Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Impact of prescribed burning on diffuse knapweed (centaurea diffusa) infestations and floral diversity… Nicholson, Andrew Ross 1992

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

Item Metadata

Download

Media
831-ubc_1992_spring_nicholson_andrew_ross.pdf [ 1.08MB ]
Metadata
JSON: 831-1.0086724.json
JSON-LD: 831-1.0086724-ld.json
RDF/XML (Pretty): 831-1.0086724-rdf.xml
RDF/JSON: 831-1.0086724-rdf.json
Turtle: 831-1.0086724-turtle.txt
N-Triples: 831-1.0086724-rdf-ntriples.txt
Original Record: 831-1.0086724-source.json
Full Text
831-1.0086724-fulltext.txt
Citation
831-1.0086724.ris

Full Text

IMPACT OF PRESCRIBED BURNING ON DIFFUSE KNAPWEED (Centaurea diffusa) INFESTATIONS AND FLORAL DIVERSITY IN KALAMALKA LAKE PROVINCIAL PARK  by Andrew Ross Nicholson B.A. The University of Saskatchewan, 1988 B.S.A. The University of Saskatchewan, 1988  A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE  in THE FACULTY OF GRADUATE STUDIES  (Department of Plant Science) We accept this thesis as conforming to the required standard  University of British Columbia June 1992 Andrew Ross Nicholson,  1992  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes ma, be granted by the head of my department  or  by  his  or  her  representatives.  It  is  understood  that  copying  or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department  of  The University of British Columbia Vancouver, Canada Date  DE-6 (2/88)  I(I  Ji,irt’  /  _)  /  Abs tract  This study investigated the infestation of Kalamalka Lake British Columbia by diffuse knapweed  Provincial Park, diffusa Lam.).  (Centaurea  Surveys of six different types of grassland  ecosystem association/seral stages for species richness and diversity in the spring of 1990 and 1991,  and for diffuse  knapweed density in the fall of 1990 and 1991 revealed that increasing diffuse knapweed density was not associated with reduced richness or lower diversity.  A multiple discriminant  analysis on the density of diffuse knapweed showed that the level of accessibility of a site was the most important factor in the spread of knapweed, and that the amount of bare soil may be an important co-factor.  A multiple discriminant analysis for  species richness revealed that time since the last herbicide treatment was the most important factor affecting species richness of a site. A prescribed burning experiment was conducted on two sites in Kalamalka Lake Provincial Park. diffuse knapweed. (P>0.05)  One site was infested with  Neither spring nor fall burning reduced  any class of knapweed  plants, or total knapweed).  (i.e.  seedlings, rosettes, bolted  The other site initially had no  knapweed, and neither burning treatment resulted in the establishment of any knapweed up to the end of the first growing season after the burns.  Although there were initial differences  in the richness and diversity of the two sites, neither burning treatment altered (>O.05)  the seasonal pattern or level of  ii  richness or diversity by the end of the study. Both spring and fall burning treatments reduced the amount of available fuel and increased the amount of bare soil.  The  =O.792) 2 increase in percent bare soil was strongly correlated (r to fire intensity. Management recommendations include continued monitoring of the burn sites to determine if the burning treatments result in changes in the density of knapweed or species richness over the next few years.  A moratorium on the use of herbicides is also  recommended in order to maintain species richness and diversity in the park.  iii  Table of Contents Abstract  ii  .  Table of Contents  iv  .  vi  List of Tables  viii  List of Figures  ix  Acknowledgements 1.0  INTRODUCTION  1  2.0  MATERIALS AND METHODS  7  2.1  Study Area  7  2.2  Field Methods  10  2.2.1  Diffuse knapweed survey  10  2.2.2  Richness and diversity study  12  2.2.3  Prescribed burn documentation  13  2.2.4  Impact of prescribed burning on the 15  density of diffuse knapweed 2.2.5  Impact of prescribed burning on species richness and floristic diversity.  2.3 3.0  .  .  .16 17  Statistical Analysis  22  RESULTS AND DISCUSSION 3.1  The Diffuse Knapweed Problem in Kalamalka Lake 22  Provincial Park 3.2  Species Richness and Floral Diversity in Kalamalka Lake Provincial Park  3.3  33  Characteristics of the Fall and Spring 44  Prescribed Burns  iv  3.4  Impact of Prescribed Burning on the Density 50  of Diffuse Knapweed 3.5  3.6 4.0  Impact of Prescribed Burning on Species Richness and Floral Diversity  58  Summary of Findings  63 65  MANAGEMENT IMPLICATIONS 4.1  Diffuse Knapweed in Kalamalka Lake 65  Provincial Park 4.2  Species Richness and Floral Diversity in 68  Kalamalka Lake Provincial Park 4.3  The Use of Prescribed Burning in 69  Kalamalka Lake Provincial Park 5.0  LITERATURE CITED  6.0  APPENDIX I  72  Location and type of diversity/knapweed survey transects in Kalamalka Lake Provincial Park 7.0  76  APPENDIX II Location of burn plots in Kalamalka Lake 77  Provincial Park 8.0  APPENDIX III Description of study sites used in diversity/knapweed 78  study 9.0  APPENDIX IV Higher Plants of Kalamalka Lake Provincial Park.  .79  10.0 APPENDIX IV Heat Yield Constant for Fire Intensity Equation.  v  .  .82  List of Tables  Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability (P) for a two factor analysis of variance randomized within two years and six ecosystem associations for four categories of diffuse knapweed density (seedlings, rosettes, bolted plants, and total knapweed) in Kalamalka Lake Provicial Park during early May 1990 and early May 24 1991  2  Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability (P) for a two factor analysis of variance randomized within two years and six ecosystem associations for four categories of diffuse knapweed density (seedlings, rosettes, bolted plants, and total knapweed) in Kalamalka Lake Provicial Park during late August 1990 and late August 25 1991  3  Canonical correlations and canonical loadings of a three group multiple discriminant Analysis for diffuse knapweed density in Kalamalka Lake Provincial Park during the fall 29 of 1990 and 1991  4  Validation matrix for the three group, 36 case multiple discriminant analysis of the density of diffuse knapweed during the spring and fall of 1990 and 1991 in Kalamalka Lake Provincial Park  30  Group classification constants and group classification coefficients for a three group multiple discriminant analysis of diffuse knapweed density in Kalamalka Lake Provincial Park during the fall of 1990 and 1991  32  5  6  Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability () for a two-way analysis of variance randomized within two years and six ecosystem associations for species richness and the Shannon—Weinner diversity index in Kalamalka Lake Provicial Park during ea1y May 1990 and 33 early May 1991  vi  7  Canonical correlations and canonical loadings of a three group multiple discriminant analysis for species richness in Kalamalka Lake Provincial Park during the spring of 1990 40 and 1991  8  Validation matrix for the three group, 36 case multiple discriminant analysis of species richness in Kalamalka Lake 41 Provincial Park during the spring of 1990 and 1991  9  Group classification constants and group classification coefficients for a three group multiple discriminant analysis of species richness in Kalamalka Lake Provincial 43 Park in the spring of 1990 and 1991  10  Air temperature (C), wind speed (Kph), and relative humidity (%) during the two sets of prescribed burning treatments in Kalamalka Lake Provincial Park on November 2 44 1990 and March 7 1991  11  ), 2 Fuel moisture (% oven dry weight), available fuel (g/m ), rate of spread (rn/sec), and fire 2 fuel consumed (g/m intensity (Kwatts/m) for two burning treatments at two sites 45 in Kalamalka Lake Provincial Park  12  Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability (P) for a two factor analysis of variance randomized within three burn treatments (fall, spring, and control) and four dates with three replicates per cel the density of four categories of diffuse knapweed (seedlings, rosettes, mature plants, and total knapweed) in Kalamalka 54 Lake Provicial Park during 1990 and May 1991  13  Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability (P) for a three factor analysis of variance randomized within two sites, three burn treatments (fall, spring, and control), and four dates with three replicates per cell for species richness and the Shannon—Weinner diversity index in Kalamalka Lake Provicial Park during 1990 60 and 1991  vii  List of Figures  Map of British Columbia showing location of Kalamalka Lake Provincial Park  8  2  Map of Kalamalka Lake Provincial Park  9  3  Density of Diffuse Knapweed in Kalamalka Lake Provincial Park  23  Richness and Diversity in Kalamalka Lake Provincial Park  35  Canopy and Soil Temperature Profile during Spring Burns  46  6  Changes in Percent Bare Soil  48  7  Impact of Prescribed Burning on the Density of Diffuse Knapweed  55  Impact of Prescribed Burning on Species Richness and Plant Diversity  61  Impact of Accessability and Percent Bare Soil on Diffuse Knapweed Density  67  Impact of Herbicide Treatment and Diffuse Knapweed on Species Richness  70  1  4  5  8  9  10  viii  Acknowledgements Pitt for providing  I wish to express my appreciation to Dr. M.D.  encouragement and guidance during the course of this study.  I  would also like to thank the members of my committee, Dr. N. Feller,  and Dr. M.K. Upadhyaya for their very constructive input.  Special thanks to Mr.  Roy Benson, Ministry of Forests,  advice and the supervision of the controlled burns; Mr. Goertzen and Mr. support;  Dr.  Bob Scheer both of B.C.  for his  Dave  Parks for their logistic  Bruce Lawson of Forestry Canada for the loan of the  fire monitoring equipment; Mr. Reg Newman, Ministry of Forests, his computer wizardry and technical advice, Wikeem, Ministry of Forests,  and to Dr.  for  Brian  for his thoughtful review of my  methodology. Financial support was provided by B.C. and B.C.  Conservation Foundation (Mr.  Parks  (Mr.  Bob Moody).  Don Gough),  The  accommodation that was provided during the summer field season by Mr.  Ted Osborn of the Coldstream Ranch was much appreciated. Finally,  I would like to thank my friends and family; without  their optimism to hearten and inspirit me I never would have been able to complete this project.  ix  1.0  INTRODUCTION  Diffuse and spotted knapweed C. maculosa Lam.) (B.C.).  (Centaurea diffusa Lam., and  pose a serious problem in British Columbia  In 1984 the value of losses in forage production due to  knapweed was estimated at $1.5 million (Cranston 1984).  Knapweed  infestation causes other less tangible losses: deterioration of wildlife habitat; increased grazing management costs; and, decreased property values.  Of these two knapweeds, diffuse is a  more serious problem in the semi—arid grassland areas of the interior of B.C.  (Watson and Renney 1974).  In 1984, diffuse  knapweed accounted for 85% of the land infested Hansen 1984),  (Hamlen and  totalling 69,000 ha in British Columbia.  Various  chemical treatments have been shown to be effective at controlling knapweed (Watson and Renney 1974, 1982, Lass and Callihan 1989).  Shelley and Roche  In a public park, however,  where  there is a concern to maintain floristic diversity, the use of herbicides may be unacceptable. At least three stands of diffuse knapweed have been identified within Kalamalka Lake Provincial Park (KLPP) Ministry of Lands, Parks, and Housing 1984).  (B.C.  Concerns have been  expressed that diffuse knapweed invasion has adversely affected the natural quality and aesthetic value of the park. example, Tyser and Key (1988),  For  in Glacier National Park, showed  that richness and frequency of native species were inversely related to spotted knapweed  stem densities. 1  Furthermore, there  is some concern that continued knapweed invasion may threaten the value of KLPP as an example of climax bluebunch wheatgrass (Acropyron spicatum (Pursh)  Scribn.  & Smith)  grassland in British  Columbia. In order to develop the most appropriate management policies,  it is necessary to properly document characteristics of  community structure. species richness  Two of the most commonly used measures are  (i.e., simply the number of different species)  and diversity (a measure that incorporates species richness and the evenness of their distribution)  (Barbour et al.  1980).  There  is a need, therefore, to measure and record the current state of species richness and floral diversity in KLPP,  and to identify  the factors most important in maintaining a high degree of species richness. Knapweeds are not very palatable such that livestock grazing is not a feasible control method  (Watson 1977).  livestock grazing is inappropriate, as KLPP  Furthermore,  policy is to  “encourage...the recovery of the natural species and processes in the park”  (Ministry of Lands Parks,  and Housing 1984)  Prescribed burning has been recommended to reduce fuel loading and the subsequent fire hazard in KLPP, now that livestock grazing is no longer appropriate.  Cawker (1983)  established that fire was an important element in the ecology of grasslands in the Okanagan Valley since deglaciation.  The  suppression, since European settlement, of this periodic burning may have altered historic grassland composition of this area, as 2  the native plants may be well adapted to this periodic burning. Prescribed burning offers not only a chance to re—introduce a “natural process,” but it may also be a promising management alternative for controlling knapweed. Unfortunately, there is a dearth of information concerning the effects of fire on the composition of grasslands in British Columbia (B.C.).  