Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Using product-specific fuelwood yields to assess economic viability : a case study of farm-based Gliricidia… Baker, Kahlil 2012

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

Item Metadata

Download

Media
24-ubc_2012_fall_baker_kahlil.pdf [ 2.11MB ]
Metadata
JSON: 24-1.0073182.json
JSON-LD: 24-1.0073182-ld.json
RDF/XML (Pretty): 24-1.0073182-rdf.xml
RDF/JSON: 24-1.0073182-rdf.json
Turtle: 24-1.0073182-turtle.txt
N-Triples: 24-1.0073182-rdf-ntriples.txt
Original Record: 24-1.0073182-source.json
Full Text
24-1.0073182-fulltext.txt
Citation
24-1.0073182.ris

Full Text

Using product-specific fuelwood yields to assess economic viability: a case study of farm-based Gliricidia sepium and Caesalpinia velutina plantations in Nicaragua by  Kahlil Baker  B.A. (Hons.) Concordia University, 2008  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE  in  The Faculty of Graduate Studies  (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)  September 2012  © Kahlil Baker, 2012  Abstract Non-industrial fuelwood plantations are commonly seen as a means of improving rural livelihoods while helping to meet energy demand. However, for smallholders to invest in the establishment of fuelwood plantations, economic viability is required. Two of the greatest sources of uncertainty in evaluating the economic viability of fuelwood plantations are the effects that market-specific requirements can have on the stumpage price a plantation owner can hope to receive and the lack of appropriate growth and yield information. The primary objective of this thesis was to determine if Caesalpinia velutina and Gliricidia sepium fuelwood plantations in Nicaragua could be economically viable in the smallholder context if sold within the market place. To improve the accuracy of the economic viability assessment, a novel approach was used that forecasted fuelwood yields by market-specific product segments, thereby accounting for the effects of market requirements on differential revenues and costs. Data on market demand, product segment dimensions and prices were collected by measuring fuelwood logs and by interviewing fuelwood consuming business owners. To forecast fuelwood log volume by product segments, species-specific yield models based on three separate sub-models were developed: 1) mean diameter at breast height (DBH) predicted over time; 2) mean height as a function of mean DBH; and 3) taper as a function of mean DBH and mean height. Mortality was assumed to be zero, following establishment mortality. To assess economic viability, information on costs, discount rates, market requirements and fuelwood yields by product segment were combined using the net present value (NPV) and the internal rate of return (IRR). It was concluded that fuelwood plantation yields according to product-specific requirements were essential for the economic viability analysis. In the context of this study, farm-based Caesalpinia velutina and Gliricidia sepium fuelwood plantations could be economically viable over longer rotations. However, barriers to entry such as access to capital and the need for reaching economies of scale made it unlikely that fuelwood plantations could be economically viable for smallholders without institutional support.  ii  Preface A version of Chapter 2 has been submitted for possible publication: Baker, K., Bull, G.Q., and LeMay, V.M. The Importance of Product Segmentation and Merchantability Requirements in the Assessment of Market Potential for Fuelwood Plantations: A Case Study of the Rosquilla Industry in Madríz, Nicaragua. I conducted the data analysis and wrote most of the manuscript. Co-authors provided advice on methodology and made editorial comments as required. This chapter involved questionnaires, which required approval by the Behavioural Research Ethics Board. UBC BREB NUMBER: H11-03211. A version of Chapter 3 has been submitted for possible publication: Baker, K., LeMay, V.M., and Bull, G.Q. Fuelwood Log Volume by Size: Estimates for Gliricidia sepium and Caesalpinia velutina using a Nonlinear Mixed-Effects Taper Model. I conducted the data analysis and wrote most of the manuscript. Co-authors provided advice on methodology and made editorial comments as required. Dr. Valerie LeMay provided a significant proportion of the SAS code used in the statistical analysis. A version of Chapter 4 has been prepared for possible publication: Baker, K., Bull, G.Q., and LeMay, V.M. Using Product-Specific Fuelwood Yields to Assess Economic Viability: A Case Study of Farm-Based Gliricidia sepium and Caesalpinia velutina Plantations in Nicaragua. I conducted the data analysis and wrote most of the manuscript. Co-authors provided advice on methodology and made editorial comments as required.  iii  Table of contents Abstract .........................................................................................................................ii Preface ......................................................................................................................... iii Table of contents........................................................................................................... iv List of tables ................................................................................................................. vii List of figures ................................................................................................................ ix Acknowledgments ......................................................................................................... xi Dedication .................................................................................................................... xii 1.  2.  Introduction ........................................................................................................... 1 1.1.  Objectives .......................................................................................................................... 2  1.2.  Case description and silvics of species ......................................................................3  The importance of product segmentation and merchantability requirements in  the assessments of market potential for fuelwood plantations: a case study of the rosquilla industry in Madríz, Nicaragua....................................................................... 6 2.1.  Introduction ...................................................................................................................... 6  2.2.  Methods .............................................................................................................................. 8  2.2.1.  Industry description and study area ..................................................................................... 8  2.2.2.  Interviews ........................................................................................................................................ 10  2.2.3.  Fuelwood measurements ......................................................................................................... 11  2.2.4.  Market analyses ............................................................................................................................ 12  2.3.  Results ............................................................................................................................. 13  2.3.1.  Merchantability requirements and product segmentation .................................... 13  2.3.2.  Product segment procurement and preferences ......................................................... 19  2.3.3.  Fuelwood demand by value .................................................................................................... 21  2.4.  Discussion ....................................................................................................................... 24  2.4.1.  Merchantability requirements and consumer preferences ................................... 24  2.4.2.  Fuelwood prices ............................................................................................................................ 25  2.4.3.  The potential for fuelwood plantation .............................................................................. 26  2.5.  Conclusions .................................................................................................................... 27  iv  3.  Fuelwood log volume by size: estimates for Gliricidia sepium and Caesalpinia  velutina using a nonlinear mixed-effects taper model ................................................ 29 3.1.  Introduction ................................................................................................................... 29  3.2.  Methods ........................................................................................................................... 30  3.2.1.  Field measurements ................................................................................................................... 30  3.2.2.  Taper model .................................................................................................................................... 31  3.2.3.  Volume estimates ......................................................................................................................... 32  3.2.4.  Model evaluation .......................................................................................................................... 33  3.2.5.  Fuelwood volumes by tree size ............................................................................................. 34  3.3.  3.3.1.  The taper models .......................................................................................................................... 36  3.3.2.  Volume by product segment................................................................................................... 43  3.4.  4.  Results and discussion ................................................................................................ 36  Conclusions .................................................................................................................... 45  Economic viability and yield ............................................................................... 46 4.1.  Introduction ................................................................................................................... 46  4.2.  Case description ............................................................................................................ 47  4.3.  