield the Harpon K h o l a from measurements comparing this to estimates  of  of sediment accumulation in its delta. By  of sediment production upstream it should be possible to  determine a sediment delivery ratio, and hence a denudation rate for the  catchment  The recent publication of Caine and Mool's (1981) paper, the interest of the sediment yield estimation exercise, and the need to understand the fluvial  component  of the sediment transport system, suggested that this field should be investigated during the present study. Accordingly, a reconnaissance  survey of fluvial characteristics  i n the  A rockfill dam was built on the Pardi Khola i n 1942, followed by a masonry dam cemented with lime and surki (ground brick) in 1958 (Sharma 1974) or 1961 (Nippon K o e i 1976a). The masonry dam failed due to piping on 2nd January 1975 (Nippon K o e i 1976a), and a concrete dam was constructed to replace i t The sluices of the new dam were closed in June 1982 and, i n order not to flood valuable khet, maintain a high water level of 793.7 m, the same as the old ;dam, rather than the design high water level of 794.7 m . 60  107  Phewa watershed was undertaken in order to determine: (i) hydraulic geometry: (ii) sediment transport and storage characteristics; (iii) sediment yield to the lake.  4.2  METHODS  4.2.1  FLUVIAL  MORPHOMETRY  Selected channel variables were determined at 15 stations along the  Andheri  K h o l a and Harpon K h o l a at approximately 1 k m intervals (Figure 31). The variables surveyed were: * width at estimated bankful stage; • cross-section; * water surface slope (estimated from channel gradient); • roughness (as estimated Manning's n). Measurements were made using a tape, level, staff, and clinometer. The data from all 15 stations are listed in Table 9, but should be treated with caution owing to the difficulty of defining some of the hydraulic characteristics slope and n. Two further morphometric parameters, capacity (cross-sectional  such as  area below  water level at bankfull stage) and hydraulic radius (capacity/wetted perimeter), were derived from these variables. Capacity was determined by planimetry from plots of the cross-sections.  4.2.2  SEDIMENT  TRANSPORT  A t each station the dominant particle size class by weight on the stream bed was determined by the pebble-count method of Leopold (1970). Random sampling of surface materials on gravel bars results i n a bias towards larger sizes since, owing to greater surface area, these are more likely to be picked up. Leopold's method avoids  F i g u r e 31.  Phewa Valley: location of stations for channel  variables  survey.  o  oo  Table 9. Andheri  STATION  DISTANCE  Khda  and Harpon Khola:  AREA  WIDTH  CAPACITY  A  (km)  W km)  (sq  WETTED  geometry, discharge,  MEAN  PERIMETER  DRAINED D  hydraulic  (m)  f<  C (sq  DEPTH  m)  (m)  (m)  0 0 4 9  9 15 9 16  SHAPE  SLOPE  W/R  S  6 14  74 37  13 12 24  33 7 1 74  68  29  4 1  51  58  82  1 . 34 2 . 24 3 .08 4 . 16  1 07 3 03 4 58 7  68  8 15  8 13 5 18  5  5 .34  9  65  24  28  6  29.5  0 89 0 87 0 60 1 18 0 97  6  6 .02  14  45  20  15  0  22  0  7  7 .03  2  32  0  8 .04  30 145  16  8  18 . 2 3 21 4 9  62  5  42  345  2  8 .95 10 . 1 0  48 06  350 172  150+ 1 1 1  0  10  26 63  150 355 175  0  9  42 63  833 273  3 0  1 1  10 12 13 14  .95 . 73 .83  65 73 74  07 57  245 85  250 88  42 15  58 65  288 68 46 69  82  69  96  102  98  0 85 1 24 1 25 0 93 1 04  2 5 4  82  212 109 75 63  5%  10%  5%  157.  12 13 14 15  ERROR  .95 15 . 6 5  5%  5%  60 68  0  92  20%  9 3  particle  ROUGHNESS  DISCHARGE  Vm  0 (cu m/s)  .57 .47  52 . 6 84 . 2 28 . 5  0 0 0 0  065 065 065  4  0  060  6 6 5 8 4  2 2.5  0  055  2 .63  0 0  2 .68  0  050 04 3 043  0 0 0  1 .5 1 . 25 <1 . 2 5 ••  "  it •1 H 10%  ? denotes uncertainty. Terms are defined in the text Error estimates are subjective assessments of the accuracy of field measurements, proportionately increased for derived values.  .26 .94 .32  169. 3 123. 6  240+ 420 300 200-470 340  4  1200+  5  220-280  2 . 10 1.93  131 289  2 5  140 95  040  2 .40  266  4  31  2 .92 3 . 76 3 .98 3 .55  619  0 8 5 7  46  0 0  040 040 038 035  0  035  3 .82  25%  40%  223 389  50%  6  SIZE  SECONDARY (mm)  (mm)  39 43  409 298  bed  PARTICLE DOMINANT  (m/s)  12 12 13 4  060  size class on channel  MEAN VELOCITY  n  (cleg)  1 2 3 4  6 12.5  and dominant  28 <2 <2 <2  _  ?  -  2000+  -  1240  7 1750 440 7  620  9  -  110  this problem by correcting for area. The procedure is outlined i n Appendix 3. In addition, at each station the a, b, and c dimensions of 20 particles moved by recent flows and 10 particles not moved were measured as potential input data for equations  relating grain size to velocity at the stream bed, and hence  (Appendix  discharge  4).  4.2.3  SEDIMENT YIELD  Sediment yield was to have been  estimated by studying delta growth i n Phewa  Tal. However, locally available data proved to be insufficient to allow any realistic estimate of sediment delivery to be made. In case it is of interest to other workers, the methodology that was to have been adopted is noted below. The difficulties encountered  with this approach are discussed i n Section 4.3.3.  The growth of the delta of the Harpon K h o l a where it enters Phewa Tal (Figures 3, 8) has been recorded by aerial photography and, recently, satellite  imagery.  By superimposing outlines of the delta from sequential images it proved possible trace the expansion of the delta (Figure 32). The the Figure derive from the air-photo coverage  1972  and 1978  outlines shown i n  described in section 3.2.1. The  outline was taken from the Indian one-inch map, which is based on 1958 photography. The 1975 March/April  1975.  outline is from a false-colour  Scale correction and image transfer  to  1958  aerial  L A N D S A T image taken in was achieved using a zoom  transferscope. Approximate lake contours are known from two bathymetric surveys (see B y measuring the areal increase  4.3.3).  of the delta over time, correcting for sediment density,  bathymetry, and water level, it was thought that a unit rate of deposition could be calculated. This estimate of sediment yield could then be checked against suspended sediment i n the Harpon K h o l a in 1979  figures  measured by Impat (1981). If  for  1958  WL  unknown  Figure 32. Growth of delta in Phewa T a l , 1958 to  1978.  112  confident i n the result, an indication of the sediment delivery ratio i n the  catchment  could be gained by reconciling the estimate of sediment yield in the river with  the  estimates of erosion computed from the survey of hillslope processes.  4.3 R E S U L T S A N D D I S C U S S I O N  4.3.1  CHANNEL  GEOMETRY  Long profiles of the Harpon K h o l a and its two tributaries, the Andheri K h o l a and Sidane K h o l a , are shown in Figure 33, together with the profile of one of the steep, ephemeral  channels on the south side of the Kaski ridge. The profiles were  developed from the 1:25,000 topographic base map ( I . W . M . steepness of the upper parts of the catchments. the old surfaces,  1979), and emphasize  the  Here, headward erosion is destroying  and a number of nickpoints can be seen on the profiles. In these  reaches (Stations 1-6,  Figure 31),  the valley profile is strongly V - s h a p e d , with  forested  slopes of approximately 45° rising for 100-200 m to a marked break which forms lower limit of cultivation. The channels have characteristics with limited width (6-24  streams,  m, Table 9), intermittent bedrock beds, rough longitudinal  profiles with angles of 2-14° and 8 a marked change  typical of mountain  the  and numerous waterfalls (Figure 34). Between  is apparent  stations 7  This is the junction between the high and low  relief areas of the watershed, and for some 500 m the Andheri K h o l a becomes transitional, with downcutting, lateral corrasion, and minor alluvial terraces all present F r o m this point until Stations  11-12, both the Andheri K h o l a and the  Sidane  Khola are braided. The river beds are wide, choked with coarse material, and laced by anastomosing channels (Figure 35). M i n o r lateral corrasion may occur where  the  thalweg abuts valley side slopes. Gradients are low, <1.5°. However, beyond Station another change  occurs. The river reverts to a single channel, which meanders  11  across a  0  1  2  3  4  5 6 DISTANCE  7 (km)  8  9  Figure 33. Phewa Valley: long profiles of the Andheri Khola, Sidane K h o l a , and an ephemeral stream on the south side of the Kaski ridge near Pame.  10  114  Figure 34. Station 5, Andheri Khola, looking downstream. Note: (1) The variability of particles sizes in the channel bed. (2) The fresh rockfall from the base of the vegetated cliff in the background directly into the stream.  Figure 35. View upstream towards Station 9, from the confluence of the Andheri Khola and the Sidane K h o l a . The braided channel has been partially controlled by a gabion diversion built to protect khet.  115  cultivated floodplain to the lake (Figure  Channel  size. Size increase  regressions  6).  downstream was irregular (Table  of capacity (C) on distance  9), although linear  downstream from source (D) and area drained  ( A ) , were significant at the 95% and 99% levels. The use of capacity as a measure  of  size allows comparison with the results of other workers. For example, a power regression of C on A and D for the 15 stations  The exponents  C  (m )  =  4.621  C  (m )  =  3.232 D '  2  2  i n these regressions  A " 0  1  7 2 0  3 0 8  gave:  (km )  (r=0.854)  (km)  (r=0.833)  2  are very similar to those found by Caine and  M o o l i n the Kathmandu area, respectively 0.784 and 1.363, and are within the of values found i n mid-latitude areas (Caine and M o o l M o o l found their coefficients  range  1981). However, Caine and  to be 0.667 for C on area, and 1.482  for C on  distance,  respectively, which they considered high. They suggested that this might indicate a relatively high yield of runoff from their catchments. the even higher Phewa coefficients  Continuing the same argument,  may reflect the steepness of the catchment  and the  intensity of monsoon rainfall events.  Channel  shape. Channel shape was highly variable, with abrupt changes graphically  illustrated i n Figure 36, a plot of shape (defined as width, W , divided by hydraulic radius, R ) against distance downstream. This variation is a reflection of the three distinct forms assumed by the stream:  the rapidly eroding upper reach involving both  headward erosion and lateral corrasion, the central braided reach, with flow disappearing in the dry season, and the meandering reach above the lake with flow confined to an active but well-defined channel. This downstream change  i n channel  morphology and valley shape is largely a result of the presence of the lake, which  116  117  has caused aggradation. It is also due i n part to the movement of coarse sediments through the system. Variations i n hydraulic geometry appear to be a complex  response  to both the water and sediment moving through the channel and the character  of the  material i n the bed and banks (Gregory and Walling 1973). Typically, braided channels occur where stream gradients drop, hydrographs are flashy, and bedload is high (Gregory and Walling 1973). A l l these conditions apply i n the Phewa valley.  Channel  roughness.  Roughness decreased  rapidly with distance downstream, which is  consistent with a diminution i n particle size. A plot of Manning's n on distance (Figure 37) could be interpreted as sigmoid curve, reflecting the rapid changes i n the channel between stations 5 and 8. Linear regression of n on distance downstream gave: n =  0.069 -  0.0024 D (km)  (r=0.958)  significant at the 95% level.  Stream  discharge.  The mean velocities ( V ) shown i n Table 9 were calculated using m  Manning's equation, an empirical equation which relates velocity to water surface  slope  (S), roughness (n), and hydraulic radius (R):  V  m  = — R / <?