There is a great deal of information available  on very similar grassland communities in the northwestern United States.  Many of these studies have concentrated on using fire to  increase forage production, reduce the density of shrubs, and limit the encroachment of conifers onto grasslands al.  1987,  1984).  Clark and Starkey 1990, Gruell et al.  (Bunting et  1986, and Mathews  Other studies have tended to focus on the effects of fire  on a few important species rather than on species richness and floral diversity (Blaisdel 1953,  Conrad and Poulton 1966,  Daubemuire 1968, Daubenmire 1975, Uresk et al. et al. B.C.,  1988).  1976,  and  Patton  Similarly, Wilms and Bailey (1980), who worked in  examined only bluebunch wheatgrass.  Furthermore, the  papers which looked at the impact of fire on individual species have tended to not documented the characteristics of the burn, which precludes comparisons among results. Schwecke and Hann (1989)  found that fire could be used to  increase the species richness of an area which they studied in the intermountain region of Washington state. Wright  (1984)  Furthermore,  showed that most forbs tolerate fire well if burned  in the spring or fall.  He concluded (Wright 1984) 3  that  fire can  enhance the diversity of forbs because they have a hard seed that can be scarified by fire. composition is complex.  The effect of the fire on the species It will depend on the pre-existing  vegetation mosaic, and the site’s potential, as well as the intensity and the timing of the fire  (i.e.  spring vs fall).  A number of researchers have looked at using fire to control knapweed (Popova 1960,  Zednai 1968, Watson and Renney 1974,  Watson 1977, Chicoine and Fay 1984,  Shelley and Roche 1982).  Most of the published results are anecdotal.  The emphasis has  been placed on how much better chemical control is than burning (Chicoine and Fay 1984,  Shelley and Roche 1982).  No study has  investigated how much control burning gives compared to untreated areas  Even the two most detailed reports of using fire to  .  control knapweed are incomplete (Popova 1960, intensity,  Zednai 1968).  Fire  fuel load, weather conditions, and other important  factors were not documented.  Quantitative data on the pre— and  post-burn knapweed population densities are lacking in both reports. Popova  Furthermore,  (1960)  the results of these two studies conflict.  found that within two years of fall burning in a  mixed diffuse knapweed/grass stand, diffuse knapweed disappeared and was replaced by a luxuriant sward of grass.  Zednai  (1968),  in a laboratory study, demonstrated that flaming reduced spotted knapweed seed germination from 69 to 3%. however,  Zednai (1968)  chemical treatments  In a field study,  found that burning combined with various  (i.e.,  2,  4, or 6 oz/ac picloram) provided no  better control than the herbicide treatment alone. 4  Spotted  knapweed may not be susceptible to control by fire because it is more prone to vegetative reproduction (Watson and Renney 1974). Furthermore, none of Zednai’s  (1968)  treatments were replicated,  so the results may not be reliable. Despite these findings, there are no studies in B.C.  on the  effectiveness of fire for controlling knapweed, perhaps because of perceived risks regarding escape into forested areas.  In  KLPP, however, where fire control personnel and equipment can be concentrated on localized areas, prescribed burning of knapweed stands is quite feasible. prescribed burns at KLPP  Moreover, the information from these would provide valuable experience and  data regarding the potential of fire to control knapweed on infested grasslands throughout British Columbia.  It is also  important to document the potential of prescribed burning for maintaining desired successional stage(s)  and floristic  diversity. This research project proposes to investigate the potential of prescribed burning in Kalamalka Lake Provincial Park to control diffuse knapweed and to enhance floristic diversity. The specific research objectives of this project were: 1.  To determine the extent of current diffuse knapweed infestation throughout the park, and to identify the factors that are most important in making an area susceptible to infestation by diffuse knapweed;  5  2.  To quantify the current species richness and floral diversity of the major grassland ecosystem associations and seral stages, and identify the factors most important in maintaining a high degree of species richness;  3.  To fully document the characteristics of fall and spring prescribed burns,  so that this kind of information is  available when comparing different prescribed burning studies, 4.  and developing management guidelines.  To determine the feasibility of using prescribed burning to control diffuse knapweed;  5.  To quantify the effects of prescribed burning on species richness and floral diversity; and  6.  to develop management guidelines for Kalamalka Lake Provincial Park.  6  2.0  MATERIALS AND METHODS  2.1  Study Area  The study site was located in south central British Columbia 10 km south of Vernon, B.C.  (50°15’N,119°12’W),  shore of Kalamalka Lake (Fig. is 890 ha in size. frontage,  1).  along the eastern  Kalamalka Lake Provincial Park  It is bordered on the west by 5 km lake  on the north by the village of Coldstream, on the east  by orchards and small mixed farms, and on the south by steep, tree—covered slopes (Fig. 700 m.  2).  Elevation ranges from 330 to  The topography is dominated by a single large hill,  Rattlesnake Hill, with a very steep and rocky south face.  A  small grass—covered valley runs parallel to Rattlesnake Hill on its south side. of the park,  The north side of Rattlesnake Hill, and the rest  is made up of gentler and more accessible slopes,  which are covered in a patchwork of grasslands and forests. Average annual precipitation is 41 cm, the winter and late spring. days.  most of which comes in  The frost—free period averages 153  The January mean maximum temperature is —2.6 C while the  mean minimum temperature for that month is —9.4 C.  The mean  maximum and mean minimum for July are 27.0 C and 11.2 C, respectively.  Meteorological data comes from the nearest  Environment Canada Weather Observation Station on the Coldstream Ranch (Environment Canada 1984),  5 km east of the park. 7  lama ika al Park  Fig.  1  Map of British Columbia showing the location of Kalamalka lake Provincial Park.  8  Fig. 2 Map of Kalamalka Lake Park which is bordered on the west by 5 km lake frontage, on the north by the village of Coldstream, on the east by orchards and small mixed farms, and on the south by steep, tree—covered slopes. Location of diversity! knapweed transects (Appendix I) and burn plots (Appendix II) can be found at the end of this report. 9  The vegetation is in the Okanagan very dry hot interior Douglas—Fir Zone (IDFxhla)  (Lea et al.  dominated by Kentucky bluegrass thread grass needlegrass  1990).  The valley is  iratensis L.), needle-and-  (Stipa comata Trin.  & Rupr.), and Nelson’s  (Stipa occidentalis var. nelsonii  (Scribn.) Hitchc.).  Important herbaceous associates include arrowleaf balsamroot (Balsamorhiza sagittata (Pursh)Nutt.), silky lupine sereceus Pursh), and sulphur cinquefoil Dougl.).  (Potentilla gracilis  Slopes with northern exposure are dominated by Douglas—  fir (Pseudotsuga menziesii (Pinus ponderosa Dougi.), Nutt.),  (Lupinus  (Mirbel)  Franco.), ponderosa pine  Saskatoon (Amelanchier alnifolia  and common snowberry (Svmphoricarpus albus  (L.)Blake).  Plant taxonomy throughout this report is based on the work of Hitchcock and Cronquist (1973). The grasslands in the park generally occurred on soils from a morainal or fluvial origin. from the Kalamalka series  The soils are black chernozems  (Kelly and Spilsbury 1949).  Soil  texture throughout the park ranged from sandy loam to silty clay loam.  2.2  Field Methods  2.2.1  Diffuse knapweed survey  The number of knapweed seedlings, rosettes and bolted plants were determined along 3 transects (82 m)  in each of six different  bunchgrass seral stage/ecosystem association.  10  These six  bunchgrass seral stage/ecosystem associations represent the most common types of grassland in the park.  They were bluebunch  wheatgrass—Idaho fescue (Festuca idahoensis Elmer) morainal habitat unit successional stages fair (WFm2)  and good (WFm3),  bluebunch wheatgrass-Idaho fescue fluvial habitat unit successional stage fair (WFf2), bluebunch wheatgrass—pasture sage (Artemisia frigida Willd.)  southerly aspect habitat unit  successional stage good (WS3), Kentucky bluegrass—bluebunch wheatgrass moist grassland habitat successional stage fair (BW2), and Douglas—fir—common snowberry northerly aspect habitat unit in the shrub-herb stage (DS1). The location of the transects was determined using a map of the park’s ecosystem associations.  Transects were placed in the  center of each ecosystem association of interest on the map first, and then located on the ground by measuring distances from land marks (1970)  (i.e. power lines, roads, and trails).  Daubenmire  suggests that sampling of 4.0 m 2 should be enough to  characterize a grassland.  Therefore,  forty 0.1—rn 2 (20 x 50 cm)  plots were placed along each transect systematically to-provide equi-distant interspersion of sampling units.  The results of all  forty 0.1—rn 2 plots on each transect were used to calóulate an average knapweed density for each transect.  All transects were  sampled for knapweed density in May and again after bolting, approximately August 20. The amount of ground cover in each of four classes soil,  (bare  litter, rock, and bryophytes) was visually estimated in 10% 11  increments within each 0.1—rn 2 plot by using measurements that were painted on the side of the quadrat frame.  The results from  the forty 0.1—rn 2 plots were averaged to provide a single estimate for the various classes of ground cover for each site. Soil texture was determined by taking two cores (2.5 x 15 cm)  at equally—spaced distances along each transect (i.e. 27.3  and 54.6 m).  The soil was analyzed using the hydrometer  technique (Gee and Bauder 1986). Aspect was determined by using a hand—held compass. transects were laid out along the contour of the slope.  All The  percent slope was determined by taking two measurements with a hand held clinorneter.  One measurement was taken upslope from the  center of the transect and one measurement was taken downslope from the center.  These two measurements were averaged to derive  one measurement of percent slope for each site.  2.2.2  Richness and diversity study  Species richness,  simply the number of species, and floral  diversity, a measure that combines species richness with evenness of distribution, were determined along 3 transects  (82 m)  in each  of six different bunchgrass seral stage/ecosystem associations. The same 18 transects that were used in the diffuse knapweed survey were used for this part of the study (section 2.2.1). Forty 0.1-rn 2 (20 x 50 cm)  plots were placed along each transect  systematically to provide equi-distant interspersion of sampling units.  Species richness was determined by recording all plant 12  species in each plot.  Floral diversity was assessed with the  Shannon-Wiener Index of Diversity (H’)  (Barbour et al.  1980),  which is sensitive to rare species: H 1 -  ) (logp 1 (p ), 1  where  (1)  =proportion of individuals of species i. 1 p  Diversity was assessed by recording total number of 0.1—rn 2 plots that each species was present in.  Sampling for floral diversity  and species richness occurred in May, coinciding with the peak period of spring ephemerals all forty 0.1—rn 2 plots  (Pitt and Wikeem 1990).  The data for  along a transect were pooled to provide  a single estimate of richness, and diversity for each site.  2.2.3  Prescribed burn documentation  In the fall of 1990 two sets of nine 10 x 45—rn rectangular plots were established,  one on a knapweed—infested site and one  on an uninfested Kentucky bluegrass grassland.  The diffuse  knapweed infested study site was located in a large level opening on the north east side of Rattlesnake hill, ecosystem association/seral stage.  in an area in the BW2  It was dominated by Kentucky  bluegrass, needle—and—thread grass, bluebunch wheatgrass, and cheat grass  (Bromus tectorum L.).  Important herbaceous  associates included diffuse knapweed and sulphur cinquefoil. There were small patches of  Saskatoon berry and common snowberry  13  throughout the area.  The Kentuôky bluegrass study site was  located in the valley on the south side of Rattlesnake Hill, an area in the Wfm2 ecosystem association/seral stage.  in  It was  dominated by Kentucky bluegrass, needle—and—thread grass, and Columbia needlegrass,  Important herbaceous associates included  silky lupine and sulphur cinquefoil. In the fall of 1990 and in the spring of 1991 3 replicate plots were burned at both sites. to ignite the burns.  A hand held drip torch was used  A single strip along the windward side of  the plot was ignited.  These headfires were then allowed to burn  until they reached the fireguard on the other side of the plot. Gravimetric soil and fuel moisture content were assessed before each burn.  Fuel moisture content was assessed for one class of  fuel only, grasses and other fine fuels  (f orbs and small twigs).  The burn plots had only very small amounts of woody fuels in them.  Soil moisture was determined for each site by taking four  randomly located cores before the burn.  (2.5 x 15 cm)  on each plot immediately  Fuel and soil moisture samples were oven dried  and moisture content is expressed as weight of water removed by drying over the oven-dry.weight of the fuel. Frontal fire intensity (I)  (Barbour et al.  1980) was  assessed for each plot: I  =  H x W x R,  where I H W  = = =  (2)  fire intensity (kwatts m ), 1 heat yield (kjoules kg ), 1 fuel consumed (kg rn ), and 2 14  The variable for heat yield (H) was determined from published data for grassland fuels with various levels of fuel moisture (Brown and Davis 1973)  (Appendix V).  W (fuel consumed) was  estimated by clipping and collecting the leaf litter of four randomly located 0.l-m 2 plots in each 10 x 45 m replicate plot immediately before burning, and immediately after burning.  This  is not an exact measurement of the fuel consumed in the front of the fire; however,  in these fires there was very little burning  other than that which was in the fire front.  The rate of spread  (R) was determined by measuring the amount of time required to burn each fire strip then dividing that by the length of the strip  in order to determine rate spread in meters per second  (m/s).  Soil temperature was monitored during each burn.  fall 1990 burn,  soil temperature probes (Forest Technology  Systems TL11 thermologger) were placed at three depths (0, and 2.50 cm)  For the  in four locations in each burn plot.  1.25,  However, these  probes do not begin to record temperatures until a temperature of 60 C is reached at the uppermost probe.  Because of the cold damp  conditions during the burn, temperatures at the soil surface did not reach 60 C.  Therefore, for the burn in the spring of 1991  the probes were placed at 10 cm above the surface, on the soil surface,  and 2.5 cm below the soil surface.  2.2.4  Impact of prescribed burning on the density of diffuse knapweed  Both sets of burn plots were sampled to determine the impact 15  of prescribed burning on diffuse knapweed.  Sampling for percent  bare soil, knapweed seedlings, rosettes, and bolted knapweed plants in the fall—burned replicates the fall 1990 burn, growing season.  occurred 3 times: before  in May 1991 and at the end of the 1991  Sampling for these variables also occurred 3  times in the spring-burned replicates: before the 1991 spring burn,  in May 1991 and at the end of the 1991 growing season.  Sampling for percent bare soil, knapweed seedlings, rosettes, and bolted knapweed plants also occurred in unburned plots in the knapweed—infested and Kentucky bluegrass sites adjacent to the respective burned plots.  Sampling of these variables in unburned  plots occurred four times: before the fall—burn, before the spring burn,  in May 1991,  and at the end of the 1991 growing  season. The density of knapweed seedlings, rosettes,  and bolted  knapweed plants in the burned and control plots were assessed by counting the number of individuals in 40 0.1-rn 2 quadrats. 40,  0.1—rn  2  These  plots were located systematically along a transect  running lengthwise down the replicate plot center. was 41 m long,  2.2.5  Each transect  leaving a 2.0—rn buffer.  Impact of prescribed burning on species richness and floral diversity  Sampling for floral diversity, 16  and species richness in the  fall-burned replicates occurred three times: before the fall 1990 burn,  in May 1991 and at the end of the 1991 growing season.  Sampling for these variables also occurred 3 times in the spring—burned replicates: before the 1991 spring burn, 1991 and at the end of the 1991 growing season.  in May  Sampling for  floral diversity, and species richness also occurred in unburned plots in the knapweed-infested and Kentucky bluegrass sites adjacent to the respective burned plots.  Sampling of these  variables in unburned plots occurred four times fall-burn, before the spring burn,  —  before the  in May 1991, and at the end of  the subsequent growing season. Floral diversity, and species richness in the burned and control plots were assessed using the same methods that were used in the richness and diversity study for the park as a whole (section 2.2.2).  Each transect was 41 m long,  leaving a 2.0-rn  buffer at each end.  2.3  Statistical Analysis  In order to identify significant differences in the density of all categories of knapweed (seedlings, rosettes, bolted plants, and total density) the data were analyzed with a two factor ANOVA randomized within six ecosystem associations/seral stages and two years.  A multiple discriminant analysis (MDA) was  done in order to identify the factors that were most often associated with a diffuse knapweed infestation. 17  MDA involves  deriving the linear combinations of two (or more) variables that will discriminate best between groups  (Hair et al.  1987).  ,  independent  priori defined  It is most appropriately used when  there is one categorical dependent variable (in this case, the level of knapweed density)  and two or more independent variables.  The independent variables used for the knapweed density MDA were: Access, a measure of accessibility of a site; Soil, percent bare soil on a site; slope, percent slope of a site; clay, percent clay in the soil; sand, percent sand in the soil.  The  derivation of the best possible functions for separating the groups is achieved by maximizing the between group variance relative to the within group variance.  The linear combinations  for a MDA are derived from an equation that takes the following form:  z  =  + 1 w . x 2 ..wx, w  where  (3)  Z=the discriminant score, W=the discriminant weights, and X=the independent variables. The result of this equation is a single composite discriminant score.  The mean discriminant score for each  knapweed density group (referred to as the centroid) calculated.  can then be  The centroids show the typical location of a group  member along this new dimension.  Analysis involves interpreting  these new dimensions and determining whether or not the distance between groups is significant.  This analysis is based on the 18  assumptions of multivariate normality and unknown but equal dispersion, and covariance structures of groups.  There is  evidence, however, that discriminant analysis is not strongly influenced by violations of these assumptions (Hair et al.  1987).  For the purpose of this study, the interpretation of the factors will be discussed.  None of the variables used in the multiple  discriminant analyses have a normal distribution, so inferences about the distance between groups is meaningless.  The  reliability of the factors will be determined by using a validation matrix technique (Hair etal.  1987).  The results from both years were combined into a single analysis group for the purpose of the MDA.  The results from  these 36 transects were then divided into three levels of knapweed density (zero, medium,  and high).  The figures from the  fall knapweed survey were used because knapweed was more wide spread in the fall.  The medium group included all transects that  had between 0.1 and 1.0 knapweed plants (at any growth stage) per square meter.  The high knapweed group included all transects  that had a density of more than 1.0 plants per square meter. In order to identify significant differences in species richness (R)  and floristic diversity (H’)  the data were analyzed  with a two factor ANOVA randomized within six different ecosystem associations/seral stages and two years.  The Student Newman—  Kuels multiple range test was used to identify significant differences in richness and diversity between all pairs of the six different ecosystem associations/seral stages. 19  In order to  identify the factors that were most often associated with high levels of species richness a multiple discriminant analysis (MDA) was done.  The results from both years were combined into a  single analysis group for the purpose of the MDA.  The results  from these 36 transects were then divided into three levels of richness  (low, medium, and high).  The low richness group  included all cases that were at least one standard deviation below the mean level of richness for all 36 cases.  The medium  group included all cases that were within one standard deviation of the mean,  and the high richness group included all cases that  were at least one standard deviation above the mean.  The micro  computer statistical package SYSTAT was used for all of the above mentioned statistical analysis. The objectives of the prescribed burn study required that descriptive statistics be calculated for the documentation of the burn characteristics.  Means and 95% confidence intervals have  been calculated where possible. the burn, however,  In the case of wind speed during  it was only possible to report ranges because  only a small hand held anemometer was available to measure wind speed.  Two models  (one for predicting fire intensity and one for  predicting the impact of fire on ground cover) were calculated using multiple linear regression options in the micro—computer package SYSTAT. Differences in knapweed density before and after the burn were examined using a repeated measures ANOVA (Little and Hills 1978).  Initially a two site by three burning treatment by four 20  date design with three replicates per cell was used.  There was  no knapweed on the Kentucky bluegrass site before the burn and none was found in the two post burn surveys of the initially uninfested plots.  Therefore,  it was decided to analyze the  infested site separately, using a three treatment by four date ANOVA with three replicates per cell.  Data for missing cells  were estimated using a least squares method.  Using this method  does not affect the total sums of squares because estimates are based on differences from the grand mean that are calculated from the non—missing cells. Changes in species richness, classes of ground cover  floral diversity,  (bare soil,  and various  litter, rock, and bryophytes)  were investigated using the two site by three burning treatment by four dates  (three dates in the case of the ground cover  variables) ANOVA on LOTUS.  21  3.0 RESULTS AND DISCUSSION  3.1  The Diffuse Knapweed Problem in Kalamalka Lake Provincial Park  Diffuse knapweed was found throughout the park.  It was found  in each type of ecosystem association/seral stage at one time or another during the study (Fig.  3).  There was a high density of  knapweed along some of the main trails.  All of our transects  were placed in the middle of the site (usually away from trails); therefore, our measurements may tend to under estimate the extent of the problem in well—travelled areas of the park. measurements, however,  Our  should more accurately describe the extent  of the problem in the park as a whole. There was no difference in the density of knapweed in the different ecosystems from one year to the next (i.e.  interaction  between ecosystem and year was not significant >O.lO) and 2).  Knapweed density did not differ (>O.O5)  (seedlings, rosettes, or bolted plants)  (Tables 1  in any category  amongst the different  types of ecosystem associations in either the fall or spring surveys. In the spring of 1990 one of the transects in the WS3 ecosystem association/seral stage had a very large number of knapweed seedlings and rosettes  (over 30 2 plants/m ) ; however,  only 4.75/ni 2 of these survived to the fall 22  (fig.  3).  In the  Spring  i  990  Spring  10  10  8  8  6  I.  1991  6  4  4  2  2  0 BW2  DS 1  0  WFf2 WFm2 WFn,3 W33  8W2  DSi  FaIl 1990  FaIl 1991  10  10  8  8  6  6  4  4  2  2  >  0 2  DS 1  WFI2 WFm2WFm3 WS3  WFf 2 WFm2 WFm3 WS3  0  I  8W2  DS I  WFf2 WFm2 WFm3 WS3  Mean (n=3) density (plants/rn 2) of diffuse knapweed in Fig. 3. six different grassland ecosystem association/seral stages of Kalamalka Lake Provincial Park during the spring and the fall of 1990 and 1991. These six grassland ecosystem associations /seral stages are bluebunch wheatgrass—Idaho fescue morainal habitat successional stages fair (WFm2) and good (WFxn3), bluebunch wheatgrass—Idaho fescue fluvial habitat successional stage fair (WFf2), bluebunch wheatgrass-pasture stage southerly aspect habitat successional stage good (WS3), Kentucky bluegrass bluebunch wheatgrass moist grassland habitat successional stage fair (BW2), and Douglas—fir—common snowberry northerly aspect Differences between habitat in the shrub-herb stage (DS1). ecosystems in the same year and for the same ecosystems in different years were not significant (>0.05) 23  Sources of variation, degrees of freedom (df), sums of Table 1 squares (ss), mean squares (ms), F—ratio, and associated probability (P) for a two factor analysis of variance randomized within two years and six ecosystem associations for four categories of diffuse knapweed density (seedlings, rosettes, bolted plants, and total knapweed) in Kalamalka Lake Provicial Park during early May 1990 and early May 1991.  Source  df  ins  ss  F—ratio  P  seedling density ecosystem year ecosystem*year  5 1 5  55.307 9.766 55.307  11.061 9.766 11.061  error  24  267.042  11.127  0.994 0.878 0.994  0.442 0.358 0.442  0.849 0.007 0.812  0.529 0.935 0.553  rosette density ecosystem year ecosystem*year  5 1 5  9.863 0.016 9.423  1.973 0.016 1.886  error  24  55.750  2.323  bolted plant density ecosystem year ecosystem*year  5 1 5  3.951 0.063 0.395  0.790 0.063 0.079  error  24  11.167  0.465  1.699 0.134 0.170  0.176 0.717 0.971  total diffuse knapweed ecosystem year ecosystem*year  5 1 5  110.411 9.252 116.384  22.082 9.252 23.277  error  24  559.000  23.292  24  0.948 0.397 0.999  0.468 0.534 0.439  Table 2 Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability () for a two factor analysis of variance randomized within two years and six ecosystem associations for four categories of diffuse knapweed density (seedlings, rosettes, bolted plants, and total knapweed) in Kalamalka Lake Provicial Park during late August 1990 and late August 1991.  