Methods ........................................................................................................................... 48  4.3.1.  Growth and yield models ......................................................................................................... 48  4.3.1.1.  Data description .......................................................................................................................... 48  4.3.1.2.  Model components ..................................................................................................................... 49  4.3.1.3.  Diameter at breast height (DBH) ........................................................................................... 50  4.3.1.4.  Height prediction ........................................................................................................................ 51  4.3.2.  Volume by product segment................................................................................................... 52  4.3.3.  Methods for economic viability assessment .................................................................. 53  4.4.  4.3.3.1.  Input data for economic assessment .................................................................................... 55  4.3.3.2.  Sensitivity analysis ..................................................................................................................... 58  Results ............................................................................................................................. 58  4.4.1.  Mean DBH and height models................................................................................................ 58  4.4.2.  Merchantable yield by product segment ......................................................................... 61  4.4.3.  Economic viability ....................................................................................................................... 63  4.4.3.1.  Stumpage prices .......................................................................................................................... 63  4.4.3.2.  Internal rate of return (IRR) and net present value (NPV) ........................................... 65  4.4.4. 4.5.  Sensitivity analysis ...................................................................................................................... 67  Discussion ....................................................................................................................... 68  v  4.5.1.  Growth and yield models ......................................................................................................... 68  4.5.2.  Economic viability of fuelwood plantations ................................................................... 69  4.5.3.  Economic viability in the smallholder context and the role of the buyer ....... 71  4.6.  5.  Conclusions .................................................................................................................... 72  Final conclusions .................................................................................................. 75 5.1.  Future research ............................................................................................................. 76  Works cited .................................................................................................................. 78 Appendix ...................................................................................................................... 91 Appendix A: Interview .............................................................................................................. 91  vi  List of tables Table 1 - Proportions of bakeries surveyed by community and municipality. ................. 11 Table 2 - List of species that rosquilla bakeries purchase and the number of times they were cited by respondents. ..................................................................................... 15 Table 3 - Species that rosquilla bakeries prefer to purchase and the number of times they were cited by respondents. ..................................................................................... 16 Table 4 - Species considered non-merchantable and the number of times they were cited by respondent. ....................................................................................................... 16 Table 5 - Descriptive statistics of product segment dimensions. ..................................... 18 Table 6 - Values and volumes of rosquilla fuelwood per municipality. .......................... 24 Table 7 - Descriptive statistics for C. velutina and G. sepium data. ................................ 31 Table 8 - Taper model estimates for Caesalpinia velutina and Gliricidia sepium. .......... 37 Table 9 - Fit statistics for diameters using taper models for C. velutina and G. sepium. .. 39 Table 10 - Volume bias tables for C. velutina and G. sepium for different height class... 40 Table 11 - Volume bias tables for C. velutina and G. sepium for different DBH classes. 41 Table 12 - Distribution of merchantable volume (m3) by product segments for DBH height combinations of C. velutina and G. sepium trees.......................................... 44 Table 13 – Plot-level descriptive statistics for all permanent and temporary sample plots of Caesalpinia velutina and Gliricidia sepium........................................................ 49 Table 14 – Establishment and maintenance costs and other input variables for the fuelwood plantations under the two different scenarios. ......................................... 56 Table 15 - Prices and harvest-related costs for scenarios 1 and 2. ................................... 57 Table 16 – Frequency and timing of fuelwood plantation costs for scenarios 1 and 2. .... 57 vii  Table 17 – DBH and height yield models for C. velutina and G. sepium. ....................... 61 Table 18 – Range of potential harvest prices for scenario 1. ........................................... 63 Table 19 - Range of potential harvest prices for scenario 2. ........................................... 64 Table 20 – Sensitivity analysis: effects of change in the economic parameters on the NPV with a 10-year harvest. ........................................................................................... 68  viii  List of figures Figure 1 – Nicaraguan fuelwood demand and supply. A breakdown of the market is boxed-in by the dotted line. ...................................................................................... 8 Figure 2 - Map of Nicaragua in Central America with the province of Madríz in dark grey. ................................................................................................................................ 9 Figure 3 - Man unloads fuelwood near dome shaped ovens in a rosquilla bakery. .......... 10 Figure 4 - Small-end diameter cut-off point between burusca and rolliza fuelwood logs. .............................................................................................................................. 18 Figure 5 - Percentage of bakeries that consume the various fuelwood product segments per community. ...................................................................................................... 19 Figure 6 - Traditional fuelwood purchasing unit transported to market........................... 20 Figure 7 - Average price per purchasing unit by fuelwood product segment, defined by the small-end diameter of the log, in the wet and dry season. Burusca log diameters > 0.4 cm; rolliza logs > 2.5 cm; and revuelto is a mix of both product segments. ... 21 Figure 8 - Average price per unit volume by fuelwood product segment, defined by the small-end diameter of the log, in the wet and dry season. Burusca log diameters > 0.4 cm; rolliza logs > 2.5 cm; and revuelto is a mix of both product segments. ...... 22 Figure 9 - Average dry season prices by fuelwood product segment, defined by the smallend diameter of the log, in 2007 and 2012. Burusca log diameters > 0.4 cm; rolliza logs > 2.5 cm; and revuelto is a mix of both product segments. .............................. 23 Figure 10 - Tree volume separated into product segments. ............................................. 35 Figure 11 - White noise residuals plotted against predicted diameters for A) C. velutina and B) G. sepium. .................................................................................................. 37  ix  Figure 12 - Relative diameter (dij/DBHi) versus relative height (hij/Hti) for: A) measured diameters of C. velutina; B) predicted diameters for C. velutina; C) measured diameters for G. sepium; and D) predicted diameters for G. sepium. ...................... 42 Figure 13 – White noise residuals plotted for DBH yield model for C. velutina. ............ 59 Figure 14 – White noise residuals plotted for DBH yield model for G. sepium............... 59 Figure 15 – White noise residuals plotted against predicted heights for C. velutina. ....... 60 Figure 16 – White noise residuals plotted against predicted heights for G. sepium. ........ 60 Figure 17 – Fuelwood yield by product segment for C. velutina and G. sepium: A) Small product segment and B) large product segment in scenario 1; C) Small product segment and D) large product segment in scenario 2. ............................................. 61 Figure 18 – Mean annual increment (MAI) of fuelwood for both products in A) scenario 1 and B) scenario 2. ............................................................................................... 62 Figure 19 – Maximum IRR at harvest net of inflation for scenario 1. ............................. 65 Figure 20 - Maximum IRR at harvest net of inflation for scenario 2. .............................. 66 Figure 21 – Maximum NPV at harvest in real prices in scenario 1. ................................ 67 Figure 22 - Maximum NPV at harvest in real prices in scenario 2. ................................. 67  x  Acknowledgments The realization of this study would not have been possible without the research funding provided by the Social Sciences and Humanities Research Council of Canada (SSHRC) and The University of British Columbia (UBC). I offer my enduring gratitude to the faculty, staff and my fellow students at the UBC, who have inspired me to continue my work in this field. I owe particular thanks to Dr. V. M. LeMay and Dr. G. Q. Bull for their support and guidance as well as for continuously challenging me to do better. Dr. H. Nelson and S. J. Mitchell also provided ongoing support and valuable insight.  