°" n  2  J  5  Discharge (Q) was calculated from mean velocity and capacity, and increased downstream according to linear functions with the form: Q Q  (m /s) 3  (mVs) =  =  63.8 + -  4.04 A (km ) 2  3.32 +  25.9 D (km)  (r=0.770) (r=0.711)  both significant at the 99% level. Before discussing discharge estimates, it is advisable to glance at the error values in Table 9. F o r the morphometric data these error figures are subjective assessments of the accuracy o f the field measurements,  and serve to emphasize the crudeness of the  Figure 37. Regression  of Manning's n on distance. Data from  Table  119  derived values. Nevertheless, the discharge at station 16 indicated by the 400  m V s , compares well with an estimate for the 100  D a m at the outlet of the lake of 630 m V y r , a the A n d h i K h o l a at Dumrichaur (Nippon Koei to a runoff 4.9 m ' s ^ k n r  2  figure  regressions,  year flood event at the  Pardi  derived by extrapolation from  1976b). The flood estimate is equivalent  (msk) from the whole catchment,  compared to 4.8  msk  from the regression of Q on A above. The slope of the regression line indicates a unit discharge of approximately 4 msk in addition to the intercept value. Assuming 100%  runoff (a fully saturated mande), this is equivalent to a rainfall intensity of  some 350  mm/day, a possible event i n the area (see  Figure 11). Caine and M o o l  (1981) also report estimated peak flows of up to 4.0 msk i n both the streams which they investigated near Kathmandu. The absence of flow records prevented any attempt to assess runoff coefficients, or the influence of land use on stream flow.  4.3.2  SEDIMENT  TRANSPORT  Table 9 summarizes the dominant particle size class by weight  61  found at the  various stations. In general, particle size diminished with distance downstream, although not i n a regular fashion. Below station 12 (Figure 31) were  only sand and silt fractions  present A plot of dominant size class and secondary peaks (Figure 38)  distinct phenomena:  firstly,  revealed two  an abrupt reduction in particle size between stations 7 and  8 where the stream debouches onto the valley floor; secondly, the ; presence of very large material at some, but not all, stations . The reduction i n particle size 62  Equivalent sieve size class. See Appendix 3. The very high dominant size at station 6, 1200 m m + , is a genuine product of the sampling procedure, but should obviously be classed with the occasional boulders comprising the "secondary peak" column. Finer material may have been present at the site but protected by an armour layer, and so not sampled, or the! location could have been a depositional site for large debris from a point source on the hillslopes above. 61  62  120  Dominant class by weight Secondarypeak (? denotes uncertainty) Possible trend (visual fit)  \ \ \ \  \ \  \  \  \  \ \ ?  i  \  U  \  .5  i—o  \  \  6  DISTANCE  8  10  12  14  16  (km)  Figure 38. G r a p h of dominant particle size class against distance. Data from Table 9.  121  downstream can be accounted for by both attrition and sorting, and so is a result of both lithology and process. The boulders (some are visible in Figure 33)  were  invariably quartzite, rather than the phyllitic schist predominating in the bedload, and so more resistant to abrasion and comminution. Many were far too large to be moved by normal fluvial events, which suggests that they may be relicts, deposited i n the valley bottom by hillslope processes and subsequently exposed as fluvial erosion removed the matrix surrounding them. N o records of local stratigraphy were made and so it is not possible to determine whether this matrix was alluvial or the result of a debris torrent or similar event. The sudden reduction i n particle size between stations 7 and 8 can be explained in two ways. The reduction in gradient and increase  in channel width at this point  cause a marked diminution of the river's transport capacity, and so materials  are  deposited here. In addition, it is the point at which a large debris fan is spreading across the valley, built up from sediments transported by gullies from the long slopes above. Scour depths were observed to be 1 m or less along the braided section of the river, between stations 8 and 11, increasing to approximately 2 m downstream around stations 12 to 15. Upstream, along the V - s h a p e d gorge above station 6, scour  depths  again appeared to be approximately 1 m, but the extremely rough channel precluded accurate  measurement  To estimate flow velocities and depths required to move the very large particles i n the stream bed a multiple-technique approach similar to that of Bradley and Mears (1980) could be used. Bradley and Mears were interested i n determining what magnitude of flow was required to move large boulders i n a creek i n Colorado. By using six empirical and two theoretical techniques to convert particle size to  competent  velocity, and three other equations to estimate flow depths, they were able to give a  122  confident statement that, in their stream, boulders with a mean intermediate  dimension  of 1.88 m would probably be moved by flows with a velocity of 4.7-7.6 m / s and a depth o f 3.4-4.9 m (Bradley and Mears, 1980). The associated order o f 625 m / s . Their confidence 3  values predicted by the different  4.3.3  discharge was of the  was based on the relative consistency  techniques.  SEDIMENT YIELD  Figure 39, an enlargement  and simplification of Figure 32, clearly shows the  rapid growth o f the delta of the Harpon K h o l a i n Phewa Tal between 1978.  of the  1958 and  Local opinion associates much of this growth with two erosional features, the  mass movement sub-watershed by gabion  catchment  near Pame, and a deep and active gully i n the same  climbing up the Kaski ridge near Toripani (Figure 16), now stabilised  check-dams.  In order to calculate sediment yield from a delta, it was necessary to know the composition and density of the materials deposited. Visual inspection and hand-texturing indicated a predominance of sand and silt size classes on the riverbed at station 15 on the H a r p o n K h o l a . Fleming (1978) described sediment  "from the lake bottom" as  being clay for at least 2 m depth, but the location of his sampling sites was not reported. It is probable that the delta can be classified as a bar-finger  sand delta ( V .  Galay, pers. comm.), an elongated sand body of greater density than the pro-delta sediments, and with a bi-convex form i n cross-section  (Reineck  and Singh 1973). The  bi-convexity is caused by the gradual sinking o f the sand under its own weight as the sediments below  consolidate . 63  F o r further calculation, the density of the delta deposits could be assumed from these characteristics,  but before doing so an additional problem intervened. This was  For a comprehensive introduction to lake sedimentology which emphasizes the implications for environmental management, see Hakanson and Jansson (1983). 63  —  1972  post - m o n s o o n  1978  post-monsoon  0  500  I  I  Figure 39.  1000m  I  Growth of delta in Phewa T a l , enlarged.  124  that the water level in the various images was not constant  (Figure 32), and the  shallowness of the lake in the vicinity of the delta meant that any fluctuation in water levels resulted in large movements  of the shoreline, and corresponding uncertainty  concerning the size of the delta. As mentioned in Chapter 3, the Phewa Tal has been the subject of two bathymetric surveys: in October Kraayenhagen and Impat (Impat echosounder  1976  by Ferro and Swar (1978), and in late 1979  1981). Both surveys were carried out by towing an  behind a rowing boat or dugout across the lake at steady velocity,  positions on transects by compass bearings, and interpolating the echosounder to draw isobaths  by  fixing  read-out  at 2 m intervals. Ferro and Swar (1978) present a map of these  underwater contours at a scale of 1:28,100, but information on Kraayenhagen and Impat's work is limited to a table showing areas between isobaths (Impat  for both surveys  1981). Both surveys were carried out with a water level controlled by the  broken dam at approximately 790.0 m (Nippon Koei  1976a). Ferro and Swar's  map  indicates a delta front of approximately 18 m depth, but the complications of unknown sediment composition and density, uncertain bathymetry, and fluctuating water  levels  combined to make any estimation of sediment yield by the method proposed hazardous, and it is not attempted  here.  However, before leaving the subject, it is important to emphasize that  neither  rates of sedimentation i n nor rates of sediment delivery to Phewa Tal are known. This emphasis is necessary to counteract  the tendency in the region for first approximations  to be quoted as facts. With regard to the lake, two of these approximations  are  circulating i n Nepal. The first is Fleming's estimate of sediment delivery based on a mass balance  analysis of phosphorus dynamics in Phewa Tal (in Fleming 1978,  as Fleming 1983)". In March and M a y 1978  Fleming found the concentration  repub. of  In these papers Fleming also presents a simplified water balance for the watershed, but includes i n it a value for evaporation from the lake of 51 x 10 m V y r . Since 64  6  125  phosphorus i n the surface waters of the lake to be 0.16  mg/1 . By making major 65  simplifying assumptions, including constant concentration throughout the lake, no overall gain or loss from year to year, an annual throughput equivalent to the rate of flushing (defined as total flow/lake volume), and an annual loss in sedimentation on the lake bed, he concluded that the input of phosphorus to the lake was 13,862 kg/yr. Since the soils on the  "highly erodible" north side of the watershed were  "two  thirds clay", and the sediments on the lake bottom were also clay, Fleming then made the surprising assumption that the sediment delivery ratio of nutrients and soil to the lake was also about two thirds (Fleming 1978, p. 17;  Fleming 1983,  p.240).  The second approximation is Impat's estimate of the longevity of the lake based on the bathymetric surveys and on sediment sampling i n the Harpon K h o l a " .  The  difference i n estimated lake volume between the two surveys was 3%, which is appreciably less than the error inherent in the method . Despite this, Impat 67  extrapolated the difference to give an estimated useful life for the lake, and correlated this estimate with a sediment load delivered to the lake by the Harpon K h o l a some 9.84  t ha^yr  sampling i n 1979  1  of  (Impat 1981). This last value is based on suspended sediment  at Chankapur, the meander below Pame. Individual grab  were taken daily between A p r i l and December  from "a little bit below the  samples surface",  analysed for sediment concentration, and the results multiplied by an estimate of discharge (obtained from cross-sectional  "(cont'd)  area and estimated velocity) to give daily  the lake area is only some 6 k m  2  this is an evaporation rate of some 8500  mm/yT!  Usually indicative of eutrophication. Earlier limnological investigations i n the Pokhara area are reported i n Hickel (1973). Nippon K o e i C o . also estimated lake life during their design studies for the new Pardi D a m . They based their figures on data from the Poonch River in Pakistan "because of similar geographical features in both river basins" (Nippon K o e i 1976b). The Poonch River basin has an area of 2470 k m and is not directly comparable with the Phewa watershed, either i n terms of area or physiography. See Rausch and Heinemann (1984) for a description of techniques for measuring reservoir sedimentation. 65  66  2  67  126  suspended sediment transport The total sediment load for the year was then calculated, including estimates  for the first three months of the year, and an assumed 20%  contribution for bed load (Impat 1981). Readers are referred to the authors noted, under Table 1 for an explanation of the errors inherent i n such a technique.  4.3.4  SEDIMENT SYSTEMS  The sediment production and transport system in the valley reflects, on a more moderate scale, many of the features described by Brunsden et al. (1981) in eastern Nepal. The fluvial system is supplied with sediment from a number of different sources, including: * mass movement activity on river banks and lower valley slopes; * lateral corrasion by rivers and streams; * vertical incision of headwater channels; * erosion of fans on the valley floor; * mass movement *  catchments;  ephemeral streams and gullies;  * generalised surface erosion. Except for activity i n mass movement catchments, sediment movement is confined almost entirely to the monsoon season. Material is transferred during storms from the valley side slopes to streams and river channels. M u c h of this sediment arrives as pulses of debris when rivers are i n flood, and is deposited as the flood stage falls, both on the valley bottom, and as fans at the base of the slopes. Both these sediment stores are continually reworked as channels meander over their surfaces. Comminution proceeds rapidly owing to the softness of the principal rock types, with both attrition and sorting contributing to reductions in dominant particle size with distance downstream.  