Source  df  ss  ms  F—ratio  seedling density ecosystem year ecosystem*year  5 1 5  0.056 0.111 0.056  0.011 0.111 0.011  error  24  0.417  0.017  0.640 6.400 0.640  0.671 0.018 0.671  1.016 0.146 0.152  0.430 0.705 0.977  rosette density ecosystem year ecosystem*year  5 1 5  3.854 0.111 0.576  0.771 0.111 0.115  error  24  18.208  0.759  bolted plant density ecosystem year ecosystem*year  5 1 5  0.988 0.210 0.988  0.198 0.210 0.198  error  24  7.292  0.304  0.650 0.691 0.650  0.664 0.414 0.664  total diffuse knapweed ecosystem year ecosystem*year  5 1 5  7.488 0.210 0.905  1.498 0.210 0.181  error  24  39.125  1.630  25  0.919 0.129 0.111  0.486 0.723 0.989  three surveys of that area that followed over the next 16 months, the density of knapweed was not unusually high.  Moreover, even  when knapweed density was very high on that transect, there was no knapweed present on the other two transects in that type of ecosystem association/seral stage.  Although herbicides have been  used in some of the study areas in the past few years, there were no herbicides used in any of the study areas during the course of this study.  Despite this, there was no increase (>O.O5)  in the  density of knapweed from one year to the next. There was no significant correlation between the density of any category of knapweed and either species richness or Shannon’s diversity index.  This may not contradict the findings of Tyser  and Key (1988), who were working on a small scale in heavilyinfested areas.  Certainly,  if one looked at an individual  2 quadrat, certainly an increase in knapweed stem density 0.1—rn would be strongly correlated with a decreased frequency of native plants.  However,  in this study we were more concerned with the  park as a whole, and it appears that in that case there is no significant threat to species richness or plant diversity by diffuse knapweed at this time.  That is not to say that one  should stop trying to control diffuse knapweed in the park, but just that there are still a number of areas in the park that are not heavily infested at all. In order to determine what factors are most often associated with high levels of diffuse knapweed infestation a multiple discriminant analysis  (MDA) was calculated. 26  A simultaneous  computational method was used to calculate the MDA.  In this  method all independent variables are considered concurrently. Therefore, the MDA is computed based on the entire set of variables, regardless of the discriminating power of the individual variables.  Five variables were used in the MDA.  The  variable, -access, measured the ease or difficulty of access to a particular site.  A site was given a value of 1.0 if it was near  a road or main trail; therefore, very easy to access. Conversely, a site was given a value of 4.0 if it was more than 500 m from any trail or could only be accessed by climbing a steep slope (Appendix III).  It seemed from casual observation  that knapweed was most dense in heavy traffic areas.  Using this  variable in the MDA would be one way of determining how important the level of access is in relation to other variables. variable, bare soil.  The  soil, measured the percent ground that was covered with Diffuse knapweed seems to thrive on disturbed sites.  One way of determining the level of disturbance on a site is to measure the percent bare soil.  The theory being that with more  frequent or more severe disturbances there would be more bare soil.  The variable,  slope,  (percent slope of a site) will  dictate the value of several other important abiotic variables (i.e. runoff,  erosion, and drainage).  Clay (percent clay in  soil), and sand (percent sand in the soil) were used to determine the importance of soil texture. The maximum number of functions that can be calculated in a three group MDA is two.  These functions are linear combinations 27  of the original independent variables.  Only the first factor was  strongly correlated (cannonical correlation>O.6) in knapweed density (Table 3).  to the variation  The first factor was very  strongly negatively correlated with the variable that measured access.  This clearly shows how important traffic is in spreading  knapweed, and it confirms our suspicion that as the level of traffic increases so increases the level of infestation.  knapweed  Percent slope was also fairly strongly negatively  correlated with the first factor.  Percent slope was strongly  correlated with the access variable (r=O.732); steeper slopes generally mean a more inaccessible site.  Percent slope may also  have implications for the physical characteristics of the site in that it will affect drainage, soil texture, and the amount of runoff. Although the second factor was only weakly correlated with the differences amongst the groups,  it may provide some additional  information on what is important in creating a knapweed infestation.  The second factor is most simply expressed as a  combination of the percent bare soil and soil texture. Apparently, percent bare soil and soil texture are only weakly correlated with knapweed density.  A soil disturbance may have a  small influence on increasing knapweed density, particularly if the site has a high percent clay.  These interpretations are  based on the canonical loadings (correlations between conditional dependent variables and canonical factors).  28  Table 3. Canonical correlations and canonical loadings of a three group multiple discriminant analysis for diffuse knapweed density in Kalamalka Lake Provincial Park during the fall of 1990 and 1991. factor 1  factor 2 Canonical Correlations 0.604  0.242  Canonical Loadings variable ACCESS SOIL SLOPE CLAY SAND  —0.938 —0.201 —0.526 0.449 —0.189  0.006 0.483 —0.291 0.337 —0.389  A validation matrix (Table 4) was constructed to determine how well these functions correctly predicted the level of diffuse knapweed density.  The success rate or ‘hit ratio’  by summing the values on the diagonal  is determined  (correctly predicted cases)  and dividing them by the total number of transects.  The factors  calculated in this MDA correctly predicted the level of diffuse knapweed in 24 out of 36 cases (66.67%).  This is well above the  53% correct that would be possible if all cases were simply assigned to the largest group.  Hair et al.  (1987)  suggest that  classification accuracy should be at least 25% greater than what is possible due to chance alone (in this case 66.25%).  29  Table 4. Validation matrix for the three group, 36 case multiple discriminant analysis of the density of diffuse knapweed during the fall of 1990 and 1991 in Kalamalka Lake Provincial Park.  Actual Group (Rows) by Predicted Group (Columns) Predicted Group Actual Group  1  2  3  1  17  2  0  19  2  1  7  1  9  3  5  3  0  8  12  1  36  Total  23  Total  Ideally one would split the cases into two groups: the analysis group for calculating the factors; and a hold-out sample to test the validity of the factors.  This should only be done if the  total number of cases is over 100 only 36 cases in our study.  (Hair et al.  1987).  There were  This small number of cases meant  that all 36 cases had to be used in both the analysis group and validation group.  This results in an inflated predictive  accuracy, but it is better than not testing the factors at all (Hair et al.  1987).  These functions may not be valid if used to  analyze areas outside KLPP.  However, they are very useful in  identifying the factors most often associated with diffuse knapweed infestation in KLPP.  Perhaps the greatest weakness of  this model is that it tends to underestimate the level of knapweed infestation on a site (i.e. all high density sites were 30  predicted to have either medium or zero levels of knapweed density).  While the level of accessibility may be one important  factor in creating a knapweed infestation, there may be some other factors that we have not yet identified  —  especially in  causing high levels of knapweed density. In order to classify new areas using the raw data, group classification function coefficients and group classification constants (Table 5)  can be used.  The raw data are simply fit to  the equation: z  =  . 1 c+wx+wx  .  .  Wnxn,  where  (4)  Z  =  group classification score,  c  =  group classification constant,  W  =  group classification coefficient for a given variable, and  X  =  raw data for a particular variable  31  Table 5. Group classification constants and group classification coefficients for a three group multiple discriminant analysis of diffuse knapweed density in Kalamalka Lake Provincial Park during the fall of 1990 and 1991.  Group 1  Group 2  Group 3  Group Classification Constants —158.537  —163.721  —160.955  Group Classification Coefficients Variables ACCESS SOIL SLOPE CLAY SAND  16.918 0.058 —1.673 7.260 3.777  15.110 0.092 —1.594 7.458 3.885  16.549 0.167 —1.847 7.358 3.800  This process is repeated for all three groups of constants and coefficients; the new case should be assigned to the group with the highest score.  This process will help determine the  potential of a site to become infested with knapweed.  Sites that  are classified in medium or high knapweed density group are at risk of becoming infested and should be managed appropriately.  32  3.2  Species Richness and Floral Diversity in Kalamalka Lake Provincial Park  Throughout this two year study, recorded in KLPP  95 species of plants were  (Appendix 1); however,  43  number recorded along any one transect.  was the highest  The various ecosystem!  associations did not differ in species richness from one year to the next  (i.e. the ecosystem by year interaction was not  significant >O.O5) (R>O.05)  (Table 6).  There were no differences  in species richness from one year to the next  (Fig.  4).  Table 6 Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability () for a two—way analysis of variance randomized within two years and six ecosystem associations for species richness and the Shannon—Weinner diversity index in Kalamalka Lake Provicial Park during early May 1990 and early May1991.  Source  df  ss  ms  F—ratio  P  species richness ecosystem year ecosystem*year  5 1 5  error  24  1606.556 5.444 98.889  321.311 5.444 19.778  817.333  34.056  9.435 0.160 0.581  0.000 0.693 0.714  Shannon—Weinner diversity index ecosystem year ecosystem*year  5 1 5  1.040 0.026 0.032  0.208 0.026 0.006  error  24  0.594  0.025  33  8.403 1.040 0.259  0.000 0.318 0.931  In the spring of 1990 the average richness for the eighteen transects was 21.83,  and in the spring of 1991 it was 20.39.  There were significant differences  (<0.05)  in species richness  amongst the six different types of communities that were studied (Fig.  4).  The DS1 ecosystem association was richer (<O.O5) than  any of the other ecosystem associations.  This ecosystem  association is actually an early successional stage of a forested ecosystem association.  It is found only on north facing slopes.  It contains most of the grassland species such as Kentucky bluegrass, bluebunch wheatgrass, diffuse knapweed,  june grass,  silky lupine, and  as well as common chokecherry (Prunus  virginiana L.), round leaf alumroot (Heulchra cvlindrica Dougi.), lanceleaved stonecrop (Sedum lanceolatum Torr.),  and woodsia  (Woodsia scopulina D.C.  Eat.) which were found only in this  ecosystem association.  These species may be better adapted to  forest conditions.  These areas may be so rich because they are  transitional between grassland and forests; therefore, they can support many grassland species and some forest species. The BW2 ecosystem association had the lowest number of species. All three of the sites in this ecosystem association had been sprayed with herbicide (piclorain) within the last two years. This type of herbicide eliminates all the broadleaved f orbs for at least a few years after application and seriously reduces the species richness (i.e. the number of species)  of an area.  WFm3 ecosystem association had the second lowest species richness.  It also contained areas that had been sprayed. 34  The  Richness 1091  Richness 1 990  U)  cn  45  4’  40  40  35  35  30  30 v  2 20  20  15  15  10  10  5  5 0  0 BW2  DS1  ecosystem  BW2  WFI2 WFrri2WFm3 WS3 assodatiorvseral  DS1  WFf2 WFm2 WFm3 WS3  ecosystem ossocation/sera stage  stage  Diversity 1 99 1  Diversity 1 990  15  1.5  x C V .  V C  1.0  1.0  >.  >  c  C  S  05  05  00 BW2  OS1  8W2  WFf2 WFm2 WFm3 WOO  OS1  WFI2 WFm2 WFm3 WS3  ecosystem associettorvseral stage  ecosystem assoclatlorvserat stage  Fig. 4. Mean (n=3) species richness and Shannon’s diversity index in six different grassland ecosystem association/seral stages of Kalamalka Lake Provincial Park during the spring of 1990 and 1991. The Douglas-fir-common snowberry northerly aspect habitat in the shrub-herb stage (DS1) was richer and more diverse (>O.O5) than all other ecosystem types (i.e. bluebunch wheatgrass—Idaho fescue morainal habitat successional stages fair (WFm2) and good (WFm3), bluebunch wheatgrass—Idaho fescue fluvial habitat successional stage fair (WFf 2), bluebunch wheatgrass-pasture stage southerly aspect habitat successional stage good (WS3), Kentucky bluegrass—bluebunch wheatgrass moist grassland habitat successional stage fair (BW2)). The BW2 ecosystem was less diverse (>0.05) than the four most diverse ecosystems. Differences between the same ecosystems in different years were not significant (>0.05) 35  Species diversity measures are a product of both species richness and the evenness of distribution of these species.  For  example, two comnu.inities may both have ten different species Community A may have 91 members from one species and one from each of the other nine.  Community B may have 10 members from  each of the ten species.  In this case both communities are  equally rich, but community B is more diverse than community A (i.e. H’aO. , H’b=l.OO). 22  Several measures of diversity have  been proposed and many are presently in use.  