xi  Dedication  To the people who have made the Limay Community Carbon Project possible  xii  1. Introduction One-half of the world’s annual wood harvest is used for fuel (FAO, 2010a) and 74% of this consumption takes place in developing countries (World Energy Council, 2007). Fuelwood, an aggregate of wood-based fuels, also accounts for 67% of global energy consumption when compared to all other forms of renewable energy production combined (FAO, 2010a). In regions of high demand, fuelwood consumption can lead to deforestation or forest degradation, particularly in dry forest environments where growth rates are low (Sanchez-Azofeifa et al., 2011). Given this demand for wood energy and the need to find alternatives to fossil fuels, it is alarming that fuelwood consumption is still frequently considered unsustainable (IEA, 2009). Since it appears that fuelwood consumption will continue to rise (Arnold et al., 2003), it is important to find renewable ways of increasing production. Having access to energy is particularly important for improving the quality of life of many people in developing countries (Pachauri and Spreng, 2003). The establishment of tree plantations can play an important role in both helping to meet their energy demand while reducing poverty (Mayers, 2006; Street and Price, 2009). In Nicaragua, a study found that growing Eucalyptus spp. compared to using oil for electricity generation could be cost competitive while making a greater contribution to the local economy (van den Broek et al., 2000). In addition to the employment created on industrial plantations, the production of farmbased fuelwood is seen as a means of improving rural livelihoods while improving the environment (Regmi, 2003). This is especially true when structured through inclusive business models (FAO, 2010b) and when established on marginal lands not suited for agriculture (Khamzina et al., 2012). While the use of planted trees by communities and industry for productive purposes is increasing throughout Latin America, only a relatively small number of species have been widely planted (Del Lungo et al., 2006). Due to different biophysical and socioeconomic environments in different tropical regions, there is merit in extending the choice of species available to growers (Brown et al., 1997). The need for farm-based tree plantations is particularly important in countries like Nicaragua where poverty is acute, fuelwood is the country’s predominant energy source (FAO, 2004) and deforestation has led to a critical wood supply deficit (OAS, 1997). However, for landowners to invest in the establishment of fuelwood plantations, economic viability must be expected. Economic viability is defined as the ability to sustain operations on 1  the basis of current and projected revenues equal to or in excess of current and planned expenditures. In the context of plantation forestry, economic viability, including the environmental and social benefits derived from plantations, is a pre-requisite for wider adoption of sustainable forest management practices (FAO, 2005). Smallholders, subsistence oriented farmers with limited land, capital and input technologies with high exposure to risk and low market orientation, generally have a competitive advantage in fuelwood production since the transaction costs in accessing and supervising motivated family labour are low (Poulton et al., 2010). They also have a low opportunity cost of their landholdings, especially when tree plantations are established on portions of the farm not suitable for agriculture. However, they face major disadvantages related to the production and sale of their products. These include: lack of access to capital; lack of technical knowledge (Poulton et al., 2010); few connections to established markets and actors; lack of information on prices; low bargaining power; and the inability to reach economies of scale (Markelova et al., 2009). As such, some research has recently broadened from a focus on improving farmers’ production capacity to improving ways to link smallholders to markets (FAO, 2007). The economic viability of tree plantations is commonly assessed by determining the net present value based on future cash transactions. However, for such an assessment to be feasible and credible, a good understanding of yield by product segment and its associated market value is indispensible. Lack of accurate growth and yield information is often one of the largest sources of risk in evaluating the economic viability of tree plantations (Taylor, 1991). Forecasting growth and yield is commonly oversimplified and is often integrated from external research unintended for economic considerations (Siregar et al., 2007; Niskanen, 1998; Griess and Knoke, 2010). Multiplying merchantable volume by a single price regardless of market requirements is an oversimplified but common practice (Felker and Gevera, 2003; Piotto and al., 2010; Griess and Knoke, 2010). Furthermore, when the performances of Central American tree species are studied (Piotto et al., 2004b; Wishnie et al., 2007; Hall et al., 2011a, 2011b), they have seldom have been linked to market requirements. 1.1.  Objectives The primary objective of this thesis was to determine if plantations of Caesalpinia  velutina and Gliricidia sepium plantations, two tree species native to Nicaragua, could be an 2  economically viable endeavour for smallholders if sold as fuelwood. To improve the accuracy of the economic viability assessment of fuelwood plantations, I used a novel approach to forecast yields by market-specific product segments, thereby accounting for the effects of market requirements on differential revenues and costs. The term fuelwood generally refers to all woodderived products used as a fuel source including charcoal, by-products from wood processing, post-consumer recovered wood and processed wood-based fuels. However, in the context of this thesis from here onwards, the term fuelwood is used to exclusively refer to unprocessed wood used as firewood. In Chapter 2, the market potential for plantation-based fuelwood was assessed by examining product segmentation, prices and merchantability requirements. More specifically, the main questions to be addressed were as follows: 1) What are the various product segments of the fuelwood market and how is each product segment defined? 2) What are the prices that plantation fuelwood could reasonably expect to be sold for given the various product segments? 3) What is the total size and value of market demand? These questions were explored using the largest industrial consumer of fuelwood in the province of Madríz, Nicaragua as a case study. In Chapter 3, the distribution of fuelwood log volumes within Caesalpinia velutina and Gliricidia sepium trees of different diameter at breast heights (DBH; 1.3 m above ground) and total heights were estimated using the product segments determined in Chapter 2. To do so, species-specific taper models were developed based on a version of the Kozak (2004) taper model and fitted using a nonlinear mixed-effects modelling approach to account for within-tree correlations, as well as possible heterogeneity of variances. In Chapter 4, growth and yield models for Caesalpinia velutina and Gliricidia sepium trees were developed. Furthermore, the work from the previous chapters was integrated with the growth and yield modelling in order to analyse the economic viability of farm-based fuelwood plantations of these tree species as a function of different harvest ages. 1.2.  Case description and silvics of species The plantation assessments made in this thesis were made in the region of San Juan de  Limay, Nicaragua. The region is defined as a tropical dry forest ecosystem with temperatures ranging between 24-34o C with two distinct seasons, wet and dry (INIFOM, 2002a). The mean  3  annual precipitation of 1,394 mm falls mostly in the rainy season that typically begins in May and ends in November. Average monthly precipitation is below 100 mm 7 months per year. The region is quite poor and the local economy consists predominantly of agriculture and cattle ranching (Alcaldia de Limay, 2009). Approximately 30% of the land area is covered in shrubby vegetation, which is largely underutilized and provides the community with little economic benefit (Baker et al., 2011). An initiative in the region, the Limay Community Carbon Project (LCCP), supports farmers to establish productive, multi-purpose forest plantations on the underused/unproductive portions of their farms (FAO, 2010c). At the time of writing, there were a total of 135 farming families participating in the project with a cumulative area of 371 noncontiguous hectares of land newly afforested and managed for the production of ecosystem services and forest products (Taking Root, 2012). The plantation establishment and maintenance methods in this study are based on the ones used in the LCCP growing Caesalpinia velutina and Gliricidia sepium trees. In 1979, the Tropical Agricultural Research and Higher Education Centre (CATIE) funded by the United States Agency for International Development (USAID), along with the different forestry departments of the various countries in Central America, developed a research project to identify tree species with promising potential for sustainable fuelwood production in the region. Caesalpinia velutina and Gliricidia sepium, both native to Central America and known to perform well in arid regions and on marginal soils, were identified as species with promising potential (Cannon and Galloway, 1995). Over 20 years ago, multiple linear growth and yield models for both species were developed using data from relatively short time periods and using site indices (Hughell, 1990; Hurtarte, 1990). Caesalpinia velutina is a deciduous tree that produces a single straight bole that reaches 10 to 12 m in height and up to 20 to 30 cm in DBH (Cordero et al., 2003a). With a specific density between 0.7 and 0.75 g/cm3, the dense wood is highly durable making it ideal for fences, posts and fuelwood (Cordero et al., 2003a). The species has a calorific value of 20,000 KJ/KG (Hurtarte, 1990). Gliricidia sepium is a nitrogen-fixing tree that produces high-protein forage and is easily propagated from stakes (Cordero et al., 2003b). As such, it is one of the most commonly used tree species in Central America due to its multiple uses in agroforestry systems. It is a small to medium sized tree reaching comparable dimensions to Caesalpinia velutina (Cordero et al., 2003b). Its specific density ranges between 0.5 and 0.8 g/cm3 and is considered excellent for 4  fuelwood because it burns slowly, produces a lot of ember and little smoke (CATIE, 1986a). The species has a calorific value of 19,228 kJ/kg (Herrera, 1990).  