127  Currently, the transport capacity been exceeded.  of the fluvial system in the valley bottom  has  The river is energy-limited owing to the perched base level caused  by  the Pardi dam. Very large quantities of coarse material remain to be moved from  the  upper end of the valley floor to Phewa Tal. Data were insufficient to allow the estimation of sediment  residence  time i n fans or gravel bars. However, the remote  sensing images provide an excellent  record of channel changes in the valley over time,  and could be very useful in a study of bank recession meander  bends.  and downstream migration  of  Chapter 5 SUMMARY  5.1  AND CONCLUSIONS  THE PROBLEM As elsewhere  in the Himalaya, the Middle Mountains of Nepal are  settled, and support an agroecosystem subsidies from the forest  densely  which is dependent on energy and nutrient  for its continuation. The land surface  has been  extensively  modified in order to allow arable agriculture, and increasing population and unfavourable institutional arrangements  have resulted in degradation of the  forest  resource. Recurrent, and supposedly worsening, landslides, floods, and associated deposition downstream have resulted i n concern being expressed  sediment  that some critical  environmental threshold has been reached, and further deterioration becomes inevitable. The principal cause of this environmental deterioration is widely perceived to  be  deforestation. Owing to remoteness and lack of infrastructure, little work has been done to define rates of natural processes in the Himalaya, or how these processes are currently affecting the landscape. Original studies on hillslope and fluvial processes are few i n number, sometimes  5.2  of limited availability, and often of doubtful accuracy.  EROSION IN THE PHEWA  VALLEY  The Phewa Valley lies i n the Middle Himalaya of Nepal at the foot of the Annapurna massif. It has an east-west structural trend, and is formed in moderately hard to weak metamorphosed rocks, with phyllite predominating. The area of the watershed is approximately 122  k m , and elevations range from 800 m to 2500 m. 2  M e a n annual precipitation is dependent on elevation, and averages 4202 m m .  128  129  Between  500  and 1100  years ago alluvial deposits accumulated rapidly, or  catastrophically, i n the Pokhara Basin, and blocked the a lake, Phewa Tal. Sediments  by ephemeral  entrance to the valley, forming .  from incising streams i n the headwaters  have accumulated i n the base of the valley, creating  above  the  lake  a flat valley floor. Debris carried  streams on the steep valley sides has been deposited at the  side slopes, forming alluvial  perhaps  foot of the  fans.  Mass movement processes identified in the area included rockfalls, rockslides, shallow translatidnal failures, flows, and creep. or structural discontinuities, complex movement  activity was seasonal,  groundwater discharge are  In areas associated  with incompetent  rock  failures with steep debris tracks developed. Mass  with dry season movement  confined to flows in  zones. N o data relating individual failure events to precipitation  available. 90% of all the material displaced by mass wasting in the watershed originated i n  large failures, which had a mean estimated  age o f 24 years. Shallow debris slides, the  most common form of failure, had a mean volume of approximately 400 mean estimated  age of 5.5  due to landsliding is 2.5  3  years. A n estimate of surface lowering i n the Phewa Valley  mm/yr, based on the morphometric attributes  ages of the sample surveyed, and on air photo Surface erosion i n the watershed, estimated runoff plots, is approximately 5-6 protected pasture, and < 1  m , and a  and  estimated  interpretation. from very limited data from small  m m / y r on overgrazed areas, 1-2  m m / y r on  m m / y r under forest Soil loss from cultivated areas has  not  been measured. The total amount of soil lost by surface erosion remains unknown. Active gullies existed on the Kaski ridge, but no reliable production from them are available. Losses known.  figures  on  sediment  due to solution and shallow creep are  not  130  If not delivered directly to the valley floor, materials displaced by mass movement activity are transferred downslope in high angle channels which are transitional between  streams and gullies. The ratio of material displaced to material  delivered to the valley floor is not known, but is probably high, owing to the steepness of the relief and the intensity of evensL Once i n the valley bottom, sediments  undergo fluvial sorting. The tributaries  the main river, the Harpon Khola, have braided channels  of  for some 4 km after they  debouch on to the valley floor, and sediment transport here is energy-limited. Hydrographs are flashy. Flow estimates suggest a unit runoff at bankfull stage of 4 m s k n r , which, with 100% 3  -1  about 350  2  runoff, is equivalent to a 24 hr precipitation value of  m m . Intensity-duration-frequency  curves developed from 7-10  years of data  at Pokhara Airport indicate that values i n the region of 300 mm/day are unreasonable.  Reliable records of rainfall approaching this figure  not  exist  The delta of the Harpon K h o l a i n Phewa Tal has grown rapidly over the 30 years, possibly associated  some  last  with specific erosional sites on the slopes on the north  side of the valley. The records of its growth provided by remote sensing could be used to determine  sediment yield, given information on delta volume and density not  currently available. Several authors have attempted  to quantify rates of erosion and sedimentation in  the Phewa Valley, but often the original data have been extrapolated unwisely. The author of the present study would like to emphasize that his own figures for failure age, volume, and frequency (and hence the estimate of surface  lowering by landsliding),  should never be used without the prefix "based on a small sample". They  are  probably as valid, or invalid, as the estimates made by Caine and M o o l (1982), and Starkel (1972a, 1972b).  131  5.3 T H E E R O S I O N S Y S T E M  IN THE MIDDLE  HIMALAYA  The few studies of Himalayan geomorphology that have been carried out to date suggest a dynamic environment in which orogenesis, relief and climate combine to give rates of denudation in large catchments  of up to 5 m m / y r . Locally this rate may be  exceeded. The principal mechanism by which this high rate of denudation is achieved appears to be an integrated slope development and sediment transfer system which is dominated by mass wasting on slopes, and which is synchronized with high discharges i n river- channels. The frequency of formative events is unknown, but values ranging from 10-25  years have been proposed. A t less frequent intervals, perhaps on a scale  of centuries, catastrophic changes  occur, with major slope failures precipitating both  channel scour and terrace accumulation, as i n the Pokhara Basin. The erosional system is characterised by extreme  seasonality, with virtually all  work being carried out during the monsoon months of July, August, and  September.  Within this period, slope failure and fluvial sediment transfer occur episodically, triggered by intense and prolonged rainfalls which commonly exceed  150 mm/day. The  coincidence of seismic shaking with one of these rain events, or simply with high groundwater conditions, causes widespread failure release,  but the magnitude of shock  needed is not known. Hillslopes in the Middle Himalaya display a wide range of mass  movements,  including translational and rotational failures, flows, creep, and transitional forms. Volume and velocity of movement are equally diverse. The type of material involved varies according to lithology and relief. Shallow failures generally remove  mantle  material, which consists of either untransported regolith, or colluvium from earlier erosion/deposition cycles. Large failures often involve rock weakened by deep weathering.  132  The most common failures are shallow translational slides with extended runout paths. Typically, these occur in mid-slope positions, and heal within 5-10  years. Larger  failures are associated with undercutting, unfavourable geology, and structural discontinuities. Particularly dynamic failure complexes develop where these combine, and may become  self-reinforcing through the interaction of mass  and mass transport processes, catchments  factors  area exposed, and precipitation. Such mass  movement  movement  are closely associated with, and sometimes transitional to, high-angle fluvial  features. Runoff generated  in the larger slides often results in the rapid integration of  failure scars into the drainage  net  Sediment transferred to the valley bottom by mass movement or mass is either deposited in fans where gradients lessen, or enters stream  transport  channels.  Down-channel movement is then proportional to clast size and river stage. Turnover times for the various sediment stores are not known. Intense precipitation and steep slopes combine to cause high runoff, and this is reflected in the rapid rise and fall i n stage noted on the few rivers i n Nepal for which discharge measurements  are available. The entry into the channel of displaced  materials from point sources during storm events results in pulses of sediment moving downstream. Suspended sediment concentrations  of <25,000 ppm have been recorded in  the Narayani. These materials are deposited i n the Terai, forming vast alluvial fans. The principal controls on this system are geological and climatic, and disturbance o f the land surface is unlikely to have any great effect on long-term rates of • landscape change. However, i n the short term, deforestation and construction activities have both been implicated in increasing rates of sediment production. Deforestation enlarges the area of marginal agricultural land, which is then subject to surface  I  erosion, gullying, and possibly shallow landsliding. Construction activities cause deeper failures. Insufficient data are available to quantify man's influence on the magnitude  133  and frequency of erosional events, but on the basis of the evidence available, it seems reasonable  to suggest that man may have caused a slight increase  in rates of mass  wasting, but has had a marked effect on surface erosion and gullying. Although  the  latter have a more insidious effect on the productivity of the village agroecosystem,  the  former are nore immediately hazardous, and provide much of the sediment input to the fluvial system.  5.4 C O N C L U S I O N S  A N D RECOMMENDATIONS  Rates of erosion and sedimentation i n the Himalaya are extremely high. T o live in harmony with such a dynamic geomorphological environment, man must learn to accommodate amenable  these natural processes,  and to discriminate between those which  are  to modification and those which are not The following points should be  made: (1)  In Nepal, the causes of mass movement, sensu stricto, are primarily  geological, and so cannot be influenced by man. Intervention is extremely  expensive,  and can only be justified where high-value infrastructure is threatened. Even then, it is not always successful. (2)  Engineering structures must allow for this environment; they cannot hope to  subdue i t Briefly, particular consideration should be given to: * Siting: linear features (roads, canals) the most unstable areas.  require careful alignment to avoid  * Lower specifications: environmental impact can be reduced by, e.g., reducing road widths, so that a smaller area of ground is affected. * Adjusting the design and management of hydraulic installations to cope with high sediment loads. F o r example, sediment intake to pumps and irrigation canals can be reduced by adjusting pumping schedules and incorporating sediment bypass features i n the intake structures. * Adapting specifications to ensure the survival of structures when major hazards cannot be avoided, albeit at the cost of slightly lower performance. F o r example, low level road crossings survive floods better than bridges,  134  and  can be constructed at much lower cost.  • Planning for high levels of siltation behind dams. In most situations, catchment conservation programmes will have negligible impact on fluvial sediment load (although benefitting individual farmers).  (3)  Mass movements and associated  erosional features are the principal  contributors of material to valley bottom sediment transport systems. They are probaby responsible for the very high peak sediment loads recorded in Himalayan rivers. (4) Infrequent catastrophic events, up to several orders of magnitude larger than the majority of failures, have had and will continue to have a major effect on the landscape in Nepal. These usually involve slope failure and subsequent landslide-dams and  dam-bursts. (5) Deforestation is unlikely to affect the scale and timing of large slope  failures, but may increase the incidence of shallow debris slides. In volumetric terms, these small slides do not appear to be major contributors of sediment to river systems. (6) Throughout the Middle Mountains, deforestation and the abandonment marginal arable land are associated with an increase  of  in the area of barren land and  unproductive communal grazing areas. These sites are rapidly degrading, and, locally, are responsible for high rates of sediment production through surface erosion and gullying. (7)  Soil loss from terraces,  khet or well managed bari, is probably low, although  not as low as the loss from forested areas. (8) The nutrients carried by eroded soil are useful, i f not essential, i n maintaining fertility i n fields at lower elevations. (9)  Notwithstanding the role of erosion i n nutrient transfer, the degradation of  surface soils is a serious problem in the M i d d l e Mountains. This is due to not only the loss of soil off-site, but also deterioration i n some of the physical and chemical soil properties which affect fertility. The most important of these are structure,  135  permeability, and organic matter content  The consequences are a lower production  potential, and a higher susceptibility to erosion. (10)  If the productivity of marginal areas drops sufficiently, due to soil  deterioration and loss, they are abandoned. In the absence of management livestock increase  site degradation by trampling and the suppression of  Owing to the generation  of excessive  free-ranging  vegetation.  runoff these areas then become the sites of gully  initiation. (11)  Once started, gullying is difficult to control, and can rapidly destroy terraces  and other productive land by lateral and headward expansion. The large volume of coarse sediments produced by gullies cutting into mantle materials forms a hazard to bottom lands. G u l l y prevention is simpler than cure. (12)  The emphasis in conservation programmes should be to reduce  erosion and gullying by improving management (13)  on these low-productivity areas.  Forests, which supply the nutrients to maintain crop yields, have  due to excessive excessive  surface  retreated  harvesting of forest products and damage to young growth. Localized  harvest occurs where, for socio-economic  or political reasons,  normal methods  of distributing demand over a larger area have broken down. Concentric rings of  forest  degradation then spread out from consumption centres, i.e. villages. Such a pattern is widespread i n the Middle Hills of Nepal. Forest productivity can be improved by reinstating methods for spreading demand over the larger area. Where this improvement is insufficient owing to an absolute limit on yields imposed by the degraded condition of the forest  improved silvicultural practices can assist geographical control of  harvesting i n increasing productivity. (14)  The majority of forest products go towards maintaining the  livestock  population. Supplying alternative sources of fodder, such as forage crops, will  reduce  the pressure on the forest  must  However, any attempt to diversify fodder sources  136  confront the problems of the extra labour demand which it may create. W h o will harvest the grass for the stall-fed cow, which previously collected its own feed? H o w will the new crops be protected from unrestrained animals? A n y alteration of land management  practices  at the village level (the only valid level for improving the lot  of the individual), requires both an objective to the village, (15)  68  appraisal of the land resources  available  and a thorough appreciation of the social dynamics of the community.  Although not insoluble, the scale of the problem of environmental  deterioration in N e p a l defies imposed solutions. The most effective  kind of  involves d a y - t o - d a y  decisions, and should devolve directly to resource  case the panchayats.  The community forestry programme  69  management  users, in this  is a useful step i n this  direction. * * * » * » » * » » * * It is interesting that, in addition to man affecting erosion processes, the hazardous environment has also affected  man. In order to live in and off a  where dangerous high magnitude geomorphological events occur on a  quasi-continuous  basis, man has had to adapt his behaviour. By definition, such behavioural cannot involve complete  landscape  adaptations  hazard avoidance, but they do include both physical  damage-limitation. and control techniques religious defence mechanisms -  (see  Johnson  et al.  1982), and cultural and  the H i n d u philosophy of Majaburi  or "things we must  bear" (Carson, i n press). Physical damage-limitation involves practices such as land use deintensification, e.g. changing irrigated khet to rainfed bari to reduce water  saturation  when signs of slope movement appear, and the construction of levees along river banks. Majaburi  developed as a philosophical response  to the landslides,  earthquakes,  See, e.g., Carson (1985) for a technique for "rapid rural appraisal", based on the use of large scale aerial photographs, and applicable to Nepal. See also Shah and Shreier (1985) and Whiteman (1985) for recent exercises i n , respectively, land evaluation in Kailali District, and experimental agronomy in the Jumla area. See G i l m o u r and Applegate (1984); Pelinck et al. (1985). 68  69  137  floods,  droughts, and other recurring but unpredictable disasters common on the  subcontinent,  and performs an essential  function i n providing moral support i n the  of calamity. It is the appropriate attitude to adopt when faced by a large, landslide. However, it is not appropriate to apply the same philosophy to erosion and gullying. Both constitute resource  face  deep-seated surface  an immediate and severe threat to the rural  base, and, as a product of man's misuse of the environment, are intrinsically  controllable.  138  Bibliography Ahuja, P.R., and P.S. Rao. 1958. Hydrological aspects of floods. Proc. Symp. on Meteorological and Hydrological Aspects of Floods and Droughts in India (New Delhi). Cited i n : Brunsden et al. 1981 (q.v.). A k i b a , C . 1980. Classification of the fluvial terraces in Central Nepal. J. Hokkaido Univ. of Education (Section II B), V o l . 30, pp 179-188. Cited i n : Yamanaka al.  1982  et  (q.v.).  Anonymous. 1978. Tropical forest ecosystems: a state-of-knowledge U N E S C O / U N E P / U N F A O , Paris. 683 pp.  report  . 1980. Management plan for the integrated development of Phewa Tal watershed, 1980-1985. Integrated Watershed Management Project D e p t of Soil Conservation and Watershed Management M i n . of Forests, Kathmandu. Phewa Tal Tech. Rep. N o . 12. Asian Development Bank. 1982. Nepal Agriculture Sector Strategy Study: V o l . II: Detailed sector review and analysis. A D B / H M G , Kathmandu. 231 pp, 131 app. Bajracharya, D . 1981. Implications of fuel and food needs for deforestation: an energy study of a hill village panchayat of Eastern Nepal. Unpub. P h . D . thesis, Univ. of Sussex. . 1983 Deforestation i n the food/fuel context: historical and political perspectives from Nepal. Mountain Res. and Dev., V o l . 3, N o . 5, pp 227-240. Bansode, R.B., and S.R. Pradhan. 1975. Landslides i n Nepal Himalaya and their influence on the K o s i D a m . In: Proceedings of a Seminar on Landslides and Toe Erosion Problems with Special Reference to Himalayan Region, Gangtok, Sikkim, 24-26 Feb. 1975. Indian Soc. of Eng. Geol., Calcutta, pp 247-254. Barry, R . G . , and J.D. Ives. 1974.  Arctic and alpine environments.  999  pp.  Blaikie, P., J. Cameron, & D . Seddon. 1980. Nepal i n crisis: growth and stagnation the periphery. Clarendon Press, Oxford. 311 pp. Blong, R.J. 1973. A numerical classification of selected landslides of the slide-avalanche-flow type. Eng. G e o l , V o l . 7, pp 99-114. Bordet  P. 1961. Recherches geologiques C . N . R . S . , Paris. 275 pp.  at  debris  dans l'Himalaya du Nepal, region du Makalu.  Bordet  P-, M . Colchen, M . Krummenacher, P. L e F o r t R. Mouterde and J . M . Remy. 1971. Recherches geologiques dans l'Himalaya du Nepal, region de la Thakkhola. C . N . R . S . , Paris. 279 pp. Cited i n : Fort and Freytet 1982 (q.v.).  Bradley, W . C . , and A.I. Mears. 1980. Calculations of flows needed to transport coarse fraction o f Boulder Creek alluvium at Boulder, Colorado: Summary. G e o l . Soc of Amer. Bull., Part I, V o l . 91, pp 135-138. Brahski, J.  1981.  Accuracy of estimating  basin denudation processes from  suspended  139  sediment measurements. In: Erosion and Sediment Transport Measurement: Symposium. I.A.H.S. Pub. N o . 133. pp 213-218. Brown, L . R . , E . W . Wolf, L . Starke, W . U . Chandler, C . Flavin, and S. Postel. 1984. State of the World 1984. WorldWatch Institute Report. W . W . Norton & Co., N . Y . 252 pp. Brune, G . M . 1948. Rates of sediment production in the midwestern United States. U . S . D . A . S.C.S. Tech. Pub. N o . 65. 40 pp. . 1953. Trap efficiency of reservoirs. Trans. A m . Geophys. Union, V o l . 34, N o . 3, pp 407-418. Brunsden, D . 1973. The application of systems theory to the study of mass movement Geologica Applicata e Idrogeologia. Univ. of Bari. V o l . 8, N o . 1, pp 185-207. Cited i n : Hansen 1984 (q.v.). Brunsden, D . , D . K . Jones, R.P. Martin and J.C. Doornkamp. 1981. The geomorphological character of part of the Low Himalaya of Eastern Nepal. Z . Geomorph. N . F . , S u p p l . - B d . 37, pp 25-72. Brunsden, D . , and J.B. Thornes. 1979. Landscape sensitivity and change. Trans. Inst B r i t Geographers. V o l . 4, pp 463-484. Bull, W . B . 1975. Allometric change 1489-1498.  of landforms. Gepg. Soc Amer. Bull. 89, pp  Caine, N . 1980. The rainfall intensity-duration control of shallow landslides and debris flows. Geogr. A n n . V o l . 62A, N o . 1-2, pp 23-27. Caine, N . , and P . K . M o o l . 1981. Channel geometry and flow estimates for two small mountain streams in the Middle Hills, Nepal. Mountain Res. and Dev., V o l . 1, N o . 3-4, pp 231-243. . 1982. Landslides in the K o l p u K h o l a drainage, M i d d l e Mountains, Nepal. Mountain Res. and Dev., V o l . 2, N o . 2, pp 157-173. Carson, B. 1985. Aerial photography as a base for village level planning i n Nepal. Pap. presented at Regional Symp. on the Remote Sensing Applications of Soicioeconomic Aspects of the Environment M a r . 4-9, Jogjakarta, Indonesia. Land. Res. Mapping Proj., Kathmandu. 19 pp.  Int  . In press. Erosion and sedimentation processes Centre for Integrated Mountain Dev., Kathmandu.  Carson, M . A . 1971.  i n the Nepalese Himalaya.  The mechanics of erosion. Pion, London.  . 1915. Threshold and characteristic angles of straight slopes. In: E. Yatsu, A . J . W a r d and F . Adams (eds.), Mass Wasting: Proceedings of the Fourth Guelph Symposium on Geomorphology. G e o Abstracts, Norwich, pp 19-34. . 1976. Mass wasting, slope development and climate. In: E Derbyshire (ed.), Geomorphology and Climate. Wiley-Interscience, London, pp 101-136.  140  Carson, M . A . , and M . J . Kirkby. 1972. pp. Cedergren, H . R . 1967.  Hillslope form and process.  C . U . P . , London. 475  Seepage, drainage and flow nets. J. Wiley, N . Y . 534 pp.  Chakrabarti, D . C . 1971. Investigation on erodibility and water stable aggregates of certain soils of Eastern N e p a l . J. Indian Soc. Soil Sci., V o l . 19, N o . 4, pp 441-446. Cited i n : Brunsden et al. 1981 (q.v.). Champion, H . G . , and S.K. Seth. 1968. Govt, of India, Delhi. 511 pp.  General silviculture for India. Pub. Branch,  Chatra Research Centre. 1976. Results of runoff and erosion experiments on small catchments. Chatra Forest Research Centre, Nepal. In: Annual report, 1976. Cited in: Laban 1978 (q.v.). Chaudhury, H . M . 