The two most common  are Simpson’s diversity index and Shannon’s diversity index (Barbour et al.  1980).  Simpson’s diversity is based on the  probability of choosing two different species at random.  It  requires counts of the number of individuals in a community and tends to give very little weight to rare species.  For the  purpose of this study it was decided that Shannon’s diversity index (H’)  should be used because it can be used with any measure  of abundance (i.e. number of individuals in each species, frequency of each species, or relative cover), and it is more sensitive to rare species. ‘uncertainty’  H’  in a community.  can be thought of as the amount of The more variable a community the  more uncertain (or unpredictable)  each sample of it would be.  In  the end, diversity is simply one more descriptive measure of a plant community on par with biomass, or leaf area index. Shannon’s diversity index (H’) was highly correlated with species richness (r=0.953).  This is to be expected as richness  is a major component of this index. 36  The results of the ANOVA for  H’ were therefore very similar to that for R (Table 6).  There  was no difference in the diversity of the different communities from one year to the next (i.e. the ecosystem by year interaction was not significant >O.O5). in H’  There was no significant difference  from one year to the next.  The DS1 ecosystem association  was significantly more diverse than any of the other ecosystem associations.  The BS2 ecosystem association was significantly  less diverse than the four most diverse ecosystem associations. There was no significant correlation between the density of any category of knapweed density and either species richness or Shannon’s diversity index.  The frequency of Kentucky bluegrass,  however, was negatively correlated with diversity (r=—O.587) richness  (r=—O.534).  and  Kentucky bluegrass forms a very dense sod  which may be able to eliminate the possibility of other plants becoming established.  Furthermore, Kentucky bluegrass would not  be affected by the herbicide treatment; therefore, when the herbicide was applied it freed this grass from competition with the broadleaved f orbs. Litter was negatively correlated to diversity (r=-O.564) richness  (r=-O.509).  and  Percent of ground covered by leaf litter  was strongly positively correlated with the frequency of Kentucky bluegrass  (r=O.809).  Obviously communities with a dense stand of  Kentucky bluegrass have a an extensive cover of leaf litter which restricts the establishment of many other species.  Silky lupine  (r=O.652), swamp saxifrage (Saxifragia integrefolia Hook.) (r=O.577), and Saskatoon (r=O.623) were strongly positively  37  correlated with diversity.  Silky lupine was found in almost all  types of ecosystem associations, and it may be a good indicator for richness and diversity.  Saskatoon and swamp saxifrage were  found mainly in the richest and most diverse early successional forest ecosystem association, DS1.  therefore, they may not be  good indicators of richness or diversity in grassland communities.  It is only an indicator of the DS1 ecosystem  association. In order to determine which factors were most often associated with high levels of species richness a multiple discriminant analysis  (MDA) was done.  A simultaneous  computational method was used to calculate the MDA.  In this  method all independent variables are considered concurrently. Therefore,  the MDA is computed based on the entire set of  variables, regardless of the discriminating power of individual variables.  Eight variables were used in the MDA.  The first  variable was herbicide, the number of years since herbicide treatment.  The herbicide treatment history of each site was  determined from maps provided by B.C. Parks (Appendix III).  Not  all of the sites in this study had been treated with herbicide. The sites that had been treated, however, were given a value of one, two, or three depending on the number of years since they were last treated. a value of seven.  Sites that had never been treated were given This assumes that the herbicide will not  persist in the soil for more than seven years and/or sites that have not been treated for seven years will have the. same richness 38  as sites that have never been treated.  Under ideal soil and  climatic conditions picloram is only able to prevent reinvasion of knapweed for seven years  (Cranston et al.  The problem  1983).  with this assumption is that picloram may affect other species for longer.  Therefore picloram could retard the level of  richness for more than seven years.  If anything the impact of  picloram may be underestimated using this method, one must keep this in mind when looking at the interpretation of the MDA.  The  second variable was access, a variable that measured the ease or difficulty of access to a particular site.  A site was given a  value of one if it was near a road or main trail; therefore, very easy to access.  Conversely, a site was given a value of four if  it could be accessed only by hiking more than 500 trail or  by climbing a steep slope.  in  from any  Two dummy variables were  used to designate if a site had a north or south aspect, and/or a east or west aspect (e.g.  a site that faced almost due north  would be coded —1,0 for these two variables; whereas, easterly facing site would be coded 1,—l). slope  (percent slope of a site), soil  a south  Other variables were  (percent ground that was  covered with bare soil), clay (percent clay in soil), and sand (percent sand in the soil). The maximum number of factors that can be calculated in a three group MDA is two.  These factors are linear combinations of  the original variables.  Both factors were strongly correlated  (cannonical correlation richness  (Table 7).  >  0.55)  to the variation in species  The first factor was most strongly 39  correlated with the number of years since herbicide treatment. Clearly the number of years since herbicide treatment was the most important variable in determining the richness of a site. The second factor is a combination of the importance of north vs south aspect, and soil texture.  This factor reaffirmed that the  northerly facing sites, and the sites with a higher sand content in the soil are more rich.  Northern facing slopes are likely  more rich because they are more moist. may not persist as well in sandier soil.  The herbicide picloram These are obviously  characteristics that can not be changed by management,  but they  might help identify sites with the potential for high levels of floral richness. canonical loadings  These interpretations are based on the (correlations between conditional dependent  variables and canonical factors). Table 7. Canonical correlations, and canonical loadings of a three group multiple discriminant analysis for species richness in Kalamalka Lake Provincial Park during the spring of 1990 and 1991. factor 1  variable Herbicide Access aspect N/S aspect E/W SLOPE SOIL CLAY SAND  factor 2  Canonical Correlations 0.865 0.552 Canonical Loadings 0.512 0.395 0.242 0.091 0.242 0.178 —0.471 0.335  0.279 0.109 —0.544 0.196 —0.086 0.217 —0.388 0.579  40  A validation matrix was constructed to determine how well these factors predicted the level of richness success rate or ‘hit ratio’ on the diagonal  The  is determined by summing the values  (correctly predicted cases)  the total number of cases.  (Table 8).  and dividing them by  The factors calculated in this MDA  correctly predicted the level of richness 91.67%.  This is well  above the 68% correct that would be possible if all cases were simply assigned to the largest group.  Hair et al.  (1987)  recommend that classification accuracy should be at least 25% greater than what is possible due to chance alone (in this case 83.34%). Table 8. Validation matrix for the three group, 36 case multiple discriminant analysis of species richness in Kalamalka Lake Provincial Park during the spring of 1990 and 1991.  Actual Group (Rows)  by Predicted Group (Columns)  Predicted Group Actual Group  1  2  3  Total  1  5  0  0  5  2  1  22  1  24  3  0  1  6  7  Total  6  23  7  36  Ideally one would split the cases into two groups: the analysis group for calculating the factors; and a validation group to test the validity of the factors. This should only be 41  done if the total number of cases is over 100 We had only 36 cases in our study.  (Hair et al. 1987).  This small number of cases  meant that we had to use all 36 cases in both or analysis group and validation group.  Therefore, this prediction percentage may  overestimate the true accuracy of the functions.  These functions  may not be valid if used to analyze areas outside KLPP.  However,  they are very useful in identifying the factors important for maintaining species richness in KLPP. The group classification function coefficients and group classification constants (table 9) cases using the raw data.  can be used to classify new  The raw data are simply fit to the  equation: Z  =  +WX...WX, X 1 c+W  where  (4)  Z  =  group classification score,  c  =  group classification constant,  W  =  group classification coefficient for a given variable, and  X  =  raw data for a particular variable.  42  Table 9. Group classification constants and group classification coefficients for a three group multiple discriminant analysis of species richness in Kalamalka Lake Provincial Park during the spring of 1990 and 1991.  Group 1  Group 2  Group 3  Group Classification Constants —187.046  —207.275  —204.310  Group Classification Coefficients Variables Herbicide. Access aspect N/S aspect E/W SLOPE SOIL CLAY SAND  1.577 16.907 4.447 5.567 —2.063 0.142 8.846 4.575  3.712 16.325 8.816 7.794 —2.161 0.218 9.034 4.911  5.217 13.355 13.004 8.632 —1.289 0.250 8.729 4.790  This process is repeated for all three groups of constants and coefficients; the new case should be assigned to the group with the highest score.  This process will allow one to predict  the potential richness of new areas.  Areas which are classified  as high may be of special interest and it may be decided to give them special protection.  43  3.3  Characteristics of the Fall and Spring Prescribed Burns  Weather conditions on the days of the burns were similar (Table 10).  It was slightly warmer and drier during the spring  burn than it was during the fall burn.  The winds were light on  both days but they were almost dead calm during the spring burn. Table 10. Air temperature (C), wind speed (Kph), and relative humidity (%) during the two sets of prescribed burning treatments in Kalamalka Lake Provincial Park on November 2, 1990 and March 7, 1991.  Air temp. Wind Speed Relative Humidity  Fall burn Nov. 2, 1990  Spring Burn Mar. 7, 1991  7.0  12.0  3 to 8  0 to 5  44%  34%  The amount of available fuel was nearly the same at both sites (Table 11).  The fuel on the infested site in the spring,  however, was somewhat patchy and less abundant.  The  spring  fires were more intense than the fall fires, and the fires spread much faster in the spring. (Table 11).  This is likely due to the dryer fuels  High fuel moisture slows down combustion and the  rate of fire spread (Wright and Bailey 1982)..  44  Table 11. Fuel moisture oven dry weight), available fuel ), fuel consumed (g/m), rate of spread (rn/sec), and fire 2 (g/m intensity (Kwatts/m) for two burning treatments at two sites in Kalaiualka Lake Provincial Park.  Infested Plots  Kentucky bluegrass  Plots Fall Burn  Spring Burn  Fall Burn  Spring Burn  Fuel Moisture  107.2±9.51  30.9±1.6  88.1±5.8  25.7±2.1  Available Fuel  40.0±4.1  27.4±2.9  42.9±3.5  44.4±2.7  Fuel Consumed  12.2±15.71  21.4±1.5  15.7±2.2  34.5±4.8  Rate of Spread  0.053±0.015  0.070±0.025  0.053±0.032 0.072±0.032  Fire Intensity  77.8±60.8  177.2±109.6  153.3±47.2  1  287.8±191.8  mean±95% confidence interval Temperatures in the grass canopy,  at the soil surface, and  below the surface were similar at both sites, burn  (Fig.  5).  Plants of dry sunny places can survive heating  for 30 minutes to temperatures of 50-60 C temperatures,  during the spring  over 60 C,  (Larcher 1983).  High  occurred on all spring burn plots and  generally lasted from 1.5 to 2.0 minutes.  On all of the spring  burn plots at least one probe recorded maximum temperatures of over 200 C for 10 to 30 seconds.  Plant damage due to heating is  a product of both temperature and time.  Such extreme  temperatures, however, would certainly damage if not destroy any plant material that was not directly consumed by the fire. Temperatures at the soil surface were above 50 C for 1.0 to 3.0 minutes.  This may have been high enough to damage some seeds  that were lying on the soil surface. 45  Temperatures below the soil  Uninfested Plots  Infested Plots  160  160  160  160 -  40  140  120  100  120  120  110  110  0120  ‘20  Q0  00  —  —  —  -2.5 cm sface +10 cm  e0 70  70 60  I  60  60  40  40  20  20  20  20  ‘3  10  0  0 0  1  2  3  4  6  mirutes  8  7  8  at ter sensor  10 ii  12  3  14  16  1 __.. :”1f._..”T .:3:t_.T 012346878  mñjtes at ter sensor  iritatior  •,  101112131416 wi  ietion  Fig. 5. Temperature (C) profiles at three heights (10 cm above the soil surface, 0 cm, and 2.5 cm below the soil surface) for two sites in Kalamalka Lake Provincial Park during the spring burning treatment (March 7, 1991). The soil temperature probes initiate recording when a temperature above 60 C is sensed by the uppermost electrode. 46  changed little from ambient temperatures, and it is unlikely the burns had any direct effect on plant material below the soil surface. All four burning treatments resulted in reductions of fuel (i.e. non—overlapping confidence intervals).  The reduction in  fuel on the infested site by the fall burns was quite inconsistant.  