5  2. The importance of product segmentation and merchantability requirements in the assessments of market potential for fuelwood plantations: a case study of the rosquilla industry in Madríz, Nicaragua. 2.1.  Introduction In regions of concentrated demand, fuelwood consumption can lead to deforestation,  otherwise it can contribute to forest degradation, particularly in dry forest environments where growth rates are low (Sanchez-Azofeifa et al., 2011). Studies on fuelwood demand and supply are usually carried out at the national level but the patterns of fuelwood production and consumption and the associated social, economic and environmental impacts are complex and site specific (de Montalembert and Clement, 1983). In the developing world specifically, fuelwood is generally consumed at the household level and by small cottage industries that, in the aggregate, consume very large quantities (FAO, 2003). As such, if we want to transition towards sustainably managed and produced fuelwood, it is imperative that we develop a better understanding of region-specific fuelwood demand patterns and requirements. Furthermore, if we expect landowners to invest in growing fuelwood, it is reasonable to expect that they would like to know that the appropriate price signals are in place before making the investment. In a case study involving a biomass plant in Nicaragua, van den Broek et al. (2000a) concluded that growing Eucalyptus spp. trees could be cost competitive with oil while creating two times more jobs and contributing four times as much to the GDP of the country. Furthermore, if the trees for that market were grown on farms as opposed to industrial plantations, 77% of the value would end up with low income groups (van den Broek et al., 2000b). However, as with many of the previous studies examining the feasibility of fuelwood plantations, a single market price was used for all fuelwood regardless of its characteristics (Ramadhani et al., 2002; Stille et al., 2011; Wicke et al., 2011). Although this might be appropriate under certain market conditions, the inaccuracy arises when such prices are applied to the entire fuelwood plantation. Studies have found that small industry fuelwood markets are commonly highly developed and complex (McCrary et al. 2005), and that fuelwood market demand is segmented (Mayorga and Urbina, 1993). Furthermore, Brouwer and Falcão (2004) found that fuelwood is commonly traded in highly variable non-precise units such as bins, bundles, etc., making it difficult to relate to volume or biomass (e.g., m3 and tonnes) measured for forest stands. As such, if fuelwood 6  supply is to transition towards being plantation based, market-specific equivalencies need to be developed so that supply can be linked with demand. Efforts to better understand the structure of region-specific markets and prices are thus justified so that the economic viability of plantations can be precisely evaluated and so that management can be optimized in accordance with market criteria. Furthermore, such research should prioritize regions like those found in Nicaragua where people are highly dependent on fuelwood and where sustainable sources of fuelwood supply are scarce (Guevara, 2004). To improve our assessment abilities I suggest that there are two key issues that need to be addressed: 1) transparency in fuelwood market prices expressed in units understandable by both consumers and producers; and 2) clarity in market requirements and size. As such, using a case study, the objective of this chapter was to assess the market potential for plantation-based fuelwood if sold to the largest industrial consumer of fuelwood in the province of Madríz, Nicaragua. This was done by examining product segmentation, prices and merchantability requirements. More specifically, the main questions to be addressed in this study were as follows: a. What are the various product segments of the fuelwood market and how is each product segment defined? b. What are the prices that plantation fuelwood could reasonably expect to be sold for in the market given the various product segments? c. What is the total size and value of market demand? An underlying assumption of this research is that merchantable fuelwood is sold according to different product segments which are valued differently. In this context, a fuelwood product segment is defined as a classification of different merchantable fuelwood characteristics with recognized quality differences and a unique trade name. Within each classification, the product is considered homogenous whereas between classifications, the product is considered heterogeneous. In Nicaragua, annual industrial roundwood production amounts to 54,000 m3 compared to fuelwood production of 6.1 million m3 per year (excluding an additional 27,991 tonnes of charcoal production) (FAO, 2010d). Fuelwood is Nicaragua’s largest source of energy (FAO, 2004) and is predominantly sourced from secondary forests, partially regenerated pasturelands and regions undergoing land-use change (PROLEÑA, 2000). At the national level, 90% of this consumption is at the household level and 10% by cottage industries that produce goods such as 7  bricks, baked goods, charcoal, pottery and sugar (PROLEÑA, 2000). However, of the 90% that is consumed at the household level, only a small minority is actually purchased in the market place (Chavarria, 2002). This study looks at the sub-segment of the market that is consumed by industry.Although this share of total consumption is small, industrial fuelwood consumption is much more concentrated compared to residential consumption due to much larger purchase orders and represents 5.7% of total demand but 57% of the market. Figure 1 illustrates a schematic of the industry’s breakdown.  Figure 1 – Nicaraguan fuelwood demand and supply. A breakdown of the market is boxedin by the dotted line. 2.2.  Methods  2.2.1.  Industry description and study area This study was conducted in the province of Madriz, Nicaragua where the problem of  deforestation is considered a high priority (Urbina, 2005). Bordering Honduras and located at a latitude of 13°28'10 N and a longitude of 86°33'50 W (see Figure 2), the province is characterized as dry with an average annual precipitation of approximately 1,000 mm and a mean annual temperature of 22 degrees centigrade. The region is mountainous with an average 8  elevation of 700 metres above sea level. There is a distinct dry season and a rainy season, which starts in the month of May and lasts until the month of November. The vegetation is semi-arid and predominantly consists of woody shrubs, deciduous trees and a few remnant patches of pine (Pinus oocarpa) and oak (Quercus oleoides) forests at higher elevations. Overall, the province is largely deforested and poor with crop production as the predominant economic activity followed by livestock production.  Figure 2 - Map of Nicaragua in Central America with the province of Madríz in dark grey. The industry that produces rosquillas, a popular Nicaraguan cookie-like product, is the largest non-residential consumer of fuelwood in Madriz (PROLEÑA, 2000). These baked goods predominantly consist of corn meal, fresh cheese and sugar. They are produced in bakeries that are low technology small businesses owned and managed by women. Rosquillas are generally baked in large wood-fired ovens located just outside of the owner’s residence (see Figure 3). In many communities, these bakeries are the largest form of industry present; the baked goods are sold all over Nicaragua and exported internationally. Within the province of Madríz, rosquillas are only produced in two of the nine municipalities: Somoto and Yalagüina. In the municipality of Somoto, all of the rosquilla bakeries are located in the city by the same name with an urban population of 15,974 inhabitants (INIFOM, 2002b). In the municipality of Yalagüina, the poorest municipality of Madríz with a total population of 8,741 inhabitants, rosquilla production is spread out throughout six different communities (INIFOM, 2002c).  9  Figure 3 - Man unloads fuelwood near dome shaped ovens in a rosquilla bakery. 2.2.2. Interviews Using a directory of all the rosquilla businesses in the province of Madríz (GERSON R.L., 2010), each business was randomly assigned a number. During the month of January 2012, rosquilla businesses within the different communities were visited for semi-structured interviews (see Appendix A) with either the business owner or manager following the approach by Yin (2002). When a business was unavailable to participate in the study, the next bakery on the randomly sequenced list within that community was visited until at least 20% of the sample population was reached within each community. Of the seven rosquilla producing communities, two remote communities with only two bakeries each in the municipality of Yalagüina were not interviewed due to time constraints. In total, 21 businesses in five communities were surveyed representing a total of 24% of the entire industry. Table 1 presents survey descriptive statistics within each community and municipality. Furthermore, every interviewee answered every question with only a few exceptions. For the few questions that were not answered, results were compiled with a slightly reduced sample size.  10  Table 1 - Proportions of bakeries surveyed by community and municipality. Community Somoto Yalagüina Los Encuentros La Esperanza Salamasí Samascunda Limon Total  Municipality Somoto Yalagüina Yalagüina Yalagüina Yalagüina Yalagüina Yalagüina  N 25 15 3 24 18 2 2 89  n 6 4 1 6 4 0 0 21  % surveyed 24% 27% 33% 25% 22% 0% 0% 24%  N = number of rosquilla bakeries registered in the municipality; n = number of rosquilla bakeries surveyed.  