1981. Earthquakes i n the Himalaya. In: J.S. L a l l (ed.)., The Himalaya: Aspects of Change. O . U . P . , Delhi, pp 99-119. Coates, D . R . 1977. Landslide perspectives. In: D . R . Coates (ed.), Landslides. Reviews in Eng. Geol., V o l . I l l , G e o l . Soc. Amer. pp 3-28. Coburn, B. 1982. Nepali aama: portrait of a Nepalese Santa Barbara, C a l . 165 pp.  hill woman. Ross Erikson Inc.  Coulomb, C . A . 1776. Essais sur une application des regies des maximis et minimis a quelques problemes de statique relatif a l'architecture. Memoirs presentees par divers Savants. Acad, des Sci., Paris. Cited i n : Gardiner and Dackombe . 1983 (q.v.).  Crozier, M . J . 1973. Techniques for the morphometric analysis of landslips. Z . Geomorph. N . F., V o l . 17, N o . 1, pp 78-101. . 1984. F i e l d assessment of slope instability. In: D . Brunsden and D . B . Prior (eds.), Slope Instability. Wiley-Interscience, Chichester, pp 103-142. Crozier, M . J . , and R.J. Eyles. 1980. Assessing the probability of rapid mass movement 3rd A u s t r a l . - N . Z . Conf. Geomechanics, Wellington. N . Z . Inst. Engin., Proc. Tech. Groups, 6(1G), Part 2, pp 2-47 to 2-51. Cited i n : Caine and M o o l 1982 (q.v.). Curray, J.R., and D . G . Moore. 1971. Growth of the Bengal deep-sea fan and denudation i n the Himalayas. G e o l . Soc. Amer. Bull. V o l . 82, N o . 3, pp 563-572. Dalai, S.S., W . U l l a h , and B. Singh. 1961. Fires in relation to watershed Proc X t h Silvi. Conf. Dehra D u n . pp 236-241.  management  Das, K . N . 1968. Soil erosion and the problem of silting in the Kosi catchment J. Soil and Water Conservation i n India, V o l . 16, N o . 3-4, pp 60-67. Davidson, C . F . 1965. Mass-movements on the Gisborne coastline. U n i v . of Canterbury, N . Z . C i t e d i n : Crozier 1973 (q.v.).  of  Unpub. M . A . thesis,  141  D e Terra, H . , and T.T. Patterson. 1939. Studies on the Ice Age i n India and associated human culture. Carnegie Institution of Washington, Pub. N o . 493. Washington, D . C . 354 pp. Dickinson, W . T . 1981. Accuracy and precision of suspended sediment loads. In: Erosion and Sediment Transport Measurement: Symposium. I.A.H.S. Pub. N o . 133. pp 195-202. Dietrich, W . E . , and T. Dunne. 1978. Sediment budget for a small catchment mountainous terrain. Z . Geomorph. N . F., Suppl. Bd. 29, pp 191-206.  in  Dietrich, W . E . , T. Dunne, N . F . Humphrey and L . M . Reid. 1982. Construction of sediment budgets for drainage basins. In: F.J. Swanson, R.J. Janda, T. Dunne, D . N . Swanston (tech. eds.), Sediment Budgets and Routing i n Forested Drainage Basins: Workshop Proceedings. U . S . D . A . For. Serv. Gen. Tech. Rep. P N W - 1 4 1 . Portland, Oregon, pp 5-23. Dobremez, J.P., and C . Jest 1974. Carte ecologique de la region Annapurna-Dhaulagiri (Nepal): 1:250,000. Cahiers nepalais, C . N . R . S . , Paris. Dollfus, O., and P. Usselmann. 1971. Recherches geomorphologiques dans le Centre-Ouest du Nepal. Cahiers nepalais, Musee de r H o m m e , Paris. 56 pp. Donovan, D . G . 1981. Fuelwood: how much do we need ? Newsletter ( D G D 14), Inst of Current W o r l d Affairs, Hanover, N . H . Cited i n : Thompson and Warburton 1985  (q.v.).  Dunne, T., and L . B . Leopold. 1978. Water i n environmental planning. W . H . Freeman and C o . , San Francisco. 818 pp. D u t t G . N . 1966. Landslides and soil erosion in the Kalimpong Sub-division, Darjeeling District and their bearing on the North Bengal floods. In: Landslides and hill stability i n the eastern Himalayas. Bull. G e o l . Surv. India, Series B, N o . 15, p t 1, pp 61-78. Dwivedi, A . P . 1980.  Forestry i n India. J. Kishore & Co., Dehra D u n . 445 pp.  Eckholm, E.P. 1975a. Losing ground i n S h a n g r i - L a : report from Kathmandu. Saturday Review, O c t 4, 1975. . 1975b. The deterioration of mountain environments. Science, 764-770.  V o l . 189, pp  . 1976. Losing ground: environmental stress and world food prospects. WorldWatch I n s t / U . N . E . P . W . W . Norton and Co., N . Y . 223 pp. Elwell, H . A . 1984. Soil loss estimation: a modelling technique. In: R . F . Hadley and D . E . Walling (eds.), Erosion and Sediment Y i e l d : Some Methods of Measurement and Modelling. G e o Books, Norwich, pp 15-36. Fairbridge, R . W . (ed.). 1968. The encyclopaedia of geomorphology. Encycl. of Earth Sciences Series, V o l . III. Dowden, Hutchinson, and Ross. Stroudsburg, Pa. 1295 pp.  142  Ferguson, R.I. 1984. Sediment load o f the Hunza River. In: K . J . M i l l e r (ed.). International Karakoram Project, V o l . 2. C . U . P . , Cambridge, pp 581-598. Ferro, W . , and D . B . Swar. 1978. Bathymetric maps from three lakes i n the Pokhara Valley (Nepal). J. Inst. Sci., V o l . I, pp 177-188. Kathmandu. Fetzer, D . K . , and L . Jung. 1978/79. Factors affecting soil erosion by water within a research area of northwestern mountain ridge of Kathmandu Valley. J. Nep. Res. Centre, V o l . 2/3 (Sciences), pp 219-236. Fleming, W . M . 1978. Phewa Tal Watershed Management Proposals. Integrated Watershed Management Project, Dept. of Soil and Water Conservation, Ministry of Forests, Kathmandu. Phewa Tal Tech. Rep. N o . 5. 90 pp. . 1983. Phewa Tal Catchment Management Programme: benefits and costs of forestry and soil conservation i n Nepal. In: L.S. Hamilton (ed.)., Forest and Watershed Development and Conservation i n Asia and the Pacific. Westview Press, Boulder, C o l . pp 217-289. Fort, M . B . 1979. Etudes sur le Quaternaire de l'Himalaya: la haute vallee de la Buri Gandaki, Nepal. Editions du C . N . R . S . , Paris. 234 pp. . In press. Contribution of sedimentary and geomorphic data to the knowledge of palaeoclimates i n Nepal Himalayas. In: D . P . Agrawal (ed.)., Proc. of the International Workshop on the Late Cenozoic Palaeoclimatic changes i n Kashmir and Central Asia, 19-23 O c t 1982, Ahmedabad. Today and Tomorrow's Pub., N e w Delhi. Fort, M . B . , and P. FreyteL 1979. L'evolution sedimentaire recente du bassin intramontagnard de Pokhara (Himalaya, Centre Ouest Nepal). Comptes Rendus, Acad. Sci. Paris, L 289, Serie D , pp 1195-1198. . 1982. The quaternary sedimentary evolution of the intra-montane basin of Pokhara i n relation to the Himalaya Midlands and their hinterland (West Central Nepal). In: A . K . Sinha (ed.), Contemporary Geoscientific Researches i n Himalaya, V o l . 2, Dehra D u n , India, pp 91-96. Fort, M . B . , P . G . White, and B . L . Shrestha. 1984. 1:50,000 geomorphic hazards mapping in Nepal. Communications du Colloque "Mouvements de Terrains", Caen, March 1984. Centre de Recherches en Geographie Physique de l'Environnement, Assoc. Francaise de Geographie Physique, pp 186-194. Fox, J . M . 1982. Managing public lands i n Nepal: a village level look at land-use planning. Paper presented at 11th A n n . Conf. on S. Asia. S. Asian Area Center, Univ. of Wisconsin, 5-7 N o v . 1982. . 1983. Managing public lands i n a subsistence economy: the perspective from a Nepali village. Unpub. P h . D . thesis, D e p t of Agric. Econ., U n i v of Wisconsin, Madison. 246 pp. Freytet, P., and M . B . Fort. 1980. Les formations plio-quaternaires de la K a l i Gandaki et du bassin de Pokhara (Himalaya du Nepal). Bull. Assoc. Geogr., Paris, N o . 471. pp 249-257.  143  Gansser, A . 1964.  Geology of the Himalayas. W i l e y - Interscience,  Gardiner, V . , and R. Dackombe. 1983. U n w i n , London. 254 pp.  New York. 289 pp.  Geomorphological field manual. Allen and  Gilmour, D . A . 1984. Keynote address: people, forests and erosion - experience from the Middle Hills of Nepal. In: C . L . O'Loughlin and A . J . Pearce (eds.), Symposium on Effects of Forest Land Use on Erosion and Slope Stability. I U F R O / N . Z . For. S e r v . / U S D A For. Serv./East-West Center, Honolulu, Hawaii, pp 15-22. Gilmour, D . A . , and G . B . Applegate. 1984. Community forestry as an option for containing environmental degradation - a case study from Nepal. In: C . L . O'Loughlin and A . J . Pearce (eds.), Symposium on Effects of Forest Land Use on Erosion and Slope Stability. I U F R O / N . Z . For. S e r v . / U S D A For. Serv./East-West Center, Honolulu, Hawaii, pp 41-46. Glacken, C J . 1967. Traces on the Rhodian shore; nature and culture i n Western thought from ancient times to the end of the eighteenth century. Univ. of C a l . Press, Berkeley. 763 pp. Goldstein, M . C , J . L . R o s s and S. Schuler. 1983. F r o m a mountain-rural to a plains-urban society: implications of the 1981 Nepalese census. Mountain Res. and Dev. V o l . 3, N o . 1, pp 61-64. Gole, C . V . , and S.V. Chitale. 1966. Inland delta building of Kosi River. J. Hydraul. Div., A . S . C . E . , V o l . 92, N o . H Y 2 , pp 111-126. Graham, J. 1984. Methods of stability analysis. In: D . Brunsden and D . B . Prior (eds.), Slope Instability. Wiley-Interscience, Chichester, pp 171-215. Gray, D . H . , and W . F . Megahan. 1981. Forest vegetation removal and slope stability in the Idaho Batholith. U S D A For. Serv. Res. Pap. I N T - 2 7 1 . 23 pp. Green, R . 1978. The construction and management of bench terracing systems i n the hill areas of Nepal. Integrated Watershed Management Project, D e p t of Soil Conservation and Watershed Management, M i n . of Forests, Kathmandu. Working Pap. N o . 4. Gregory, K . J . , and D . E . Walling. 1973. and Sons, N . Y . 456 pp. Gurung, H . B . 1970. pp 37-49.  Drainage basin form and process. John Wiley  Geomorphology of Pokhara Valley. Himalayan Review, V o l . II-III,  Hagen, T. 1965. Ubereine Uberschiebung der tertiaren Siwaliks uber das rezente G a n g e s - A l l u v i u m i n Ostnepal. Geogr. Helvetica, V o l 11, N o . 3, pp 217-219. . 1969. Report on the geological survey of Nepal: V o l . 1: preliminary reconnaissance. Denkschriften der Schweizerischen Naturforschenden Gesellschaft, V o l . 86, N o . 1, pp 1-185. Cited i n : Fort and Freytet 1982, and Yamanaka et al.  1982  (g.v.).  144  Hakanson, L . , and M . Jansson. 1983. N . Y . 316 pp.  Principles of lake sedimentology. Springer-Verlag,  Hansen, M . J . 1984. Strategies for classification of landslides. In: D . Brunsden and D . B . Prior (eds.), Slope Instability. Wiley-Interscience, Chichester, pp 1-23. Hansen, W . R . 1965. Effects of the earthquake of March 27, 1964, Alaska. U . S . Geol. Surv. Prof. Paper N o . 5 4 2 - A . 68 pp.  at Anchorage,  Heber, R. 1849. Narrative of a journey through the Upper Provinces of India. 2nd. ed. John Murray, London. V o l . 1, p 274. Cited i n : Tucker 1982 (q.v.). Heuberger, H . , L . Masch, E. Preuss, and A Schrocker. 1984. Quaternary landslides and rock fusion i n Central Nepal and in the Tyrolean Alps. Mountain Res. and Dev., V o l . 4, N o . 4, pp 345-362. Hewitt, K . 1972. The mountain environment and geomorphic processes. In: H . O . Slaymaker and H . J . McPherson (eds.), Mountain Geomorphology. Tantalus Res., Vancouver, pp 17-34. . 1983. Seismic risk and mountain environments: the role of surface conditions i n earthquake disaster. Mountain Res. and Dev., V o l . 3, N o . 1, pp 27-44. Hickel, B. 1973. Limnological investigations in lakes i n the Pokhara Valley, Nepal. Int. Revue ges. H y d r o b i o l , V o l . 58, pp 659-672. Hitchcock, J.T. 143 pp.  1980.  A mountain village i n Nepal. Holt, Rinehart and Wilson. N . Y .  Hooker, J . D . 1854. Himalayan Journals; or, notes of a naturalist i n Bengal, the Sikkim and Nepal Himalaya, the Khasia Mountains. 2 vols. John Murray, London. Hormann, K . 1974. D i e Terrassen an der Seti-khola; E i n Beitrag zur Quartaren Morphogenese in Zentral Nepal. Erdkunde, B d . 28, Heff. 3, pp 161-175. Hudson, N . W . 1971.  Soil conservation. Cornell Univ. Press, Ithaca, N . Y . 320 pp.  Hutchinson, J . N . 1968. Mass movement In: R . W . Fairbridge (ed.), Encyclopaedia of Earth Sciences. Reinhold, N . Y . pp 688-695. Impat, P. 1980. Phewa Watershed: soil survey interpretation and land capability. Integrated Watershed Management Project, D e p t of Soil Conservation and Watershed Management, M i n . of Forests, Kathmandu. Phewa Tal Tech. Rep. N o . 14. 87 pp. . 1981. Hydrometeorology and sediment data for Phewa Watershed: 1979 data. Integrated Watershed Management Project, Dept. of Soil Conservation and Watershed Management, Ministry of Forests, Kathmandu. Phewa Tal Tech. Rep. N o . 15. 29 pp. Ives, J . D . 1981. Applied mountain geoecology: can the scientist assist i n the preservation of the mountains? In: J.S. Lall (ed.), and A . D . Moddie, The  145  Himalayan Environment:  Aspects of Change. O U P , D e l h i , pp 377-402.  