There are two reasons why this happened. Firstly,  the fuel on that site on that date was quite moist; therefore, the fire did not burn well. on this site was patchy.  The second reason is that the fuel  The greater variability in the samples  made it harder to detect any reduction. Increased fire intensity strongly correlated (r =O.792) with 2 subsequent increases in bare soil.  The more intense spring burns  resulted in a greater increase in bare soil than the fall burns (Fig.  6).  Two models have been developed to help predict the impact of future burns in KLPP.  The first model predicts intensity of a  fire based on three easily measurable variables fuel in g/m , 2 speed in Kph).  (i.e. available  fuel moisture in percent on weight basis, and wind This model is based on the two sets of burns (12  individual plots)  conducted in KLPP.  It should be useful for  providing a rough estimate of fire intensity because it accounted for a large amount of variation in measurements of fire intensity (adjusted squared multiple R=O.856).  However,  it is only valid  when used for analyzing conditions that are similar to those  47  Infested  40  35  Uninfested  40-  I  .  spring bun  30 =  ————  25  falbu-n ed  25  Cl)  20  20 15  5 Sept. 90 Mar. 91  May 91  Sept 91  Seç 90 Mar. 91  May 91  Sept. 91  Fig. 6. changes in percent bare soil from September 1990 to September 1991 for three burning treatments (spring burned, fall burned, and unburned) at two sites in Kalamalka Lake Provincial Park. The fall burn was conducted on November 2 1990, and the spring burn was conducted on March 7 1991. Increases in percent bare soil differed (>0.05) amongst the three treatments. Fire intensity was highly correlated (r =0.792) to subsequent 2 increases in bare soil.  48  which we experienced (i.e. a grassland site with predominately fine moist fuel and light winds).  Fortunately, these are the  kinds of burns which are most commonly done in KLPP. I  =  6.504A—2.742M+20.875W  where;  (5)  I  =  Fire Intensity (Kwatts/m)  A  =  Available Fuel (g/m ) 2  M  =  Fuel Moisture  W  =  Wind speed (Kph)  (standard error  =  81.89)  (% oven dry wieght)  The second model estimates the increase in percent bare soil created by a given level of fire intensity,  Similarly, this  model should only be used for conditions which are similar to those found in this study (adjusted multiple squared R=0.792). It would be of no use for predicting such changes caused by a fire in a forested area. %BS  =  —4.019+0.1161  where;  (6)  %BS  =  I  Fire intensity (Kwatts/m)  =  percent increase in bare soil  (standard error  =  4.98)  These models will help managers determine the best conditions for burning and predict the impact of prescribed burns on the vegetation of the park.  49  Impact of Prescribed Burning on the Density of  3.4  Diffuse Knapweed  It was apparent by visual observation that the knapweed plants did not burn well in our fires; however, the temperatures in the canopy during the burns may have been high enough (i.e. over 200 C) (1968)  to damage much of the areal seed bank.  Daubenmire  showed that temperatures around 100 C in a five-minute dry  heat treatment can damage or even destroy the viability of seeds from some grass species.  During the burning the knapweed seeds  would not have been completely dry.  Therefore, they may have  been more susceptible to damage by heat. (1974)  Watson and Renney  pointed out that vegetative reproduction does not  naturally occur in diffuse knapweed populations.  This primary  reliance on seeds for reproduction may make diffuse knapweed susceptible to control by fire.  Daubenmire  (1968)  pointed out  that a fire that occurs when a plant is desicating but before seeds are released may destroy that year’s seed crop. large window of opportunity (i.e.,  There is a  several months) to use fire to  destroy the current year’s seed crop in diffuse knapweed since the majority of the seed heads remain closed in the fall and Renney 1974).  Nolan (1989)  documented the size of the areal  and soil seed banks for diffuse knapweed.  The areal seed bank  (2770 2 seeds/rn ) . was much larger than that of the soil seeds/rn ) 2 .  (Watson  (614 viable  All the seeds in the areal seed bank may not be  viable, but this nevertheless emphasizes the relative importance 50  of the areal seed bank.  If the burns were able to destroy a  significant proportion of the areal seed bank, we might expect a significant reduction in the number of diffuse knapweed plants in the next few years. two gall-fly larvae  Roze (1980)  found that the combined effect  (Urophora affinis and U. guadrifasciata)  reduced seed numbers by 80%.  Roze, however, concluded from life  table studies that this sizable reduction in seed numbers would not have a significant effect in slowing the rate of knapweed spread.  Fire, however, will affect knapweed at all stages of  growth not just seed production.  Perhaps, the combined effect of  burning and gall-fly larvae will help reduce the spread of knapweed.  The fire may have a negative effect on the gall-fly  larvae (i.e. the heat may kill them)  reducing the effectiveness  of this type of biological control.  The role of fire in  controlling knapweed may be related to its indirect effects (i.e. it may change microclimatic factors making the habitat unsuitable).  Only future monitoring of these sites will reveal  if this comes to pass. There was little temperature change in the soil.  Therefore,  temperatures may not have been high enough to reduce the viability of knapweed seeds; however, Nolan (1989)  found that 77%  of the viable seeds in the soil seed bank were within 1 cm of the surface where they would be most susceptible to damage. (1968)  Zednai  showed that the percent germination of spotted knapweed  seeds was reduced from 69 to 3% after they were subjected to flaming with a propane torch.  We might expect a similar response 51  to flaming from diffuse knapweed because it is quite closely related (i.e., same genus) characteristics  and has similar physical  (i.e. size, and texture).  Unfortunately,  Zednai  (1968) made no attempt to document the temperature and duration of his flaming treatments, so we do not know if they accurately simulated grass fire conditions.  Knapweed seeds do not have any  kind of hard protective covering; therefore,  it is possible that  they could be seriously damaged by temperatures much lower than those that damage grass seeds.  Furthermore, knapweed seeds do  not possess any kind of self burial mechanism; therefore, they would tend to remain near the soil surface where they would be most vulnerable to damage by fire.  Emergence was significantly  reduced when seeds were placed at a depth of 2.5 cm (compared to the surface or 1.3 cm depth)  and no plants emerged from seeds at  depths greater than 2.5 cm (Spears et al.  1980).  Therefore,  there is no need to be concerned about seeds which are buried below 2.5 cm if the site remains undisturbed. have also found similar results 1974); therefore,  Other researchers  (Popova 1960, Watson and Renney  if the fire only impacted seeds  and those in the top 1.3 cm of the soil,  at the surface  it could significantly  reduce diffuse knapweed emergence. Although fire may reduce knapweed plants and seáds,  it may  also encourage an invasion of knapweed on previously uninfested sites because it disturbs and creates more bare soil.  There was  initially no knapweed on the Kentucky bluegrass site and no knapweed was found during the two post—burn vegetation surveys of 52  that site.  Clearly,  in the short duration of this study, neither  spring nor fall burning encouraged an invasion of knapweed on this Kentucky bluegrass site.  Therefore,  it was decided to  analyze the knapweed density response to fire from the infested site only. An analysis of variance was done on the density of seedlings, rosettes,  bolted plants, and all of these categories combined  (Table 12).  The different burning treatments did not result in  changes in the seasonal pattern of knapweed density (date by burn interaction)  (>O.O5)  for any class of knapweed (i.e.  rosettes, or bolted plants). amongst the three treatments unburned)  (>O.O5).  seedlings,  No significant differences occurred (fall burn,  spring burn, and  There were significant differences, however,  for the density of rosettes and bolted plants on different dates (P<O.05). density.  One would expect seasonal differences in plant In all cases there was a great deal of variability  amongst the replicate plots differences shown in Fig.  (<O.Ol).  Therefore, the absolute  7 are not significant.  The spring  burn; however, did result in a large absolute reduction of rosettes (2.75 to 0.75 rosettes/rn ) while the unburned plots 2 showed an increase in rosettes  (0.40 to 1.00 2 rosettes/rn ) .  This  may lead to a significant decrease in bolted plants in the next year.  Schirman (1981)  found that seedlings emerging after May 15  had a low survival rate and almost no flower production.  A fire  in the spring may kill or damage early emerging seedlings and later emerging ones would not be able to produce seed. 53  Table 12 Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability () for a two factor analysis of variance randomized within three burn treatments (fall, spring, and control) and four dates with three replicates per cel the density of four categories of diffuse knapweed (seedlings, rosettes, mature plants, and total knapweed) in Kalamalka Lake Provicial Park during 1990 and May 1991.  Source  df  ss  •ms  F—ratio  seedling density burn treatment 2 replicates 2 date 3 date * burn 6 date*repl icates 16 missing values 6  0.898 4.536 1.531 0.485 4.423 0.000  0.449 2.268 0.765 0. 081 0.276  0.198 8.222 2.775 0.293  >0.10 <0.01 0.08 >0.10  0.207 12.436 5.219 1.120  >0.10 <0. 01 0.01 >0.10  rosette density burn treatment 2 replicates 2 date 3 date*burn 6 date*replicates 16 missing values 6  3.110 15.022 9.456 4.058 9.663 0.000  1.555 7.511 3 . 152 0.676 0.604  mature plant density burn treatment 2 replicates 2 date 3 date * burn 6 date*repl icates 16 missing values 6  3.680 10.412 3.720 0.225 4.763 0.000  1.840 5.206 1.240 0.068 0.298  0.353 17.484 4.168 0.126  >0.10 <0.01 0. 03 >0.10  total knapweed density burn treatment 2 replicates 2 date 3 date*burn 6 date*replicates 16 missing values 6  17.858 66.494 2.489 4.704 12.045 0.000  8.829 33 .247 0.830 0.785 0.763  54  0.269 43.585 1.088 1.028  >0.10 <0.01 >0.10 >0. 10  Seedlings  Rosettes  3.0 - —  — —  181  bun  — Lltuflecj >1  20  2.0  UI  V  1.5  1.0  1.0  0.5  0.5  0.0  0.0  -Os 0  2 Sept. 90 Mar. 91  4  3 May 91  5  -as  1  0  Sept 91  5  Total Knapweed  3.5  3.5  3.0  3.0  2.5  2.5  2.3  2.0 15  UI  1J  4  3  Sept 90 Mar. 91 Mey 91 Sept. 91  Bolted Plants  >.t  2  10  1.0  as  0.5  00  00  0  -05 1  2  Sept 90 Mar. 91  3  4  May 91  Sept 91  -  0  5  1  Sept. O  2 .  3  91 May 91  4  sept.  5 91  Fig. 7. Density (individuals/rn ) of four classes of diffuse 2 knapweed (seedlings, rosettes, bolted plants, and total knapweed) from September 1990 to September 1991 for three burning treatments (spring burn, fall burn, and unburned) on an infested site in Kalamalka Lake Provincial Park. The fall burn was conducted on November 2 1990, and the spring burn was conducted on March 7 1991. Differences in the density of knapweed on different dates or for different treatments are not significant (P>0.05) 55  During the course of this study neither spring nor fall burning produced reductions in knapweed density when compared to the changes in the unburned plots.  There were probably two main  reasons why we could not detect any significant changes in knapweed density.  Firstly, the high degree of variability  amongst the replicate plots within the same treatment.  Three of  the plots at the north end of the study site initially had quite dense stands of knapweed. a few knapweed plants each.  The other six plots initially had only During the summer before the first  burn, a large ponderosa pine blew down over one of these six less infested plots.  The tree was removed before the start of the  study, but the disturbance caused by its removal may have been largely responsible for the sharp increase in knapweed on this plot.  This plot was given the spring burning treatment which  reduced the knapweed on one of the other heavily infested plots. The second reason was related to the natural seasonal variability in plant density.  It would probably be best to have sampled  these plots at the same time each year for at least two years after the burn. =O.792) between 2 There was a strong positive correlation (r fire intensity and increased bare soil.  The more intense spring  burns resulted in a greater increase in bare soil than the fall burns.  This increase in percent bare soil may provide an  opportunity for knapweed to spread.  The type of disturbance may  be of critical importance, the disturbance caused by traffic may be quite different than that caused by burning.  56  Only continued  be of critical importance, the disturbance caused by traffic may be quite different than that caused by burning.  Only continued  monitoring of these sites will reveal if this actually happens. Bare soil, however, tends to warm faster and dry faster (Barbour et al.  1980).  spears et al.  (1980)  found that emergence rate and  percentage emergence decreased significantly as initial soil moisture content decreased (70 to 55%).  If a fire resulted in a  warmer and dryer soil,  it could reduce the emergence of diffuse  knapweed and possibly  shift the competitive advantage to another  species.  Berube and Myers  (Agropyron cristatum (L.)  (1982)  showed that crested wheatgrass  Geartn.) was better able to suppress  knapweed under dryer conditions.  