At this stage, each participating business owner was briefed on the nature of the research; those who agreed to participate signed a free and informed consent form according to the ethical conduct for research involving humans set by the Canadian Behavioural Research Ethics Board. Interviewees were asked to free-list the common names of the fuelwood species present within the units that the wood was sold in (either cartloads or human carried loads hereafter referred to as purchasing units), the species that they preferred and the species that they refused to purchase. Species mentioned more frequently were assumed to be more commonly used compared to species mentioned at a lower frequency. Interviews also revealed the number of people working in the bakery, the fuelwood product segments consumed, average weekly fuelwood demand by product segment, price paid per product segment per purchasing unit (all of which contained a mix of numerous species) during the dry season, the wet season and the prices they remember paying for the same products five years ago. Furthermore, interviewees were asked about their expectations regarding the evolution of prices over the next five years and the rationale for their price forecast. 2.2.3. Fuelwood measurements Fuelwood measurements were conducted in the various municipalities during the months of February to May 2012 in order to calculate the solid-wood volume equivalents of traditional purchasing units in cubic metres and to determine the fuelwood log dimensions of the various fuelwood product segments consumed. Purchasing units of fuelwood were purposively sampled at the point of delivery to various rosquilla bakeries to obtain a distribution of fuelwood log dimensions from the different sized purchasing units (Creswell, 2008). For each selected purchasing unit, first, the buyer and seller identified the product segment. Next, the total weight 11  was recorded using a hanging dial scale. Then, a subsample of fuelwood logs representing approximately 10% of the weight of each purchasing unit was selected and weighed. Each fuelwood piece in this subsample what then measured for diameter at each extremity and length. Diameter outside bark at each extremity was obtained by calculating the geometric mean of two diameters at 90-degree angles each measured using a caliper. Length of each fuelwood log was measured using a measuring tape. With these measurements, volume of each fuelwood log was calculated using Smalian’s equation (Husch et al., 1972): Equation 1 where Vf l = volume of a fuelwood log in the subsample (m3); A1= area of the small end of the fuelwood log in m2; A2= area of the large end of the fuelwood log in m2 ; L = length (m). The volume of each subsample (Vsub) was then obtained by summing all fuelwood logs in the subsample. The solid wood volume of each sampled purchasing unit was then calculated as: Equation 2 where Vs = volume of the sampled purchasing unit (m3); Wsub = weight of the subsample (kg); Ws = weight of the sample (kg). In total, 24 purchasing units were sampled and 1018 individual fuelwood logs were measured. Using these sampled purchasing units, the average solid wood volume per purchasing unit (i.e., metric equivalents; m3/purchasing unit) was calculated by product segment. These values were coupled with average price per purchasing unit by product segment from interviews to obtain the average price per unit volume of solid wood ($/m3 ) by product segment. This was repeated using prices for the wet versus dry seasons. Also, frequencies of fuelwood pieces by small-end diameters were used to indicate “cut-off” points for the various product segments. 2.2.4. Market analyses The common names of fuelwood species provided in the interviews were matched with scientific names obtained using Nicaragua’s Ministry of the Environment’s ethno-botanical guidebook (MARENA, 2006). Species were then grouped into three categories according to market preferences: merchantable, preferred and non-merchantable. The average weekly consumption reported in the interviews was converted to metric equivalents and annualized for each product segment and bakery. Some bakeries reported 12  consuming purchasing units that had a mixture of product segments. In such cases, it was assumed that each load had an equivalent proportion of each product segment and the average solid wood volume of the different product segments per purchasing unit was used. Fuelwood demand for each product segment per community was estimated by multiplying the total demand per product segment of the sampled bakeries times their proportionate share of the total number of bakeries in that community. Using these results, demand at the municipal level was estimated by summing the demand of each community. For the bakeries in the two communities where no data were available, municipal averages were used. Demand at the provincial level was simply the sum of demand per municipality. Workforce per community was estimated by multiplying the average number of workers reported per bakery by the total number of bakeries present within that community. To indicate whether consumption by product segment was related to size of work force, Spearman’s correlation coefficients between the percentage of bakeries consuming the various product segments per community versus the total workforce per community were used. Market values at the community and provincial levels were obtained by multiplying the average price times the average demand for each product segment. Unless otherwise stated, prices were expressed as the average price across both the wet and dry season and reported in USD. Throughout the article, the exchange rate used to convert from the local currency into USD was NIO 1.00 = USD 0.04271, the average exchange rate in January 2012. 2.3.  Results  2.3.1. Merchantability requirements and product segmentation Across all communities, rosquilla owners listed 29 different species of fuelwood that they considered merchantable and the number of times that each species was listed ranged from 1 to 14 (see Table 2). Eighteen species were identified as preferred wood species (see Table 3) and 12 species were identified as non-merchantable (see Table 4). The characteristics of the preferred species were: duration of combustion, high heat generation, minimal smoke production and ease of combustion or high ember production. Unfortunately, measures of related to  1  Source: http://www.oanda.com/currency/historical-rates/ (accessed March 14, 2012)  13  combustion (i.e., BTUs and specific gravity) were not available for the majority of the species mentioned. The described characteristics of the species considered non-merchantable were the exact opposites. However, 25% of the respondents stated that in spite of their preferences, due to scarcity, they would purchase any species of fuelwood. Furthermore, 95% of respondents reported that should new fuelwood species with characteristics similar to those of the preferred species become available, they would purchase them. Seven of the fuelwood species considered non-merchantable by some of the bakeries were considered merchantable by others, two of which were even considered preferred species. Neomillspaughia paniculata and Pinus oocarpa were the most contentious species and Gliricidia sepium was the most commonly cited species to be considered non-merchantable. However, two respondents stated that mature Gliricidia sepium was an excellent fuelwood species, but the juvenile wood was of low quality. Another respondent stated that they would purchase this species, but only when present as a minority species within the purchasing unit.  14  Table 2 - List of species that rosquilla bakeries purchase and the number of times they were cited by respondents. Common name Amarguito Quebracho Carbon Pino (sawlog residue) Chaperno Paracay/ Tapatamal Lengua de vaca Madero negro Cornizuelo Flor amarilla Laurel Roble Huesito Ron ron Cacho novio Escoba negra Eucalyptus Frijollio Espino negro Neem Jjiñocuado Maria blanca Chilca Tiguilote Guanacaste Miliguiste Bambayan Acacia Varrilla  Number of times cited 14 11 10 9 7 7 5 4 3 3 3 3 3 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1  Scientific name Tecoma stans Lysiloma divaricata Acacia pennatula Pinus oocarpa Albizia adinicephala Neomillspaughia paniculata Solanum atitlanum Gliricidia sepium Acacia collinsii Caesalpinia vesicaria Cordia alliodora Tabebuia rosea Trichilia sp. Astronium graveolens Chloroleucon mangense Cordia spinescens Eucaliptus sp. Leucaena shannoni Adela Barbinervis Azadirachta indica Bruñera simaruba Calophyllum brasiliense Cascabela ovata Cordia dentada Enterolobium cyclocarpum Pithecellobium seleri Rehdera trinervis Senna siamea Ryania speciosa  15  Table 3 - Species that rosquilla bakeries prefer to purchase and the number of times they were cited by respondents. Common name Carbon Amarguito Quebracho Tapatamal/ Paracay Cacho de novillo Chaperno Flora maria Huesito Pino (sawlog residue) Ron ron Varria negra Cornisuelo Escoba negra Frijollio Lengua de vaca Marria blanca Miliguiste Roble  Number of times cited 8 7 7 4 2 2 2 2 2 2 1 1 1 1 1 1 1 1  Scientific name Acacia pennatula Tecoma stans Lysiloma divaricata Neomillspaughia paniculata Chloroleucon mangense Albizia adinicephala Caesalpinia vesicaria Trichilia sp. Pinus oocarpa Astronium graveolens Ryania speciosa Acacia collinsii Cordia spinescens Leucaena shannoni Solanum atitlanum Calophyllum brasiliense Pithecellobium seleri Tabebuia rosea  Table 4 - Species considered non-merchantable and the number of times they were cited by respondent. Common name Madero Negro Guanacaste Tapatamal/ Paracay Jjiñocuado Pino/Ocote (sawlog residue) Nancite Cedro Ceiba Berberilla Laurel Tiguilote Llamarada del bosque  Number of times cited 7 6 3 2 2 1 1 1 1 1 1 1  Scientific name Gliricidia sepium enterolobium cyclocarpum Neomillspaughia paniculata Bruñera simaruba Pinus oocarpa Byrsonima crassifolia Cedrela odorata Ceiba pentandra Cochlospermum vitifolium Cordia alliodora Cordia dentada Spathodea campanulata 16  Of the merchantable fuelwood consumed, four product segments were reported: burusca, rolliza, revuelto and ripio. The latter is sawlog residues (i.e., trim, edgings, bark, etc.), a byproduct of a small sawmill industry using pine (Pinus oocarpa) trees. Transporters, including some rosquilla bakeries themselves, purchased and resold the sawlog residue to other rosquilla bakeries. Conversely, the three other product segments were defined by small-end diameters and not by species, provided that they had desirable combustion characteristics. Table 5 presents descriptive statistics for the dimensions for each product segment. The lengths of the fuelwood logs did not play an important role between product segment differentiation. Average lengths varied between 0.98 m and 1.13 m and were more related to the type of transport used than the product segment. Burusca was the product segment with the smallest diameters followed by rolliza. Revuelto fuelwood logs were essentially a mix of rolliza and burusca fuelwood logs that had not been sorted. The diameter cut-off point between burusca (the smaller diameter fuelwood logs) and rolliza (the larger diameter fuelwood logs) was not precisely defined, as there was some overlap between the small-end diameters of both product segments. The 5th percentile of the small-end diameter distribution of rolliza fuelwood logs measured from cartload purchasing units corresponded to a diameter of 2.5 cm. At this small-end diameter, there was approximately the same relative frequency between burusca and rolliza fuelwood logs (see Figure 4). As such, this can effectively be used as a cut-off point between these two product segments. For burusca fuelwood logs, 95% of small-end diameters were > 0.4 cm, which is its appropriate lower diameter cut-off point for that product segment. Despite the flexibility in diameters across product segments, there was a clear differentiation, each recognized with unique trade names.  