Ives, J.D. 1984. Preparation of a position paper on the Himalayan/Ganges problem in the context of peace and resource-use conflict resolution. U . N . U . H i g h l a n d - L o w l a n d Interactive Systems Proposal. Ives, J.D., and R . G . Barry (eds.). 1974. London. 999 pp.  Arctic and alpine environments.  Methuen,  Ives, J.D., and B. Messerli. 1981. Mountain hazards mapping i n Nepal: introduction to an applied research project Mountain Res. and Dev., V o l . 1, N o . 3-4, pp 223-230. . 1984. Stability and instability of mountain ecosystems: lessons learned and recommendations for the future. Mountain Res. and Dev., V o l . 4, N o . 1, pp 63-71. Iwata, S., T. Sharma, and H . Yamanaka. 1984. A preliminary report of Central Nepal and Himalayan uplift J. N e p . G e o l . Soc, V o l . 4, pp 141-149. I . W . M . 1979. Base map of the Phewa Tal Watershed; 1:25,000. Integrated Watershed Management Proj., D e p t of Soil Conserv. and Watershed Management M i n . of Forests, Kathmandu. Jahn, R., C . Kellner and H . Stamm. 1979. Auswirkungen der Landnutzung i n Nepal bezueglich der Bodenerosion und Erfassung der Faktoren zur Landnutzungsplanung unter Beruecksichtigung der entwicklungspolitischen Problemstellung (Effects of land use i n N e p a l , particularly consideration of erosion, parameters for land use planning and development policy). U n p u b . D i p l o m a thesis, D e p t of Landscape Planning, Tech. Univ. of Berlin. Cited i n : Mohns, 1981 (q.v.). Jennings, A . H . 1950. World's greatest observed point rainfalls. Monthly Weather V o l . 78, pp 4 - 5 . Cited i n : Caine 1980 (g.v.). John,  Rev.,  K . W . 1968. Geographical stability analysis of slopes in jointed rock. J. Soil Mech. and Foundation D i v . , Proc. of the Amer. Soc. C i v . Eng. S M 2 , pp 498-521. Cited i n : Wagner 1983 (q.v.).  Johnson, K . , E . A . Olson, and S. Manandhar. 1982. Environmental knowledge and response to natural hazards i n mountainous Nepal. Mountain Res. and Dev., V o l . 2, N o . 2, pp 175-188. Jones, R . L . , and S.K. Jones. 1976. The Himalayan W o m a n : a study of L i m b u women in marriage and divorce. M a y f i e l d Pub. C o . Palo Alto, C a l . 155 pp. Junor, R.S. 1981. Impact of erosion on human activities, Bagmati catchment Nepal. J. Soil Cons. Service of N . S . W . , V o l . 37, N o . 1, pp 41-50. Kandel, G . P . 1978. Report on suspended sediments i n Kathmandu Valley. Integrated Watershed Management Project D e p t of Soil Conservation and Watershed Management Ministry of Forests, Kathmandu. 17 pp. Karunakaran, C . 1975.  The importance  of landslide studies. In: Proc. of a Seminar  on  146  Landslides and Toe Erosion Problems with Special Reference to Himalayan Region, Gangtok, Sikkim, 24-26 Feb. 1975. Indian Soc. of Eng. G e o l , Calcutta, pp v i i i - i x . Kelsey, H . M . 1978. Earthflows i n Franciscan melange, California. Geology, V o l . 6, pp 361-364.  V a n Duzen River basin,  . 1982. Influences of magnitude, frequency, and persistence of various types of disturbance on geomorphic form and process. In: F.J. Swanson, R.J. Janda, T. Dunne, D . N . Swanston (tech. eds.), Sediment Budgets and Routing in Forested Drainage Basins: Workshop Proceedings. U . S . D . A . For. Serv. Gen. Tech. Rep. P N W - 1 4 1 . Portland, Oregon, pp 150-153 Kenney, T . C . 1975. Weathering and changes i n strength as related to landslides. In: E. Yatsu, A . J . Ward and F . Adams (eds.), Mass Wasting: Proceedings of the Fourth Guelph Symposium on Geomorphology. G e o Abstracts, Norwich, pp 69-78. Khosla, A . N . 1953. Silting of reservoirs. Central. Board of Irrig. and Power of India, Pub. N o . 51. 203 pp. Cited i n : Menard 1961 (q.v.). Kienholz, H . , H . Hafner, S. Schnieder, and R. Tamrakar. 1983. Mountain hazards mapping in Nepal's M i d d l e Mountains: maps of land use and geomorphic damages (Kathmandu-Kakani area). Mountain Res. and Dev., V o l . 3, N o . 3, pp 195-220. Kienholz, H . , H . Hafner, and G . Schnieder. 1984. Stability, instability and conditional instability; mountain ecosystem concepts based on a field survey of the Kakani area i n the Middle Hills of Nepal. Mountain Res. and Dev., V o l . 4, N o . 1, pp 55-62. Kienholz, H . , S. Schnieder, M . Bichsel, M . Grunder, and P . K . M o o l . 1984. Mapping of mountain hazards and slope stability. Mountain Res. and Dev., V o l . 4, N o . 3, pp 247-266. K i r k b y , M . J . 1978. The stream head as a significant geomorphic threshold. Dept. of Geog. Working Pap. N o . 216, Univ. of Leeds. Cited in Brunsden et al. 1981 (q.v.).  Kojan, E. 1978. Engineering geological evaluation of landslide problems: Western Road Project, Godavari to Dandeldhura, N e p a l . U.S.A.I.D., Kathmandu.  Hills  Kollmannsperger, F . 1978/79. Long range landscape changes under the influence man. J. N e p . Research Centre, V o l . 2/3 (Sciences), pp 11-31.  of  Krishnan, M . S . 1960.  Geology of India and Burma. Higginbotham, Madras. 604 pp.  Laban, P. 1978. Field measurements on erosion and sedimentation i n Nepal. Integrated Watershed Management Project, Dept. of Soil Conservation and Watershed Management, Ministry of Forests, Kathmandu. Working Paper N o . 5. 30 pp. . 1978/1979. A preliminary appraisal of land mass movements J. N e p . Research Centre, V o l . 2/3 (Sciences), pp 97-109.  in Nepal.  147 . 1979. Landslide occurrence i n Nepal. Integrated Watershed Management Project, Dept. of Soil Conservation and Watershed Management, Ministry o f Forests, Kathmandu. Working Pap. N o . 13. 27 pp. L a l l , J.S. (ed.), and A . D . Moddie. 1981. The Himalaya: aspects o f change. O . U . P . , Delhi. 481 pp. L e Fort, P. 1975. Himalayas: the collided range; present knowledge of the continental arc. Amer. J. Sci., V o l . 275A ("Tectonics and Mountain Ranges"), p p 1-44. Leopold, L . B . 1970. A n improved method for size distribution of stream Water Resources Res., V o l . 6, N o . 5, pp 1357-1366.  bed gravel.  Leopold, L . B . , M . G . Wolman, and J.P. M i l l e r . 1964. Fluvial processes i n geomorphology. W . H . Freeman and Co., San Francisco. 522 pp. Levenson, B. 1979. Fuelwood utilisation: a study of the demand and available fuelwood resources at six selected villages. Integrated Watershed Management Project, Dept. of Soil Conservation and Watershed Management, M i n . of Forests, Kathmandu. Phewa T a l Tech. Rep. N o . 9. 65 pp. Low, B . L . (ed.). 1968. Mountains and rivers of India. Prep, for 21st Internat Congress, India. Cited i n : Brunsden et al. 1981 (q.v.).  Geogr.  L . R . M . P . 1983. Draft Land Systems Report. Land Resource Mapping Project, C . I . D . A . / H . M . G . , Kathmandu. 234 pp. MacFarlane,. A . 1976. Resources and population: a study of the Gurungs of Nepal. C . U . P . , Cambridge. 364 pp. M c L e a n , R . F . , and C . F . Davidson. 1968. The role of mass movement i n shore platform development along the Gisborne coastline, N . Z . Earth Sci. J., V o l . 2, 15-25. Cited i n : Crozier 1973 (q.v.). Meade, R . H . 1969. Errors i n using modern streamload data to estimate natural rates of denudation. G e o l . Soc. Amer. B u l l , V o l . 80. pp 1265-1274. Mehra, S.R., and T . K . Natarajan (eds.). 1966. Handbook on landslide analysis and correction. Central Road Res. Inst, N e w D e l h i . 114 pp. Menard, H . W . 1961. Some rates of regional erosion. J. of Geol., V o l . 69, pp 154-161. Messerschmidt D . A . 1976. The Gurungs of N e p a l : conflict and change i n a village society. Aris and Phillips L t d . , Warminster, England. 151 pp. Meybeck, M . 1976. Total mineral dissolved transport by world major rivers. H y d r o l . Sci. B u l l . N o . 21, pp 265-284. Cited i n : Ferguson 1984 (q.v.). M i l l e r , H . G . 1984. Dynamics of nutrient cycling i n plantation ecosystems. In: G . D . Bowen and E K . S . Nambiar (eds.), Nutrition of Plantation Forests. Academic Press, London, pp 53-118. Mohns, B . 1981. Agroforestry practices  for improving degraded mountain  ecosystems i n  148  Nepai. Unpub. M.Sc. thesis, Dept. of Earth Resources, C o l . State Univ. Mukerji, A . B . 1975. Geomorphological study of the choe terraces of the Chandigarh Siwalik Hills, India. Himalayan Ecology, V o l . 5, pp 303-328. Mulder, R.P. 1978a. A soil survey of Phewa Tal catchment Integrated Watershed Management Project D e p t of Soil Conservation and Watershed Management, M i n . of Forests, Kathmandu. Phewa Tal Tech. Rep. N o . 4. 57 pp. . 1978b. Erosion plot measurements i n Phewa Tal catchment Integrated Watershed Management Project D e p t of Soil Conservation and Watershed Management Ministry of Forests, Kathmandu. Phewa Tal Tech. Rep. N o . 6. 11 pp. Nayava, J.L. 1974.  Heavy monsoon rainfall in Nepal. Weather, V o l . 29, pp 443-450.  Nelson, D . , P. Laban, B . D . Shrestha, and G . P . Kandel. 1980. A reconnaissance inventory of the major ecological land units and their watershed condition in Nepal. Integrated Watershed Management Project D e p t of Soil Conservation and Watershed Management M i n . of Forests, Kathmandu. Working Paper N o . 17. 292 pp. N i p p o n Koei C o . L t d . 1976a. Phewa D a m Reconstruction Project Report H M G / U N D P , Kathmandu. . 1976b. Phewa D a m Reconstruction Project H M G / U N D P , Kathmandu.  Field Investigation  Study and Design Report  Novak, M . D . , and L.J.P. van V l i e t 1983. Degradation effects of soil erosion by water and wind. In: Soil Degradation in British Columbia. Proc. 8th Annual B . C . Soil Sci. Workshop. M i n . of A g . and Food, Victoria, B . C . pp 46-70. Oilier, C . 1984.  Weathering. 2nd ed. Longman, N . Y . 270 pp.  O'Loughlin, C . L . 1972. A n investigation of the stability of the steepland forest soils in the Coast Mountains, southwest British Columbia. Unpub. P h . D . thesis. Univ. of B r i t C o l . , Vancouver. 147 pp. . 1984a. Effectiveness of introduced forest vegetation for protection against landslides and erosion i n New Zealand's steeplands. In: In: C . L . O'Loughlin and A . J . Pearce (eds.), Symposium on Effects of Forest Land Use on Erosion and Slope Stability. I U F R O / N . Z . For. S e r v . / U S D A For. Serv./East-West Center, Honolulu, Hawaii, pp 275-280. . 1984b. Tree roots and their influence on slope stability. In: Asia Pacific Forest Hydrology Newsletter N o . 6, N o v . , East/West Center, Honolulu. O'Loughlin, C . L . , and R . R . Ziemer. 1982. The importance of tree root strength and deterioration rates upon edaphic stability in steepland forests. In: R . H . Waring (ed.), Carbon uptake and allocation: a key to management of subalpine ecosystems. Corvallis, Oregon, pp 84-91. Pal, S.K., and K . Bagchi. Recurrence  of floods i n Brahmaputra and Kosi Basins: a  149  study i n climatic geomorphology. Geogr. Rev. India, V o l 37, N o . 3, pp 242-248. Cited i n : Brunsden et al. 1981 (q.v.). Pandey, A . N . , P . C . Pathak, and J.S. Singh. 1983. Water, sediment and nutrient movement in forested and non-forested catchments in Kumaun Himalaya. For. Ecol. and Management, V o l . 7, pp 19-29. Pandey, U . , and J. Singh. 1984a. Energy-flow relationships between agro- and forest ecosystems in Central Himalaya. Env. Conserv., V o l . 11, N o . 1, pp 45-53. . 1984b. Energetics of hill agroecosystems: a case study from central Himalaya. Agric. Systems, V o l . 13, pp 83-95. Pecher, A . 1978. Les mecanismes de la deformation d'un cisaillement intracontinental: l'exemple du chevauchement central Himalayen ( M . C . T . ) . These d'Etat, Grenoble, 300 pp. Cited i n : Fort and Freytet 1982 (q.v.). Pelinck, E., P . K . Manandhar, and R . H . Gecolea. 1984. Forestry extension: community development i n Nepal. Unasylva, V o l . 36, N o . 1, pp 2-12. Peltier, L . 1950. The geographical cycle i n periglacial regions as it is related to climatic geomorphology. A n n . Assoc. Amer. Geog. V o l . 40, pp 214-236. Cited in: Land Resources Mapping Project 1983 (q.v.). Penck, A . 1894. Morphologie der Erdoberflache (2 vols.), pp 471 and 696. Hansen, M . J . , 1984 (q.v.). Powell, J.W. 1876.  Report on the geology of the eastern portion of the  Mountains. U.S. Geol. and Geog. Survey Terr. 218  Cited i n :  Uinta  pp. Cited i n : Price  1981  (q.v.).  Prasad, R . C . 1975. The landslides and Kosi Catchment In: Proceedings Problems with Special Reference Feb., 1975. Indian Soc. of Eng. Price, L . W . 1981.  