Perhaps, the closely related  native species bluebunch wheatgrass will be better able to compete if the soil is dryer and warmer.  57  3.5  Impact of Prescribed Burning on Species Richness and Floral Diversity  On all of the spring burn plots maximum temperatures in the canopy of over 200 C were reached.  Plants of dry sunny places  can survive heating for 30 minutes to temperatures of 50—60 C (Larcher 1983).  High temperatures, over 60 C, generally lasted  from 1.0 to 2.5 minutes.  The temperatures created in these burns  were probably not hot enough to damage the viability of all native grass seeds as they generally have a tough outer coating (Daubenmire 1968).  Many of the small native annuals such as  bristlystickseed (Lappula redowskii  (Hornem.)  Greene),  flowered blue—eyed Mary (Collinsia iarviflora Lindi.), flowered fringecup  (Lithophraqma parviflora  annual Jacob’s ladder  small small  (Hook.) Nutt.), and  (Poleinonium micranthum Benth.) which are  important in creating the richness of the grasslands, may have been consumed by the fire.  As Daubenmire (1968)  pointed out, a  fire that occurs when a plant is desiccating but before seeds are released may destroy that year’s seed crop.  These small plants  will have to rely on their soil seed bank to maintain their populations.  Wright (1984)  showed that most f orbs tolerated fire  well if burned in the spring or fall.  He (Wright 1984)  concluded  that fire can enhance the diversity of forbs because many have hard seeds that can be scarified by the fire. There was very little change in the soil temperature, therefore, the temperatures in the soil were probably not hot 58  enough to destroy any seeds in the soil seed bank.  These fires  would probably not eliminate any of the species that were present in the area before the fire.  However, the relative composition  of the plant community may change,  as some species may be better  adapted to the conditions created by the burn (i.e. the increased bare soil may warm faster and dry out more quickly). All plots responded similarly over time.  Neither the fall  nor spring burning treatment significantly changed the seasonal pattern or the level of diversity or richness interaction P>O.05)  (Table 12) (Fig.  8).  (date by burn  Some new species may be  able to become established over the next few years on the bare soil created by the fire.  Schwecke and Hann (1989)  fire could be used to increase species richness.  found that  Their study in  the intermountain region of Washington state that both fall and spring burns resluted in increased species richness in the first and second years after the burns. Initially there were very significant (<O.O1) the richness and diversity of the two sites.  differences in  In the spring of  1991 the Kentucky bluegrass site was at its richest (R=13.33) most diverse (H’=O.84). richer (R=16.67)  and  At the same time the infested site was  and more diverse (H’=O.92)  (Fig. 8).  The  uninfested site was dominated by kentucky bluegrass and had a thick litter layer.  Both burning treatments resulted in  increases in the percent bare soil.  This increase in bare soil  was likely due to much of the leaf litter being consumed by the fire.  Disturbance of this type can lead to an increase in 59  species richness  (Armesto and Pickett 1985).  However, the timing  and level of the disturbance may be of critical importance in determining whether richness increases or not.  There may be an  increase in richness over the next few years especially on. the Kentucky bluegrass site.  An increase in the amount of bare soil  may allow some of the smaller plants to become established. Table 13 Sources of variation, degrees of freedom (df), sums of squares (ss), mean squares (ms), F—ratio, and associated probability () for a three factor analysis of variance randomized within two sites, three burn treatments (fall, spring, and control), and four dates with three replicates per cell for species richness and the Shannon—Weinner diversity index in Kalamalka Lake Provicial Park during 1990 and 1991.  Source  df  .  site 1 burn treatment 2 site*burn 2 replicates 12 date 3 date*site 3 date*burn 6 date*burn*site 2 date*replicates 28 missing values 12  ss  ms  F—ratio  P  species richness 234.722 35.964 52.481 245.457 546.404 12.031 36.355 9.246 203.812 0.000  234.722 17.982 26.241 20.464 182.135 4.010 6.059 4.512 7.280  11.470 0.879 1.181 2.811 25.019 0.551 0.832 0.620  <0.01 >0.10 >0.10 0.02 <0.01 >0.10 >0.10 >0.10  Shannon—Weinner diversity index site 1 burn treatment 2 site*burn 2 replicates 12 date 3 date*site 3 date*burn 6 date*burn*site 2 date*replicates 28 missing values 12  1.184 0.199 0.408 1.142 1.099 0.113 0.082 0.090 0.888 0.000  1.184 0.100 0.204 0.095 0.550 0.056 0.014 0.045 0.032  60  12.460 1.049 2.149 2.996 17.326 1.780 0.431 1.424  <0.01 >0.10 >0.10 0.02 <0.01 >0.10 >0.10 >0.10  infested site  uninfested site  25  25  20  20  [5  15  10  10  5  5  U,  0  1  0  2  4  3  Sept. 90 Mar. 90 May 91  0  5  0  Sept. 91  1  2  3  4  Sept. 90 Mar. 90 May 91  15  1.5  1.0  1.0  0.5  0.5  5  Sept. 91  a,  0 > C,,  a, >  U U,  C  0.0  0  1 Sept. 90  Fig.  2  3  Mar. 90  May 91  4  0.0  5  0  Sept. 91  2 Sept. 90  3  Mar. 90 May 91  4  5  Sept. 91  species richness and Shannon’s diversity index from 8. September 1990 to September 1991 for three burning treatments (spring burn, fall burn, and unburned) at two sites (a diffuse knapweed infested site, and an Kentucky The fall bluegrass site) in Kalamalka Lake Provincial Park. the spring burn burn was conducted on November 2 1990, and Neither burning treatment was conducted on March 7 1991. altered (P>O.05) the seasonal pattern or final level of richness or diversity.  61  There were also very significant differences in richness and diversity over the course of this study (i.e. significant date effect <O.O1) (Fig.  8).  However, one would expect seasonal  differences in diversity and richness.  62  3.6  Summary of Findings  All of the research objectives for this project have been achieved.  The results have been briefly summarized below and  numbered according to the specific research objective to which they relate (section 1.0). 1.  Diffuse knapweed has been found throughout the park in all six of the major grassland ecosystem association/seral stages.  The accessability of a site seems to be the most  important factor in creating a diffuse knapweed infestation in KLPP. 2.  The Douglas—fir  —  common snowberry northerly aspect habitat  in the shrub-herb stage (DS1) was more rich and diverse than all other ecosystems studied.  Sites that had been treated  with herbicide were the least rich and lowest in diversity, the number of years since herbicide treatment was the most important factor in determining the level of richness. 3.  Measures of fuel moisture, fuel availability, and fire intensity as well as soil temperature profiles were taken to document the characteristics of the burn. were more intense than the fall burns,  The spring burns  and the subsequent  increase in bare soil was highly correlated to fire intensity. 4.  During the course of this study neither spring nor fall burning produced reductions in knapweed density compared to unburned plots.  Continued monitoring is required to 63  determine the longer term affect of the burns on the knapweed population. 5.  Neither burning treatment altered the seasonal pattern or level of diversity or richness by the end of this study.  It  is unlikely that the use of prescribed burning will result in the loss of any species that was previously present on the site. 6.  Management guidelines for Kalamalka Lake Provincial Park are presented in detail in the next section.  64  4.0  MANAGEMENT IMPLICATIONS  One of the most important goals of this project was to develop some management alternatives for Kalamalka Lake Provincial Park.  A number of concerns regarding the management  of vegetation have been identified.  Perhaps the most serious  concern has to do with the spread of diffuse knapweed.  There are  a number of important questions to which park managers require answers.  How bad is the knapweed problem in the park?  is it spreading?  Moreover,  How fast  is the spread of diffuse knapweed  affecting the richness and diversity of the natural vegetation? In order to get a handle on this problem, there was a need to collect some base line data on the density of diffuse knapweed and the levels of richness and diversity in all areas of the park.  Other important concerns related to fuel accumulation and  the associated risk of wildfire.  Now that cattle grazing is no  longer appropriate in the park, there is a need to reduce this build up of fuel.  Fire seems to be a natural part of the  grassland ecosystem, but how would it affect the spread of diffuse knapweed and/or the levels of species richness and floral diversity?  4.1  Diffuse Knapweed in Kalamalka Lake Provincial Park  Diffuse knapweed was found throughout the park in all types of grassland ecosystem associations/seral stages.  65  It was very  dense along some of the main trails.  However,  very sparse in the middle of most sites.  it appeared to be  Despite the fact that  no herbicides were used on any of the study sites during this two—year study, no increase in knapweed density was found (>O.O5). There was no significant correlation between the density of any class of diffuse knapweed (seedlings, rosettes, mature plants,  or total knapweed)  diversity.  and either species richness or floral  Apparently diffuse knapweed is not yet a threat to  the species richness or floral diversity in KLPP.  This is no  reason to stop trying to control diffuse knapweed in KLPP. However, there may not be a need for any immediate or drastic action. The most important factor in the spread of diffuse knapweed seems to be the level of accessibility of any area in question (Fig.  9).  Percent bare soil may be an important co-factor for  increasing the density of diffuse knapweed in that it provides an opportunity for new plants to become established.  The most  appropriate action for the parks management to take in order to control the spread of diffuse knapweed would be to close trails that have a lot of knapweed on them and/or fence of f areas that have a lot of bare soil.  Posting signs that explain how knapweed  is spread and encouraging people not to travel in infested areas may also help.  66  5 >1 4 J  U)  A  a)  3  a)  2  U) ‘4-i -‘-4  1  L  0 0  Fig. 9. The level of accessability is the most important factor in spread of diffuse knapweed 2 (individuals/ ) . Sites near. rn roads or main trails were given a value of 1.0 for the access variable; conversely, sites which were more than 500 rn from the nearest trail or that could -only be accessed by climbing a steep slope were given a value of 4.0. The amount of bare soil may be an important co—factor. 67  4.2  Species Richness and Floral Diversity in Kalamalka Lake Provincial Park  There were significant differences in the species richness and floral diversity amongst the six different ecosystem association/ seral stages that were studied. (average of 36 species)  The richest  and most diverse (Shannon’s Diversity  index, H’=1.434) community was the early successional stage forest ecosystem DS1  (Douglas fir  —  Saskatoon).  communities are on north—facing slopes.  These  They are generally quite  inaccessible, and they have never been sprayed with herbicide. The moist grassland community BW2  (Kentucky bluegrass-Bluebunch  wheatgrass/ successional stage fair) number of species  (R=15.33)  had the lowest average  and lowest diversity (H’=O.871).  All  of these communities have been sprayed with herbicide in the past two or three years.  Furthermore, they are dominated by a thick  sward of Kentucky bluegrass, which may crowd out many of the smaller species.  The spraying of broad—leaved herbicides may  help Kentucky bluegrass establish this thick sward as it is not affected by them. The most important factor in reducing the species richness and floral diversity was the herbicide treatment.  The number of  years since a site was treated was very strongly correlated to richness (Fig.  10).  The areas that had most recently been  sprayed had the lowest levels of diversity.  As the number of  years since spraying increased the level of richness and diversity increased.  Areas that had never been sprayed were the 68  richest and most diverse.  The density of diffuse knapweed had  little impact on the species richness of an area.  Low levels of  diffuse knapweed were associated with a slight reduction in species richness.  This is likely a reflection of incomplete  knapweed control from the herbicide.  Areas that had knapweed may  have been sprayed reducing the richness but not completely controlling the knapweed, therefore,  low levels of knapweed would  be associated with lower levels of richness.  A northern aspect  and sandy soil also seem to be important factors in creating high levels of richness and diversity.  There should be a moratorium  put on the use of herbicides in the park to maintain the richness and diversity of the native vegetation in KLPP.  There must be  other methods found to control diffuse knapweed.  4.3  The Use of Prescribed Burning in Kalamalka Lake Provincial Park  The spring burns were more intense because of dryer fuels. Neither the spring nor fall burns encouraged an invasion of diffuse knapweed into a previously uninfested area.  However,  there was no significant reduction in the density of any class of diffuse knapweed associated with either spring or fall burn. Prescribed burning will certainly reduce the fuel load and the associated risk of a wild fire.  It may not reduce the density of  diffuse knapweed, but it does not seem to encourage knapweed. Neither burning treatment significantly affected the seasonal pattern or level of species richness or floral 69  0  I-.. I-’- 1-3  CD CD 0  •  )  I-i.  