17  Relative frequency  45% 40% 35% 30% 25% 20% 15% 10% 5% 0%  Small-end diameter Burrusca  Rolliza  Figure 4 - Small-end diameter cut-off point between burusca and rolliza fuelwood logs. In addition to dimensions and species, 66.7% of respondents reported other requirements for merchantability. These requirements were: dryness of the wood; absence of thorns and rot on the fuelwood logs; and absence of small branches attached to the fuelwood logs. Table 5 - Descriptive statistics of product segment dimensions. Product segment  Burusca  Rolliza  Revuelto  Purchasing units  Cartload  Cartload  Human load  Purchasing units sampled  10  8  6  Fuelwood logs subsampled  506  355  157  Dimensions Small-end diameter (cm) Large-end diameter (cm) Length (m) Small-end diameter (cm) Large-end diameter (cm) Length (m) Small-end diameter (cm) Large-end diameter (cm) Length (m)  Min  Max  Mean  Std  0.1  3.1  1.3  0.6  0.3  5.1  2.1  0.7  0.70  1.54  1.13  0.13  1.5  8.4  4.8  1.2  2.9  8.9  5.2  1.2  0.11  1.89  1.07  0.14  0.2  5.7  1.5  1.2  0.6  6.5  2.2  1.3  0.1  1.77  0.98  0.19  Min=Minimum; Max=Maximum; Std=Standard Deviation  18  2.3.2. Product segment procurement and preferences The fuelwood product segments consumed by the different communities varied substantially and no individual community consumed all of them. Figure 5 illustrates the percentage of bakeries surveyed in the different communities that consumed the various fuelwood product segments. Bakeries in Somoto most commonly consumed rolliza, the segment with the largest small-end diameters, as did those in La Esperanza whereas the other communities did not consume it at all. Aside from Los Encuentros, the other communities consumed larger diameter fuelwood but did so by consuming revuelto, a mixture of large and small diameter fuelwood. These same communities were also more reliant on ripio (pine sawlog residue) as a fuel source. Burusca, the segment with the smallest diameters, was the product segment most commonly consumed by all communities followed by ripio. Of the bakeries interviewed, only one (representing less than 5% of the sample) reported consuming exclusively ripio whereas others reported only consuming it when no other fuelwood was available. Unlike other fuelwood product segments, sawlog residue is a by-product of cutting large diameter logs rather than a direct product from plantations; therefore, no further analysis of this product segment is presented in this thesis.  Figure 5 - Percentage of bakeries that consume the various fuelwood product segments per community.  19  It was estimated that the total number of people working in the rosquilla industry, which was used as a proxy for size of the industry, varied between 363 people in the community of Somoto and 15 people in the communities of Limon and Samascunda. The total number within the entire province of Madríz was estimated to be 821 people. The correlations between the estimated size of the work force in each community and the percentage of bakeries that consume the various fuelwood product segments revealed that there was a strong positive correlation of 0.9 for the use of rolliza and a weak positive correlation of 0.2 and 0.125 for the use of revuelto and burusca respectively and the size of the rosquilla industry. The complete opposite was true for ripio that had a negative correlation in the magnitude of -0.7 with the size of the industry. These results suggest that larger industries have a preference towards larger diameter fuelwood logs. Where the industry is small, the use of ripio is much more common. In terms of procurement, the majority of bakeries reported purchasing their fuelwood from seasonal farmers or full time fuelwood merchants. In addition to these two sources, some bakeries owned their own vehicles and purchased ripio directly from sawmills. Sellers traditionally deliver the fuelwood directly to the bakeries in standardized cartloads measuring 1 x 0.85 x 2.5 metres for a total cart volume of 2.125 m3 pulled by oxen (see Figure 6). A rolliza shipment was said to contain approximately 360 fuelwood logs whereas the carts with the other product segments are simply filled to the top. Although cartloads pulled by oxen were the standard, fuelwood logs were commonly delivered in trucks but referred to and priced relative to their volume in cartloads based on the load dimensions. Additionally, in the smaller rosquilla bakeries in Salamasí and Yalagüina, revuelto was purchased in small loads carried by people or donkeys.  Figure 6 - Traditional fuelwood purchasing unit transported to market. 20  2.3.3. Fuelwood demand by value Figure 7 presents the average price paid per purchasing unit and for the different product segments delivered to the bakery for both the wet and dry season. Since numerous bakeries purchased different product segments in different purchasing units, insufficient data per segment was collected to compare prices per community. When sold by purchasing unit, the most expensive fuelwood was rolliza, the product segment with the largest small-end diameters. The least expensive fuelwood was burusca, which had the smallest diameters. Notably, the price of revuelto was similar to rolliza despite containing small diameter fuelwood logs. In terms of solid wood volume, traditional oxen pulled cartloads carried on average 0.76 m3 of rolliza, 0.33 m3 of burusca and 0.54 m3 of revuelto. Human carried loads of revuelto contained on average 0.04 m3. Figure 8 illustrates the prices per fuelwood product segment converted into cubic metres using these metric equivalencies. $29.36  $30.00  $27.22 $25.19  $/purchasing unit  $25.00  $23.45 $19.35  $20.00 $16.79 $15.00 $10.00 $5.00  $2.35 $2.21  $Cartload burusca Cartload rolliza Cartload revuelto Dry season  Human load revuelto  Wet season  Figure 7 - Average price per purchasing unit by fuelwood product segment, defined by the small-end diameter of the log, in the wet and dry season. Burusca log diameters > 0.4 cm; rolliza logs > 2.5 cm; and revuelto is a mix of both product segments.  21  $70.00  $62.63 $59.28  $58.82  $60.00  $51.03  $58.83  $51.08  US$/m3  $50.00 $38.63  $40.00 $33.15 $30.00 $20.00 $10.00 $Burusca  Rolliza Dry season  Revuelto  Human carried revuelto  Wet season  Figure 8 - Average price per unit volume by fuelwood product segment, defined by the small-end diameter of the log, in the wet and dry season. Burusca log diameters > 0.4 cm; rolliza logs > 2.5 cm; and revuelto is a mix of both product segments. In terms of seasonality, during the rainy season that takes place between the month of May and the end of October, fuelwood prices increased on average by 18.8%. To weather these price fluctuations, some bakeries reported holding fuelwood inventories while others were forced to purchase at higher prices. Based on reported prices for the same product segments in the year 2007, it was determined that between 2007 and 2012 the cost of fuelwood across all product segments increased by an average of 49.8% (or an average annual rate of 8.3%). Figure 9 illustrates that this increase in price was not constant across product segments. Rolliza, the fuelwood logs with the largest small-end diameters, had an average price increase of only 4.2% per year whereas the revuelto product segment increased at an average rate of 11.16% and burusca at a rate of 8.8% per year. This suggests that there has been a relative price increase for lower diameter fuelwood, which is expected to be a global phenomenon as energy prices rise (Roberts, 2008). However, the average rate of inflation of consumer prices in Nicaragua between that same time period was estimated to be 9.4% (IMF, 2011), which implies that in relative terms fuelwood prices actually declined slightly.  22  $100.00 $90.00 $80.00 $70.00 US$/m3  $60.00 $51.08  $51.03 $50.00 $40.00  $33.46  $33.15 $26.97  $30.00  $30.09  $20.00 $10.00 $0.00 Burusca  Rolliza 2012  Revuelto  2007  Figure 9 - Average dry season prices by fuelwood product segment, defined by the smallend diameter of the log, in 2007 and 2012. Burusca log diameters > 0.4 cm; rolliza logs > 2.5 cm; and revuelto is a mix of both product segments. In terms of future prices, 90% of respondents reported that they believe that fuelwood prices will continue increasing over the next five years. Of these respondents, virtually all of them believed that this was going to happen because of an increase in wood scarcity or because of an increase in gas prices. The link between gas prices and fuelwood prices was due to transportation costs. Table 6 outlines estimates annual fuelwood demand and value of the rosquilla industry in Madríz by product segment. As can be noted, the industry consumed an annual total of 7,721 m3 of fuelwood logs for a value of USD $360,065. At purchasing power parity, this is equivalent to USD $3.74 million2. However, the volumes and values varied by product segment and location. The value of fuelwood demand in the city of Somoto was slightly less than half the value of fuelwood consumed in the municipality of Yalaguina, with six different rosquilla producing  2  Source: http://data.worldbank.org/indicator/PA.NUS.PRVT.PP (accessed June, 2012 using 2010 data)  23  communities. Furthermore, the rolliza product segment represents approximately 47% of total fuelwood demand but only 36% of the industry’s value. Table 6 - Values and volumes of rosquilla fuelwood per municipality.  Burusca Rolliza Revuelto Total  Volume Somoto (m3)  Value Somoto (USD)  Volume Yalaguina (m3)  Value Yalaguina (USD)  Total Volume (m3)  Total Value (USD)  1402 1317 0 2719  $77,001 $47,276 $0 $124,277  1893 2293 815 5002  $103,974 $82,298 $49,516 $235,788  3295 3611 815 7721  $180,975 $129,574 $49,516 $360,065  *Value per product segment is apportion to volume  2.4.  