erosion problems with special reference to the of a Seminar on Landslides and Toe Erosion to Himalayan Region. Gangtok, Sikkim, 24-26 G e o l , Calcutta, pp 73-83.  Mountains and man. Univ. of C a l . Press, Berkeley. 506 pp.  Rausch, D . L . , and H . G . Heinemann. 1984. Measurement of reservoir sedimentation. In: R . F . Hadley and D . E . Walling (eds.), Erosion and Sediment Y i e l d : Some Methods of Measurement and Modelling. G e o Books, Norwich, pp 179-200. Reineck, H . E . , and L B . Singh. 1973. Depositional sedimentary environments; reference to terrigenous elastics. Springer-Verlag, N . Y . 439 pp.  with  Remy, J . M . 1975. Geologie du Nepal: Ouest du Nepal Himalaya: Carte Geologique du Nepal au 1:506,880. C . N . R . S . , Paris. Rice, R . H . , E.S. Corbett, and R . G . Bailey. 1969. Soil slips related to vegetation, topography, and soils i n Southern California. W a t Res. Res., V o l . 5, pp 647-659. Rieger, H . C . 1975. Himalaya Wasser, Literaturanalyse ttber die Frage der Austwirkungen von Entwaldung, Erosion und sonstigen Stbrungen i m Einzugsgebiet des Ganges  150  und des Brahmaputra. Heidelberg. . 1976. Floods and droughts: the Himalaya and the Ganges plain as an ecological system. In: Mountain Environment and Development, Swiss Assoc. for Technical Assistance i n Nepal, Kathmandu. pp 13-30. . 1978/79. Socio-economic aspects of environmental degradation i n the Himalayas. J. N e p . Research Centre, Kathmandu, V o l . 2/3 (Sciences), pp 177-184. Robbe, E . 1954. Report to the Government o f Nepal on Forestry. E T A P Report N o . 209, U . N . F . A . O . , Rome. Cited i n : Eckholm 1976 (q.v.). Roehl, J.W. 1962. Sediment source areas, delivery ratios, and influencing morphological factors. In: I.A.H.S. Pub. 59. pp 202-231. Rose, L . E . , and J.T. Scholz. 1980. Nepal: profile of a Himalayan Kingdom. Westview Press, Boulder, C o l . 144 pp. Russell, J.R. 1981. Sedimentation i n the proposed Kotmale Reservoir, Sri Lanka. In: Proc. of the S.E. Asian Regional Symp. on Problems of Soil Erosion and Sedimentation. T. Tingsanchali and H . Eggers (eds.). Asian Inst, o f Technology, Bangkok, pp 405-418. Ruxton, B.P., and I. McDougall. 1967. Denudation rates i n northeast Papua from potassium-argon dating of lavas. Amer. J. Sci. 265, pp 545-561. Cited i n : Price, 1981 (q.v.). Sastry, G . , and D . Narayana. 1984. Watershed responses to conservation measures. J. Irrig. and Drainage Eng., V o l . 110, N o . 1, A S C E . pp 14-21. Saunders, I., and A . Young. 1983. Rates of surface processes on slopes, slope retreat and denudation. Earth Surf. Proc. and Landforms, V o l . 8, pp 473-501. Schumm, S.A. 1963. The disparity between present rates of denudation and orogeny. U.S. G e o l . Survey Prof. Pap. N o . 4 5 4 - H . 13 pp. . 1979. Geomorphic thresholds: the concept and its applications. Trans. Inst B r i t Geogr., V o l . 4, pp 485-515. Schwab, J . W . 1983. Mass wasting: October-November 1978 storm, Rennell Sound, Queen Charlotte Islands, British Columbia. M i n . of Forests, Brit. C o l . , Res. Note N o . 91. 23 p p . Scoullar, B.B. 1980. Agricultural extension baseline survey and survey methodology: Phewa T a l watershed area. D i v . o f Extension and Training, Dept. of A g r i c , Kathmandu. Selby, M . J . 1967. Aspects of the geomorphology of the Greywacke Ranges bordering the lower and middle Waikato Basins. Earth Sci. J., V o l . 1, pp 37-58. Cited i n : Crozier 1973 (q.v.). Seshadri, T . N . 1960. Report on Kosi River and its behaviour. Investigation and Research, D n . Khosi Project, Birpur. Cited i n : Brunsden et al. 1981 (q.v.).  151  Shah, P.B., and H . Schreier. 1985. Agricultural land evaluation for national land-use planning i n N e p a l : a case study i n the Kailali District Mountain Res. and Dev., V o l . 5, N o . 2, pp 151-162. Shah, S.G. 1980. Phewa Tal Watershed: animal husbandry and feed resource survey: result and recommendations. Integrated Watershed Management Project D e p t of Soil Conservation and Watershed Management M i n . of Forests, Kathmandu. 130 pp. Shakya, K . M . 1982. A mathematical programming model to determine the optimum land use for minimum soil loss i n a Nepalese hill agro-ecosystem. Unpub. M.Sc. thesis, N . C . State Univ. Sharma, C . K . 1973. 164 pp.  93 pp.  . 1974.  Geology of Nepal. Educational Enterprises  (P)  L t d . , Kathmandu.  Landslides and soil erosion i n Nepal. M r s . S. Sharma, Kathmandu.  . 1975. Natural resources of the Pokhara Valley. Navana Printing Works (P) Ltd., Calcutta. 106 pp. Sharma, T., S.S. M e r h , and N . W . Vashi. 1978. The terraced conglomerates of Pokhara Valley i n central Nepal. Proc. Symp. on Morphology and Evolution of Landforms. D e p t of G e o l . , U n i v . of Delhi, pp 226-240. cited i n : Yamanaka et al. 1982 (?.v.). Shimokawa, E. 1984. A natural recovery process of vegetation on landslide scars and landslide periodicity i n forested drainage basins. In: C . L O ' L o u g h l i n and A . J . Pearce (eds.), Proc. Symp. on Effects of Forest Land Use on Erosion and Slope Stability. I . U . F . R . O . / N . Z . For. S e r v . / U . S . D . A . For. Serv./East-West Center. Simonett, D.S. 1967. Landslide distribution and earthquakes i n the Bewani and Torricelli Mountains, N e w Guinea. In: J . N . Jennings and J.A. Mabbutt (eds.), Landform Studies form Australia and New Guinea. Australian N a t Univ. Press, Canberra, pp 64-84. Singh, J.S., A . N . Pandey, and P . C . Pathak. 1983. A hypothesis to account for the major pathway of soil loss from Himalaya. Env. Conserv., V o l . 10, N o . 4, pp 343-345. Singh, J.S., U . Pandey, and A . K . Tiwari. 1984. M a n and forests: a Central Himalayan case study. A m b i o , V o l . 13, N o . 2, pp 80-87. Skempton, A . W . 1953. Soil mechanics i n relation to geology. Proc. Yorks. G e o l . Soc., V o l . 29, pp 33-62. Cited i n : Crozier 1973 (q.v.). Skempton, A . W . , and J. Hutchinson. 1969. Stability of natural slopes and embankmemt foundations: a state-of-the-art report Proc. 7th Int Conf. on Soil M e c h . and Foundation Eng., Mexico City (State of the A r t Volume), pp 291-340. Slaymaker, H . O . , and H . J . McPherson (eds.). 1972. Research, Vancouver, B . C . 274 pp.  Mountain geomorphology.  Tantalus  152  Soil Survey Staff. 1975. Soil taxonomy. U . S . D . A . S.C.S. A g . Handbook N o . 436. Washington, D . C . 754 pp. Stainton, J . D . A . 1972.  Forests  of Nepal. John Murray, London. 174 pp.  Starkel, L . 1970. Cause and effects of heavy rainfall i n Darjeeling and in the Sikkim Himalayas. J. Bombay N a t H i s t Soc. V o l . 67, N o . 1, pp 45-50. . 1972a. The role of catastrophic rainfall in the shaping of the relief of the Lower Himalaya (Darjeeling Hills). Geogr. Pol. 21, pp 103-147. . 1972b. The modelling of monsoon areas of India as related catastrophic rainfall. Geogr. P o l . 23, pp 151-173. . 1976. The role of extreme (catastrophic) contemporary evolution of slopes.  meteorological  to  events i n  Sthaphit K . M . 1982. A land use-land capability classification system for Nepal: a case study i n Phewa Lake Watershed. Unpub. M.Sc. thesis, N . C . State U n i v . Stocklin, J. 1980. Geology of Nepal and its regional frame. J. G e o l . Soc. London, V o l . 37, pp 1-34. Strahler, A . N . 1952. Hypsometric (area-altitude) Soc. Amer. Bull. N o . 63. pp 1117-1142.  analysis of erosional topography. G e o l .  Strakhov, N . M . 1967. Principles of lithogenesis. V o l . 1 J.P.F. Fitzsimmons (trans.), S.I. Tompkieff and J.E. Hemingway (eds.). Oliver and Boyd, Edinburgh and London. 245 pp. Swanson, F.J., and C.T. Dyrness. 1975. Impact of clear-cutting and road construction on soil erosion by landsliding in the Western Cascade Range, Oregon. Geology, V o l . 3, pp 393-396. Swanson, F.J., R.J. Janda, T. Dunne, and D . N . Swanston (tech. eds.). 1982. Sediment Budgets and Routing in Forested Drainage Basins: Workshop Proceedings. U . S . D . A . For. Serv. Gen. Tech. Rep. P N W - 1 4 1 . Portland, Oregon. 165 pp. Swanson, F.J., and D . N . Swanston. 1977. Complex mass-movement terrains i n the western Cascade Range; Oregon. In: D . R . Coates (ed.). Landslides. Rev. i n Eng. G e o l . , V o l . III., pp 113-124. G e o l . Soc. of Amer., Boulder, Colorado. Tamm, C O . 1979. Nutrient cycling and productivity of forest ecosystems. In: Proc. Symp. on Impact of Intensive Harvesting on Forest Nutrient Cycling. State Univ. N . Y . , Syracuse, pp 2-21. Taylor, D . W . 1948.  Fundamentals of soil mechanics.  J. Wiley , N . Y .  Ter-Stepanian, G . 1965. In-situ determination of the Theological characteristics of soils on slopes. Proc. 6th Iht Conf. on Soil M e c h . and Foundation Eng., V o l . 2, pp 575-577. Cited i n : Carson 1976 ((q.v.). Terzaghi, K . 1950.  Mechanism of landslides. Eng. G e o l . (Berkeley  Vol.). pp 83-123.  153  Thompson, M . , and M . Warburton. 1985. Uncertainty on a Himalayan scale. Mountain Res. and Dev. V o l . 5, N o . 2, pp 115-135. Thornes, J.B. 1980. Erosional processes of running water and their spatial and temporal controls: a theoretical viewpoint In: M . J . K i r k b y and R . P . C . Morgan (eds.), Soil Erosion. Wiley-Interscience. pp 129-182. Thouret J.C. 1977. Contribution a l'etude de revolution des grands versants de la montagne himalayenne: l'exemple du bassin versant de l ' A n k h u Khola, Nepal Central. These 3eme Cycle, Paris-Pantheon-Sorbonne. 419 pp. Cited i n : Fort and Freytet 1982 (q.v.). . 1981a. Geodynamique des Grands Versants de l ' A n k h u Khola, Nepal Central. C . N . R . S . . 1981b. Carte Geomorphodynamique du Bassin Versant de l'Ankhu Nepal Central. 1:75,000. C . N . R . S . , Paris.  Khola,  Toky, O.P., and P.S. Ramakrishnan. 1982. A comparative study of the energy budget of hill agro-ecosystems with emphasis on the slash and burn system ijhum) at lower elevations of north-eastern India. Agric. Systems, V o l . 9, pp 143-154. Topographical Survey Branch. 1984a. Geological M a p , 1:125,000: Nepal, Western Development Region. No's 62 P - D , 71 D - C , and 72 A - A . 1st ed. Topog. Surv. Branch, H . M . G . Surv. D e p t , Kathmandu. . 1984b. Land Systems M a p , 1:50,000: Nepal, Western Development Region. No's 62 P/15, 62 P/16. Topog. Surv. Branch, H . M . G . Surv. D e p t , Kathmandu. . 1984c. Land Utilization M a p , 1:50,000. Nepal, Western Development Region. No's 62 P/15, 62 P/16. Topog. Surv. Branch, H . M . G . Surv. D e p t , Kathmandu. Trimble, S.W. 1977. The fallacy of stream equilibrium in contemporary studies. Amer. J. Sci., V o l . 277. pp 876-887.  denudation  Trustrum, N . Z . , V . J . Thomas, and G . B . Douglas. 1984. The impact of forest removal and subsequent mass-wasting on hill land pasture productivity. In: In: C . L . O'Loughlin and A . J . Pearce (eds.), Symposium on Effects of Forest Land Use on Erosion and Slope Stability. I U F R O / N . Z . For. S e r v . / U S D A For. Serv./East-West Center, Honolulu, Hawaii, p. 308. Tucker, R.P. 1982. The forests of the Western Himalayas: the legacy of British Colonial Administration. J. For. H i s t V o l . 26, N o . 3, pp 112-123. . 1983. The British Colonial System and the forests of the Western Himalayas, 1815-1914. In: R . P . Tucker and J.F. Richards (eds.), Global Deforestation and the Nineteenth Century W o r l d Economy. Duke Press Policy Studies, Duke, N . C . pp 146-166. U . N . D . P . 1966. Feasibility report on the Chisapani H i g h D a m Project Dev. Programme, Kathmandu.  United Nations  154  U . N . F . A . O . 1965. Soil erosion by water: some measures for its control on cultivated lands. U . N . F . A . O . Agric. Dev. Pap. 81. Rome. 284 pp. Upadhyay, K . P . 1977. Contour trenching as a strategy in watershed rehabilitation: application to Nepalese conditions. Unpub. M.Sc. thesis, Utah State Univ., Logan, Utah. Van de Putte, R . A . 1979. Summary of findings of four surveys in Phewa Tal watershed area, Kaski District H i l l Agric. Dev. Proj., U N F A O / U N D P , Kathmandu. Cited i n : Shah 1980 (q.v.). Varnes, D . J . , 1958. Landslide types and processes. In: E.B. Eckel (ed.), Landslides and Engineering Practice. Highways Res. Board, Spec. Rep. 29, pp 20-47. . 1978. Slope movement types and processes. In: R . L . Schuster and R.J. Krizek (eds.). Landslides: analysis and control. Transp. Res. Board, N a t Acad. Sci., Spec. rep. N o . 176. pp 12-33. Wash., D . C . Wadia, D . N . 1957.  Geology of India. M a c M i l l a n , London. 3rd edt  536  pp.  Wagner, A . 1981. Rock structure and slope stability study of Walling area, central Nepal. J. Nep. G e o l . Soc., V o l . 1, N o . 