CD U)  I-.. CDCD U)  CD  U).  rICD I-’. CD U) rt  OU)CD r1 Q) rI o I-’- CD  d i- rt 1  I-I  rIrI-rI ctO CD  I-•.  jJU)  CD  CDC) I-CD  CDOU) CD ti I-’.  CD I-’D) ø)’< 1 U)  h<  CD CD 0  r10  I-flU)  1-’- U) CD -h 1  OCD  CD rtO U) < F CD Qj U)  -  ‘1  TJ I-..  0  4 0  species 0  richness  diversity by the end of this study.  Apparently, prescribed  burning does not adversely affect species richness or floral diversity.  There seems to be no reason that prescribed burns  should not be used to obtain other management goals.  It should  be noted that no measurements were taken to determine the effect of fire on the biological control agents of diffuse knapweed. the fire had a negative impact on these organisms, increase in knapweed over the longer term.  If  it may lead to  However, if  prescribed burning is only used in uninfested areas,  it will not  have a strong adverse affect on the biological control agents (i.e. no knapweed no biological control organism).  Only  continued monitoring of the burn sites will reveal the long term effect of prescribed burning on diffuse knapweed. The amount of bare soil on a site increased with the intensity of fire.  This increase in bare soil, especially  associated with the spring burns, may lead to an increase in the density of diffuse knapweed because knapweed seems to thrive on disturbance.  However, this increase in bare soil may also lead  to an increase in species richness and floral diversity, especially on the uninfested plots which where previously dominated by Kentucky bluegrass and covered in a thick layer of leaf litter.  71  5.0  LITERATURE CITED  Armesto, J.J., and S .T.A. Pickett. 1985. Experiments on disturbance in old—field communities: impact on species richness and abundance. Ecology 66:230-240. Barbour, M.G., J.H. Burke, and W.D. Pitts. 1980. Terrestrial Plant Ecology. The Benjamin/Cummings Publishing Company, Inc., Menlo Park, Ca. 604pp. Berube, D.E., and J.H. Myers. 1982. Suppression of knapweed invasion by crested wheatgrass in the dry interior of British Columbia. J. Range Manage. 35:459—461. Blaisdell, J.P. 1953. Ecological effects of planned burning on sagebrush—grass range on the Upper Snake River Plains. U.S. Dep. Agr. Tech. Bull. 1075. 39pp. Brown, A.A., and K.P. Davis. 1973. Forest fire control and use. McGraw Hill Book Co. New York, New York. Bunting, S.C., B.M. Kilgore, and C.L. Bushey. 1987. Guidelines for prescribed burning sagebrush—grass rangelands in the northern Great Basin. USDA, Forest Service. Gen. Tech. Rep. INT—231. Cawker, K.B. 1983. Fire history and grassland vegetation change: three pollen diagrams from southern British Columbia. Can. J. Bot. 61:1126—1139. Chicoine, T.K., and P.K. Fay. 1984. Seed longevity of spotted knapweed in Montana soils. pages 22—26. jfl Proc. of the Knapweed Symp. Plant and Soil Sci. Dept. and Cooperative Extension Services Montana State University. Bull. 1315. Clark, R.G., and E.E. Starkey. 1990. Use of prescribed fire in rangeland ecosystems, p 81—94 Waisted, J.D., S.R. Radosevich, and D.V. Sandberg (eds), Natural and prescribed fire in the Pacific Northwest forests. Oregon State University Press, Corvalis. Conrad, C.E., and C.E. Poulton. 1966. Effects of a wildfire on Idaho fescue and bluebunch wheatgrass. J. Range Manage. 19: 138—141. Cranston, R., A.H. Bawtree, and G.R. Keay. 1983. Knapweed in British Columia: a description of the problem, efforts to contain its spread and the economics of control. Unpublished report. British Columbia Ministry of Agriculture and Food. Victoria, B.C.  72  Cranston, R. 1984. Knapweed in British Columbia. pages 4-7 Proc. of the Knapweed Symp. Plant and Soil Sci. Dept. and Cooperative Extension Services Montana State University. Bull. 1315. Daubenmire, R. 1968. Ecology of fire in grasslands. Ecol. Res. 5:209—266.  Advances in  Daubenmire. R. 1970. Steppe Vegetation of Washington. Washington State University Cooperative Extension, and the U.S.D.A. Pullman, Wa. l3lpp. Daubenmire, R. 1975. Plant succession in abandoned fields, and fire influences, in a steppe area in southeastern Washington. Northwest Sci. 49:36-48. Davis, E.S., and P.K. Fay. 1988. Seed longevity of spotted knapweed (Centaurea maculosa Lam.). page 75. Proc. Western Soc. Weed Sci. Fresno Calf. Vol. 41. Environment Canada (Atmospheric Environment Service). Canadian climate normals 1951—1980: temperature and precipitation. British Columbia  1984.  Gee, G.W., and J.W. Bauder. 1986. Particle size analysis Arnold Klute (ed), Methods of Soil Analysis, Part 1 Physical and Mineralogical Methods 2nd. ed. Amer. Soc. Agron. Madison, Wis. Gruell, G.E., J.K. Brown, and C.L. Bushey. 1986. Prescribed fire opportunities in grasslands inbaded by Douglas—fir: state of the art guidelines. USDA Forest Service. Gen. Tech. Rep. INT-l98 Hamlen, C., and J.A.G. Hansen. 1984. The measurement of knapweed in British Columbia in 1983. Agriculture Canada, Regional Development Branch. (unpublished report). Hair, J.F., R.E. Anderson, and R.C. Tatham. 1987. Multivariate Data Analysis, 2nd ed. Macmillan, New York. Hitchcock, C.L., and A. Cronquist. Northwest: an illustrated manual. Press. Seattle.  1973. Flora of the Pacific University of Washington  Kelly, C.C., and R.H. Spilsbury. 1949. Soil survey of Okanagan and Similkameen Valleys: British Columbia. Can. Dep. Agr. British Columbia Surv. Rep. 3. Kings Printer. Ottawa, Ontario, Canada Lass, L.W., and R.H. Callihan. 1989. Spotted knapweed control in pasture. pages 172-173 .jj Proc. Western Soc. Weed Sci. Honolulu Ha. Vol. 42.  73  Larcher, W. 1973. Verlag. Berlin.  Physiological Plant Ecology,  2nd.  ed.  Springer-  Lea, E.C., D.A. Demarchi, and R.E. Maxwell. 1990. Ecological Resources of Kalamalka Lake Provincial Park. Wildlife Branch. B.C. Mm. of Environment. Little, T.M., and F.J. Hills. 1978. Agricultural experimentation design and analysis. John Wiley and Sons. New York Mathews, E.E. 1984. Fire on Montana’s rangelands: pilot program. Western Wildi. 10 (3):16—l9.  a successful  Ministry of Lands, Parks, and Housing. 1984. Kalamalka Lake Provincial Park. The report of the public advisory council.Queen’s printer for B.C., Victoria. 4Opp. Nolan, D. 1989. Germination Characteristics of Centaurea diffusa and C. maculosa. M.Sc. Thesis, University of BritishColumbia, Vancouver. 207pp. Patton, B.D., M. Hironka, and S.C. Bunting. 1988. Effect of burning on seed production of bluebunch wheatgrass, Idaho fescue, and columbia needlegrass. J. Range Manage. 41: 232—234. Pitt, M.D., and B.M. Wikeem. 1990. Phenological patterns and adaptations in an Artemesia/Agropyron plant community. J. Range. Manage. 43:350—357. Popova, A.Y. 1960. Centaurea diffusa Lam., A steppe—pasture weed in the Crimea. Bot. Zhur. 49:863-865. Roze, L.D. 1980. The Biological Control of Centaurea diffusa Lam. and . maculosa Lam. by Urophora affinis Frauenfeld and U. puadrifasciata Meigen. Ph.D. Thesis, University of British Columbia, Vancouver. 208pp. Schirman, R. 1981. Seed production and spring seedling establishment of diffuse and spotted knapweed. J. Range Manage. 34:45—47. Schwecke, D.A., and W. Hann. 1989. Fire behaviour and vegetation response to spring and fall burning on the Helena National Forest. Prescribed fire in the intermountain region, Symp. Proc. Washington State University, Pullman. Shelley, R.L., and B.F. Roche. 1982. Rehabilitation of spotted knapweed infested rangeland in northeastern Washington. page 31 .jfl Proc. Western Soc. Weed Sci. Denver, Cob. Vol. 35.  74  Spears, B.M., S.R. Rose, and W.S. Belles. 1980. Effect of canopy cover, seeding depth, and soil moisture on emergence of Centaurea maculosa and C. diffusa. Weed Res. 20:87—90. Tyser, R. W., and C. H. Key. 1988. Spotted knapweed in natural area fescue grassland: an ecological assessment. Northwest. Sci. 62:151—160. Uresk, D.W., J.F. Cline, and W.H. Rickard. 1976. Impact of wildfire on three perenial grasses in south—central Washington. J. Range Manage. 29:309—310. Watson, A.K. 1977. Effects of grazing, cutting, burning, irrigation and cultivation on diffuse and spotted knapweed. pages 74-79 j Proc. Knapweed Symp. British Columbia Plant Science Lead Committee. Watson, A.K. and A.J. Renney. 1974. The biology of Canadian weeds. 6. Centaurea diffusa and C. maculosa. Can. J. Plant Sci. 54:687—701. Wilms, W., and A.W. Bailey. 1980. Effects of clipping or burning on some morphological characteristics of Agropyron spicatum. Can. J. Bot. 58: 2309—2312. Wright, H.A. 1985. Effects of fire on grasses and forbs in sagebrush—grass communities. Rangeland fire effects, Symp. Proc. University of Idaho, Boise. Wright, H.A., and A.W. Bailey. 1982. Fire ecology, United States and southern Canada. John Wiley & Sons, New York. 5Olpp. Zednai, J.G. 1968. Studies on the biology and control of Centaurea maculosa Lam. B.Sc.(Agric.) Thesis, University of British Columbia, Vancouver. 6Opp.  75  6.0 APPENDIX I  Location and type of diversitY/kflaPWeed transects in Kalamalka Lake provincial Park  Hey  L4Frii2  —  •ø)Q  JFf2  t.JFn  —  a.  DS1 L.J3  -•1--  oer’  BWZ  1 i ne I.  i’cd  76  0  C  (V -4. V)  0  C  I  1,’  CD  (?  CD  ‘4,  CD  I:-’ p)  p)  I—I  ) H  (n  I-art  H I-’  0  0 I-b  0  C) p) r1  0  H  H  H  0  8.0 APPENDIX III Description of study sites used in divers ity/knapweed study  Transect  #  Ecosystem Association  Aspect  Access 1 code  year of last herbicide treatment _2  1  WFm2  S  2  2  WFm2  W  2  3  WFm2  NW  3  4  WFf 2  W  3  5  WFf 2  NW  3  6  WFf2  W  3  7  WFm3  NE  2  1989  8  WFm3  S  1  1989  9  WFm3  W  4  10  DS1  N  4  11  DS1  N  4  12  DS1  NW  3  13  WS3  W  2  14  WS3  S  4  15  WS3  S  3  16  BW2  W  1  1988  17  BW2  W  1  1988  18  BW2  W  1  1989  1  1=very easy,  2  no known herbicide treatment  2=easy,  3=remote,  78  4=very remote  —  1989  9.0 APPENDIX IV Vascular Plants of Kalamalka Lake Provincial Park  Scientific name  Conuuon name  Grasses and Grasslike 3 Acrropyron cristatum Agropyron smithii Agropyron spicatum Bromus tectorum Dactylis glomerata Elvmus cinerius Festuca idahoensis Festuca octoflora Festuca scabrella Hordeum lubatum Koleria cristata Phleum pratense Poa bulbosa Poa nervosa Poa pratensis Poa sandbergii Setaria viridis Stipa comata Stipa occidentalis Carex praegracilis  nelsonii  crested wheatgrass western wheatgrass ‘bluebunch wheatgrass cheat grass orchard grass giant ryegrass Idaho fescue six—weeks fescue rough fescue foxtail barley Junegrass common timothy bulbous bluegrass Wheeler’ s bluegrass Kentucky bluegrass Sandberg’ s bluegrass green foxtail needle and thread Nelson’ s needlegrass sedge  Forbs Achillea millefolium Agroseris heterophylla Antennaria microphylla Antennaria parvaflora Antennaria puicherrima Arabis holboèllii Arenarea serpyllifolia Astragalus miser Balsamorhiza saciittata Brodiaea douglassii Camelina microcarpa Calochortus macrocarpus Capsella bursa—pastoralis  yarrow annual agroseris rosy pussytoes Nuttall ‘ s pussytoes showy pussytoes Holboel ‘ s rockcress thyme-leaf sandwort timber milkvetch arrow—leaf balsamroot Douglas’ brodiaea hairy falsef lax sagebrush mariposa lily shepard’s purse  Nomenclature follows Hitchcock and Cronquist (1973) 79  Centaurea diffusa Claytonia lanceolata Collomia linearis Colinsia parviflora Crepis atrabarba Crepis tectorum Delphinium nutilianum Dodocantheon ullchellum Draba verna Erodium cutarium Erigeron compositus Erigeron filifolius Ericeron pumilus Fragaria virginiana Fritilaria lanceolata Fritilaria pudica Geum triflorum Heuchera cylindrica Hydrophyl lum cap itatum Hypericum perforatum Lactuca scariola Lappula redowski Lepidum nitidum Lepidum perfolatum Lespuerella douglassii Lithophraqma parviflora Lithosperma ruderale Lomatium dissectum Lomatium triternatum Luinus sericeus Microsteris gracilis Montia linearis Opuntia fragilis Opuntia polyacantha Phacelia linearis Plantago malor Plantago patagonica Polemonium micranthum Polygonium douglasii Potentilla gracilis Ranunculus cilaberrimus Saxifrãga intecrrefolia Sedüm lanceolatum Senecio canus Senecio integerrimus Sisymbrium altissimum Sonchus arvensis Taraxacum officinale Traciopogon dubius  diffuse knapweed lanceleaf springbeauty narrow—leaf collomia small flowered blue-eyed Mary slender hawksbeard rooftop hawksbeard upland larkspur dark throat shooting star spring whitlow-wort Alfilaria dwarf mountain fleabane thread-leaf fleabane shaggy fleabane wild strawberry mission bells fritilaria three flowered avens round leaf alumroot ballhead waterleaf St. John’s wort prickly lettuce western stickseed shinning pepperweed clasping pepperweed Columbia bladderpod small flowered fringecup 1 exnmonweed fern—leaved lomatium nine leaf lomatium silky lupine microsteris narrow leaved montia brittle cactus prickly-pear cactus thread leaf phacelia common plantago indian—wheat annual Jacob’ s—ladder Douglas’ knotweed sulphur cinquefoil sagebrush buttercup swamp saxifrage lance leaved stonecrop wooly groundsel western groundsel tall tumble mustard perrenial sow-thistle dandelion yellow salsify  80  Veronica arvensis Woodsia scopulania Zycradenus venenosus  common speedwell woodsia death camus  Trees and Shrubs Amelanchier alnifolia Artemisia fricrida Berberis apuifolium Eriocronum heracleoides Pinus ponderosa Prunus virgmniana Pseudotsucra menziesii Rosa nutkana Symphoricarpos albus  Saskatoon pasture sage shining Oregongrape parsnip—flowered buckwheat ponderosa pine Chokecherry Douglas—fir Nootka rose snowberry  81  10.0  APPENDIX V  Heat Yield Constant for Fire Intensity Equation 4  Fuel Moisture Content (% dry weight)  Heat Yield Constant (Kj/Kg)  10 20 40 60 80 100  15,978 15,629 14,932 14,235 13,535 12,850  adapted from Brown and Davis 1973 4 82  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0086724/manifest

Comment

Related Items