Discussion  2.4.1. Merchantability requirements and consumer preferences For the most part, the merchantability requirements for fuelwood sold to rosquilla bakeries in Madríz, Nicaragua were quite intuitive and flexible. Fuelwood consumers had a high level of acceptance for numerous species, which is consistent with findings throughout Latin America (Mayorga and Urbina, 1993; McCrary et al., 2005; Ramos and De Albuquerque, 2012). Furthermore, as observed in this study, a common strategy to sell less preferred species was to mix small proportions of them within purchasing units of preferred species (McCrary et al., 2005). Gliricidia sepium, the species most commonly cited as non-merchantable in this study was the second most commonly sold species in a fuelwood market in Masaya, in the south of Nicaragua (McCrary et al., 2005). However, fuelwood log diameters in that study were reported to be at least 10 cm and given the smaller diameters reported in this study, the findings might not be contradictory. This suggests that only larger diameter Gliricidia sepium fuelwood logs may be valued for fuelwood, which is consistent with reports that its smaller diameter wood has a lower specific gravity (FAO, 1994). This has important management implications given that it is one of the most favoured species by farmers in Central America (Piotto et al., 2004a). Aside from two mentions of Eucalyptus, an exotic species common to the region, the likely origin of the fuelwood species consumed originated from natural forests or regenerated pastures. This further supports the findings that in Nicaragua, for the most part, fuelwood is not grown in plantations (PROLEÑA, 2000). However, given the flexibility of consumer preferences, this study suggests that the Nicaraguan fuelwood market would accept different plantation species as long as they  24  had relatively favourable combustion characteristics. As such, market acceptance is not likely an important concern for the purpose of establishing renewable fuelwood plantations. In addition to species preferences, product segment preferences played an important role in consumption patterns from one community to another. Revuelto and ripio, which require little to no transformation along the value-chain, were consumed by a greater ratio of bakeries in communities with a smaller sized rosquilla industry. On the other hand, burusca and rolliza, which require greater processing in the extent that the fuelwood logs are segmented by small-end diameters, were consumed in greater proportions in communities with a larger rosquilla industry. These results suggest that on a community level, revuelto and ripio are analogous to inferior goods whereas burusca and rolliza are analogous to normal goods. In Greece, Arabatzis and Klonaris (2009) also found unprocessed wood to be an inferior good. These findings would imply that as the size of the rosquilla industry increases in a particular community, there would be a substitution effect away from ripio and revuelto towards rolliza and burusca. A potential partial explanation to this would be that fuelwood traders do not find it profitable to serve more remote communities. As such, smaller industry centres have fewer options available to them and perhaps develop preferences towards what is available. As such, the economic analysis in this thesis will focus exclusively on the burusca and rolliza product segments. Moreover, since ripio is a waste product of an independent sawmill industry, it will likely always be present and sold at a competitive price within the market as long as the sawmill industry is present. 2.4.2. Fuelwood prices Contrarily to the reporting used in other studies done in Nicaragua (van den Broek et al., 2000a; McCrary et al., 2005), I found prices to be expressed in terms of volume not weight. Specifically, fuelwood was sold in traditional purchasing units, predominantly cartloads measuring 1 x 0.85 x 2.5 metres. Purchasing units were priced based on product segments. Product segments with larger small-end diameter fuelwood logs had greater value. Moreover, when prices per carload were converted to price per cubic metre of solid wood, the opposite was true. Counter-intuitively, larger diameter fuelwood became less valuable per unit of solid wood volume compared to smaller diameter fuelwood. How could wood that takes longer to grow be worth less than comparable wood that is much younger? One possible explanation is that a large portion of the value of the fuelwood is not in the wood itself but rather a residual of the costs  25  incurred during harvest and transport. This hypothesis is reinforced by case studies in Brazil (Ramos and de Albuquerque, 2012) and Mozambique (Brouwer and Falcão, 2004) that found that transporters and traders received a larger share of profits along the value chain. However, despite the lower value per unit of wood volume, larger diameter fuelwood is more valuable per load, which is particularly relevant given that transportation costs are related to load. As such, the distance that fuelwood can be transported to market while remaining financially viable will be dependent on the product segments sold. The increase in fuelwood prices during the rainy season was likely due to a decrease in supply. Seasonal merchants return to their crops during the rainy season and transport costs likely increase due to poor road conditions. In a fuelwood market study in the south of Nicaragua, it was found that more than four times the volume of fuelwood entered the market in the dry season as compared to the rainy season (McCrary et al., 2005). Similar seasonality effects were found in a study of domestic consumers in Brazil (Ramos and de Albuquerque, 2012). 2.4.3. The potential for fuelwood plantation The findings presented in this chapter provide some of the necessary information for plantation owners to sell their fuelwood according to market requirements and the prices that they can hope to receive for the delivered product. However, in order to evaluate the economic feasibility of generating sustainably produced fuelwood for the rosquilla market, further research is needed regarding plantation establishment costs, as well as plantation growth and yield, harvesting costs, and transportation costs by product segment. Given these variables, it would be useful to calculate the maximum distance a fuelwood plantation could be located from the market before transportation costs surpass fuelwood value and how this varies according to different product segments. Given that larger small-end diameter fuelwood is less valuable per unit of volume but less expensive to bring to market, which product segment is more profitable and at what transport distance is unknown. This study is not a demonstration of the viability of such an endeavour. Nonetheless, several important insights and further considerations are provided.   The average price per m3 of small diameter fuelwood and the value of the fuelwood market consumed by the rosquilla industry is large relative to the opportunity cost of labour in Nicaragua (MITRAB, 2012). As such, an in-depth evaluation of the economic 26  viability of establishing fuelwood plantations is justifiable despite the common perception that such endeavours are not viable (McCrary et al., 2005; PROLEÑA, 2000). Furthermore, if fuelwood was grown in plantations, it is likely that the variety of species would be much smaller and it is possible that entire purchase orders of preferred species could fetch a premium price.   Fuelwood has a low energy output per unit of wood volume, which drastically limits the distance that it can be transported before transportation costs overcome the value of the fuelwood. As such, remote communities are less likely to be supplied by distant plantations. On the other hand, larger communities such as Somoto with a more concentrated and segmented demand will likely present more profitable markets.  2.5.  Conclusions In this study, I evaluated the product segmentation, prices and merchantability  requirements of fuelwood consumed by the rosquilla industry in Madríz Nicaragua. Additionally, metric equivalencies were developed so that growth and yield information could be combined with market data. Based on the data collected in this study and a review of the literature on this topic, it was concluded that: 1. The rosquilla fuelwood market in Madríz, Nicaragua is segmented into four different product segments, predominantly defined by log dimensions: burusca, rolliza, revuelto and ripio. The latter is a by-product of a small sawmill industry consisting of Pinus oocarpa, whereas the three other product segments are fuelwood logs differentiated by small-end diameters. Bususca was the trade term for smaller fuelwood logs with smallend diameters generally greater than 0.4 cm whereas rolliza represented fuelwood logs with small-end diameters greater than 2.5 cm. Finally, revuelto was an unrefined product segment consisting of a mix of rolliza and busrusca fuelwood logs. The product segments consumed varied by community but overall larger markets favoured greater segmentation. Fuelwood log length was generally close to 1 m and a large variety of species were consumed. Characteristics of preferred species were favourable combustion properties and any species that could provide such characteristics was considered merchantable. The understanding of such segmentation is essential for credible economic viability studies regarding the establishment of fuelwood plantations and their optimal management. 27  2. There were large differences in fuelwood value based on product segment. The average price paid during the dry season for a cartload of rolliza, burusca and revuelto was $25.00, $18.00 and $22.00 respectively. However, the volume of fuelwood per cartload by product was 0.76 m3, 0.33 m3 and 0.54 m3 resulting in prices per m3 for rolliza, burusca and revuelto of $35.89, $54.93 and $55.18 respectively. As such, larger diameter fuelwood fetched the highest price per cartload but the lowest price when converted to cubic metres. During the rainy season, fuelwood prices increased across product segments by an average of 18.8%. Over the last five years, fuelwood prices have increased at a rate of 1.4% below the rate of consumer prices implying that contrarily to public perception, there is no scarcity in the market. However, virtually all rosquilla bakery owners expect prices to increase in the future due to perceived scarcity in the market caused by excessive deforestation. 