2, pp 37-43.  west  . 1983. The principle geological factors leading to landslides in the foothills of Nepal: a statistical study of 100 landslides; steps for mapping the risks of landslides. Helvetas/ Swiss Technical Cooperation/ITECO, Kathmandu. 58 pp. Walling, D . E . 1977. Limitation of the rating curve technique for estimating suspended sediment loads, with particular reference to British rivers. In: Erosion and Solid Matter Transport i n Inland Waters: Symposium. I.A.H.S. Pub. N o . 122. pp 34-48. Walling, D . E . , and B . W . Webb. 1981. The reliability of suspended sediment load data. In: Erosion and Sediment Transport Measurement: Symposium. I.A.H.S. Pub. N o . 133. pp 177-194. Ward, P.R.B. 1984. Measurement of sediment yields. In: R . F . Hadley and D . E . Walling (eds.), Erosion and Sediment Y i e l d : Some Methods of Measurement and Modelling. G e o Books, Norwich, pp 37-70. Warhaftig, C . 1970. The trouble with the rate of erosion with Programs, V o l . 2. p 157. Water  [abs.]. G e o l . Soc. Amer. Abs.  and Energy Commission. 1982. Hydrological studies of Nepal, V o l . 1: Report Canadian International Water and Energy Consultants/Min. of Water Resources, Kathmandu.  Weinberg, A . 1972. Science and trans-science. Minerva, V o l . 10, pp 209-222. Cited i n : Thompson and Warburton 1985 (q.v.). White, P . G . , M . Fort, and B. Shrestha. 1983. Geomorphic mapping of the Resource Conservation and Utilization Project Areas. Resource Conservation and Utilization Project D e p t of Soil Conservation and Watershed Management Ministry of  155  Forest and Soil Conservation/South-East D e v . / U . S . A . I . D . , Kathmandu. 92 pp.  Consortium for  International  Whiteman, P . T . C . 1985. A n agronomist's perspective with a case study from Jumla, Nepal. Mountain Res. and Dev., V o l . 5, N o . 2, pp 151-162. Whyte, R . O . 1957. The grassland and fodder resources of India. Ind. co. of Agric. Res., N e w Delhi. Cited i n : Fox 1982 (q.v.). Wiart, J. 1983. Ecosysteme villageois traditionnel en Himalaya Nepalais: la production forestiere suffit-elle aux besoins de la population? Unpub. P h . D . thesis, Univ. Scientifique et Medicale, Grenoble. 259 pp, app. Willan, R . G . M .  1967.  Forestry  in Nepal. U . N . D . P . ,  Kathmandu. Cited i n : Eckholm  1976  (q.v.).  Williams, V.S. 1977. Neotectonic implications of the alluvial record i n the Sapta Kosi drainage basin, Nepalese Himalayas. Unpub. P h . D . thesis. D e p t of Geology, Univ. of Washington, Seattle. 90 pp. Wilson, L . 1973. Variations in mean annual sediment yield as a function of annual precipitation. Amer. J. Sci. V o l . 273, pp 335-349. Wischmeier, W . H . 1976. Use and misuse of the U S L E . J. Soil and Water V o l . 31, N o . 1, pp 5-9. •  . 1984. The U S L E : N o . 2, pp 105-107.  some reflections.  mean  Conserv.,  J. Soil and Water Conserv. V o l . 39,  Wischmeier, W . H . , C . B . Johnson, and B . V . Cross. 1971. A soil credibility nomograph for farmland and construction sites. J. Soil and Water Conserv., V o l . 26, pp 189-193. Wischmeier, W . H . , and D . D . Smith. 1978. Predicting rainfall erosion losses - a guide for conservation planning. Agric. Handbook N o . 537, U . S . D . A . , Washington, D . C . W o r l d Bank. 1979. Nepal: development performance and prospects. A W o r l d Bank Country Study, South Asia Regional Office, W o r l d Bank. Washington, D . C . 123 pp. Wormald, T.J. 1976a. Meteorological and soil temperature records at Lumle Agriculture Centre. Lumle Agriculture Centre, Pokhara, Nepal. 8 pp. . 1976b. A n annotated Pokhara, Nepal.  checklist of plants. Lumle Agriculture Centre,  . 1976c. Village forestry i n the Hills of West Central Nepal. Lumle Agriculture Centre, Pokhara, Nepal. W u , T . H . , W . P . M c K i n n e l l in, and D . N . Swanston. 1979. Strength of tree roots and landslides on Prince of Wales Island, Alaska. Can. Geotech. J., V o l . 16, N o . 1, pp 19-33.  156  Wyatt-Snrith, J. 1982. The agricultural system i n the hills of Nepal: the ratio of agricultural to forest land and the problem of animal fodder. A . P . R . O . S . C . Occasional Papers: 1. Agric. Projects Services Centre, Kathmandu. 17 pp. Yamanaka, H . 1982. Radiocarbon ages of upper Quaternary deposit i n Central Nepal and their geomorphological significance, Science Reports of Tohoku University, Sendai, Japan, V o l . 32, N o . 1. pp 46-60. Yamanaka, H . , M . Yoshida, and K . Arrita. 1982. Terrace landform and Quaternary deposit around Pokhara Valley, Central Nepal. J. N e p . G e o l . Soc., V o l . 2, special issue, pp 113-142. Y o u n g , A . 1969.  Present rate of land erosion. Nature, V o l . 224,  pp 851-852.  Zeitler, P . K . , N . M . Johnson, C . W . Naeser, and R . A . K . Tahirkheli. 1982. Fission-track evidence for Quaternary uplift of the Nanga Parbat region, Pakistan. Nature, V o l . 298, pp 255-257. Ziemer, R.R. 1981. Roots and the stability of forested slopes. In: T . R . H . Davies and A . J . Pearce (eds.), Erosion and Sediment Transport i n Pacific R i m Steeplands: Symposium. I.A.H.S. Pub. N o . 132. pp 343-361. Zollinger, F . 1979a. The Sapt Kosi: unsolved problems of flood control in the Nepalese Terai. Integrated Watershed Management Project, Dept. of Soil Conservation and Watershed Management, M i n . of Forests, Kathmandu. Working Pap. N o . 11. 94 pp. . 1979b. Analysis of river problems and strategy for flood control i n the Nepalese Terai. Integrated Watershed Management Project, D e p t of Soil Conservation and Watershed Management M i n . of Forests, Kathmandu. Working Pap. N o . 12. 130 pp.  APPENDIX  The following table is a recorded i n the Durbasha composite figures from a and type of failure were defined.  1. L A N D S L I D E D A T A  FROM  PRASAD  (1975)  facsimile of Table 1 in Prasad (1975), and shows events watershed near Chatra in east Nepal. The rainfall data are number of gauges at different points i n the watershed. Size not reported, and earthquake size classes were not further  TABLE—1. SHOWING  T H EM E A N  CRITERIA  O F PRECIPITATION,  DAYS, EARTHQUAKE SI. No.  Month  INTENSITY, N O .O F  A N DLANDSLIPS.  Average rainfall intcn->ity for 30 minute in mm/hr.  Average prccipitaiion mm.  Average no. of rainy days.  , Mild  RAINY  ( 1963-72 )  Average no. of earthquake Average — * Total. Feeble Slight Moderate  Total Averge N o . of N o . of Landsli- Landslidel. det.  1.  January.  7.5  12.8  1  3  8  3  —  14  —  —  2.  February.  6.S  7.2  1  4  4  4  3.  March  122  23.7  2  4  6  3  4.  April.  30.4  61.2  4  3  5  2  -  12  10  -  -  1  15  23  2.3  14  26  2.6  9  1  0.1  3.  May.  45.9  123.2  6  6  6  1  6.  June.  65.4  362.8  13  12  4  1  7.  July.  68.0  715.3  20  10  3  1  8.  August  58.0  429.7  17  8  5  1  -  13  13 17  September.  62.1  317.0  14  6  2  •  10.  October.  43.9  134.6  5  4  5  1  11.  November.  6.2  11.9  1  2  4  3  1  10  12.  December.  1.6  1.7  -  3  4  2  -  9  84  65  56  23  2  9.  Total.  2201.1  10  146  50  -  -  5.0  APPENDIX  The  following  2. E R O S I O N  PLOT  DATA,  1979, P H E W A  VALLEY  tables show:  A ) Monthly precipitation, runoff, and soil loss values for 1979 for four 10 m erosion plots at Banpale, near Naudanda, Phewa Valley, Nepal, as reported by Impat (1981). Site conditions are summarised in Table 2 (p. 22). 2  B) Precipitation, runoff, and soil loss values for 1979 for a single 10 m plot i n dense forest at Tamagi, Phewa Valley, Nepal, as reported by Impat (1981). Site conditions are summarised i n Table 2 (p. 22). 2  A PROTECTED PASTURE MIXED WITH FOREST MONTH  PLOT 1  RAINFALL (mm)  Jan Feb march Apr l1 May June July Aug Sept Oct Nov Dec  7.0 N D N D 83.5 154 9 564 . 7 1070.0 1285.5 644 . 5 285.0 30.0 0.0  TOTAL  4125.0  Runoff (11tre )  370 715 2565 1465 595  PLOT 3  Soil loss (t/ha)  O.1985 0.1904 0.2734 0. 1770 0.1662  57 10  OVERGRAZED LAND  1.0055  Runoff (1ttre)  335 385 765 7O0 130  2315  Sol 1 l o s s (t/ha)  0 . 3420 0. 1770 0 2553 0 .2294 0 0194  1 0231  PLOT 2 Runoff (1Itre)  730 3350 8445 3610 2030  3.8310 2.7340 3.2370 1.8043 0.3567  1B065  11.9630  N D  DATE (duration) 1-2 3-9 1 0 - 16 17-23 24 J u l - 6 7-20 21-27 28-30 31 A u g - 3 4-9 10 S e p t - 7  Jul Jul Jul Jul  y y y y  Aug Aug Aug Aug Sept Sept Oct  TOTAL  RUNOFF ( l i t r e )  5..2 73 . 8 231 . 3 274 .5  12 27 46  3843.1  4  78 36  0 .. 0 0 8 7 0 .. 0 2 5 9 0 .. 0 2 1 6 0 .. 0 7 7 2 0 . 1 185  33 364  0.4324  13 43  530 1820 4885 1945 995  10175  SOIL LOSS (t/ha)  0,. 0 5 9 0 0..0429 0. 0268 0. 0221 0. 0146 0 ..0151  65 7  PLOT 4 Runoff (11tre)  No d a t a  RAINFALL (mm)  682 .9 3 8 5 .4 525 . 8 4 8 .. 3 7 7 ,. 2 2 5 3 . .5 2 2 3 ., 7  S o i l loss (t/ha)  S o i l loss (t/ha)  2.2140 1.4020 1.7480 1.9819 0.3929  7.7388  159  Appendix 3. Sediment Load: Dominant Particle Size Many geomorphic and hydraulic problems involve particles sizes coarser than sand, and the samples required for the determination of particle size distribution are too large and heavy to be brought into the laboratory for weighing. Some quantitative expression of sediment size is needed to describe the material. The size of sediment particles is usually expressed as a distribution graph, showing the percentage by weight in the sample represented by each size class, determined by sieving through a nest of sieves (Dunne and Leopold 1978). A procedure described i n detail in Leopold (1970) allows the rapid, quantitative field assessment of the dominant particle size of material on a surface. This is the size class that represents the largest percentage of the total sample by weight, and it approximates the result that would have been obtained by sieving. It involves the measurment of 100 particles picked up at random from the surface under investigation. Briefly, the procedure is as follows: (1)  A relatively homogeneous  area is chosen as the sample site.  (2) The researcher walks over the site, and with eyes averted, reaches over the toe of his boot and touches whatever particle is there with an extended finger. The rock is picked up and its intermediate or b axis measured. The measurement is recorded i n m m as the lower limit of the size class into which it falls. Size classes vary by the square root of 2, so that the series progresses 2, 2.8, 4, 5.6, 8, 11 m m , etc. Material < 2 m m in diameter cannot be counted by this method, but its presence is recorded by an entry i n the < 2 m m class. (3) W h e n about 100 rocks have been measured, counting stops. The data are tabulated as a frequency against the lower limit of their size class. (4) Multiplication of the numbers in a size class by an average weight (determined experimentally in the laboratory; see Leopold 1970) gives a total weight for that size class. (5) Because large rocks present greater surface areas than small ones, they have a higher probability of being chosen, and therefore a correction is made by dividing the total weight for each class by the square of the mean diameter of the size class. These values are then transformed into percentages. (6) A final transformation to make these percentages independent of the particular sieve size used is achieved by dividing by the log of the diameter interval of the size categories. (7) These values are plotted on l o g - l o g paper against the geometric mean size of the interval, giving a curve of the percentage by weight/log sq. root of two against particle size in mm. The dominant particle size is determined as the asymptote of the plotted curve. The method is simpler to carry out once learned, but readers should consult the original reference (Leopold 1970) for a full description and a discussion of the assumptions involved.  160  A P P E N D I X 4. P A R T I C L E  DIMENSIONS  During the survey o f fluvial characteristics, the a, b, and c dimensions o f 20 o f the largest particles deposited by recent flows, and 10 particles judged not to have moved, were measured at each station, for possible use in empirical methods o f flow velocity estimation (see Caine and M o o l (1981) for application o f the technique i n Nepal). These values are given below.  STATION  PARTICLE P A R T I C L E S MOVED (mm) a  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15  42 54 46 34 53 38 43 47 39  b  DIMENSIONS P A R T I C L E S NOT MOVED (mm)  c  a  31 17 36 21 33 17 22 14 36 18 17 30 33 16 33 20 30 18 a l 1 moved  70 67 73 46 78 43 54 61 54  b 48 54 48 30 56 35 39 43 36 a l 1 moved  c 29 31 31 19 38 23 24 32 25  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items