3. At the aggregate level, fuelwood consumption by the rosquilla industry in Madríz, Nicaragua is 7,721 m3 annually, which excludes ripio. This consumption could be substituted by fuelwood produced by plantations and is valued at approximately $360,000. For terms of comparison, that is equivalent to the annual salary of 327 agricultural labourers and at purchasing power parity, it represents $3.74 million. The results from this study are based on local data for one specific fuelwood market in one geographical region and so care should be taken with respect to transferring the outcomes to different markets. Nonetheless, the level of analysis used is this study could be used by decision makers and plantation managers wishing to produce fuelwood for similar markets in a developing country context.  28  3. Fuelwood log volume by size: estimates for Gliricidia sepium and Caesalpinia velutina using a nonlinear mixed-effects taper model 3.1.  Introduction In North America, log supply managers segregate logs into different grades based  on a number of characteristics including diameter, which has a direct effect on value (Acuna and Murphy, 2007). In addition to having a different value, different diameter logs of the same tree can be employed for entirely different purposes and industries such as lumber from larger logs and pulpwood from smaller logs (Zwcic et al., 2009). In Nicaragua, the fuelwood market is well segmented; different species and logs with different small-end diameters are commonly employed by different industries (Jones and Otarola, 1981; Mayorga and Urbina, 1993). As such, even within the fuelwood industry, various diameters can produce different products for different industries. As such, studies that focus on productivity, biomass or total volume are not particularly useful for determining value or wood utilization given that different portions of the tree are sold at different prices and commonly consumed by different industries. Taper functions describe the rate of change in diameter (i.e., taper) of a tree’s central stem (bole) from ground over the height of the tree. Diameters can then be converted to areas and then areas can be integrated to obtain volume of the entire bole (total stem volume) or volumes of any part of the bole, including volume of the merchantable part of the bole (merchantable volume) or log volumes. Since products are often related to the small-end diameter and/or minimum length of logs, taper models for tree trunks can be used to obtain estimated volumes of the various potential products. Taper models commonly use diameter outside bark at breast height (DBHs; 1.3 m above ground) and total tree heights as predictor variables. As a result, the distribution of volume by log size for a given stand or plantation can be accurately estimated by using the measured DBH and height values combined with taper models. Since the late1960’s, a considerable amount of research has been done on developing and improving taper equations including popular taper functions by Kozak et al. (1969, 1988, 2004) and Max and Burkhart (1976). The variable-exponent equations presented by Kozak (2004), showed almost perfect correlations between measured and estimated upper stem diameters for the majority of the commercial tree species in British 29  Columbia, Canada. These variable-exponent taper models have been widely used (Gezan and Ortega, 2009; Li and Weiskittel, 2010; Heidarsson, 2011); however, as with the majority of the taper models presented in literature, applications have been predominantly on stand-grown coniferous tree species that tend to have a clearly defined main stem. Furthermore, I am unaware of any published research on taper functions in Central America. These models are fundamentally needed to provide required information for the economic valuation of fuelwood plantations, a prerequisite for attracting well-needed investments within the industry. The primary objective of this study was to estimate the distribution of log volumes within trees of different DBHs and heights, using product segments defined by their small-end diameters typically used within the fuelwood markets in Nicaragua (hereafter termed fuelwood log volumes). This was achieved by developing taper equations based on the Kozak 2001 variable-exponent taper model (presented in Kozak 2004) for Caesalpinia velutina and Gliricidia sepium trees. These two species were selected because they are the two most commonly planted tree species in San Juan de Limay, Nicaragua and favoured by farmers (Piotto et al., 2004a). The tools developed in this research will help reduce uncertainty around product yields of fuelwood plantations and, thus, it is hoped that this would stimulate the establishment of new plantations. With the increasing challenges facing the forest sector, the valuation of forest product options becomes an important issue in determining better wood utilization strategies for the purpose of enhancing market competitiveness and maintaining sustainable agricultural, forest and energy resources (Li et al., 2010). 3.2.  Methods  3.2.1. Field measurements Gliricidia sepium and Caesalpinia velutina trees used in this study came from 12 private woodlots smaller than 2 ha in size throughout the municipality of San Juan de Limay, Nicaragua located at 13o10’30 N and 86o36’40 W. The trees were purposively sampled with height and DBH classes for each species as recommended by Kozak and Smith (1993). Furthermore, only undamaged trees were selected. Diameters at 0.3 and occasionally 0.6 m were measured on standing trees prior to felling. Trees were felled as  30  close to the ground as possible and total heights of fallen trees were recorded. Additional diameter measurements were made every 2 m above DBH until the top of the tree (i.e., to a diameter of 0). In many cases, the main stem of the tree was not clearly defined, as there was ambiguity between what would be considered a branch versus the main stem. In such cases, the widest branch was selected. When there were no visually discernible differences in width, the branch most aligned with the rest of the stem was selected. At heights along the main stem where irregularities occurred, such as in the event of branching or irregular growth, a diameter measurement slightly above or below that point was recorded along with the associated height. Since the bark of both species is very thin and fuelwood is sold and consumed with bark, all diameter measurements were made above bark. A diameter tape was used for diameter measurements > 5 cm and a caliper was used for measurements of diameter < 5 cm. Caliper measures were taken at 90 degree angles and the geometric mean was calculated. A total of 32 and 33 trees were sampled for C. velutina and G. sepium respectively, ranging from 5 to 18.9 m tall (Table 7). Table 7 - Descriptive statistics for C. velutina and G. sepium data. C. velutina  DBH (cm) Height (m)  n  Minimum  Maximum  Mean  32 32  3.05 5.00  21.50 18.90  9.39 11.68  n  Minimum  Maximum  Mean  33 33  5.10 6.60  12.80 14.60  8.76 10.53  Standard deviation 3.72 3.20  G. sepium  DBH (cm) Height (m)  Standard deviation 1.99 1.89  n = number of trees  3.2.2. Taper model The taper model labelled as Kozak 2001 in Kozak (2004) was used as the basis in this study, specifically:  Yij  a0 DBHi 1 X ij a      b0  b1 1 / e DBHi / Hti  b2 DBH i X ij  b3 X i DBHi / Hti   eij  Equation 3  31  where Xij= [1.0 –(hij/Hti)1/4] / [1.0-p1/4]; Yij = outside bark diameter (cm) at height hij (m) from ground (cm); DBHi = outside bark diameter at breast height (cm); Hti = total tree height (m); p = 0.01, which was unmodified and taken directly from the original Kozak 2001 model; a0 , a1 and b0 to b3 are fixed-effects parameters to be estimated; eij= error term equation; i = tree; and j = measurement within a tree. Using a logarithmic transformation, Equation (3) was linearized as follows:  ln(Yij ) = a0 + a1 ln(DBH i ) + b0 ln(X ij ) + b1 ln(1/e DBH i / Hti ) + b2 ln(DBH i X ij ) + b3 ln(X ij DBH i / Hti ) + eij  Equation 4  Equation 4 was fitted and parameter estimates from this linear regression were used as starting parameters for Equation 3 by species. Models were fitted using SAS version 9.2 using PROC REG for the linear models and PROC MODEL for the nonlinear mixedeffect models (SAS Institute Inc., 2008). Since the upper stem diameter measurements of each tree were measured at irregular distances from the ground and each diameter measurement within the same tree were expected to be highly correlated with other measures on the same stem, a continuous spatial autoregressive model (i.e., CAR(x)) was used to model the covariance among errors and to obtain consistent estimates of the standard error (Pinherio and Bates, 2000). Further, a variance model was added to account for possible heteroscedasticity, which was modelled as a function of DBH divided by total height based on plots of residuals after accounting for autocorrelation. “White-noise errors” (i.e., errors after accounting for autocorrelation and heteroscedasticity) were used to visually check for any model lack-of-fit, for homoscedasticity of errors and for normality. Once the error structure was modelled, a backwards stepwise procedure was used to remove predictor variables that were not significantly different from 0 using α = 0.05. 3.2.3. Volume estimates Taper equations can also be used to accurately estimate stem volume; however in order to evaluate their reliability, they must be compared to measured (i.e., “true”) volume. For this study, true volume was calculated for each sampled tree assuming standard geometric shapes for the different sections defined by the diameter measures 32  over the tree bole. In particular, a cylinder was assumed for the first section above ground, a cone was assumed for the uppermost section, and a paraboloid frustum was assumed for all other sections. For volume assuming a paraboloid frustum, Smalian’s equation (Husch et al., 1972) was used (see Equation 1 in Chapter 2). Since the trees did not generally buttress and all measures of diameter were taken at lengths generally under 2 m, any overestimation or underestimation of volume was likely to be negligible (Husch et al., 1972). In terms of the corresponding estimated volume using the taper functions, mathematical integration of area from estimated diameter over the tree bole was used since there is no integral form of the Kozak 2001 taper model. For the mathematical integration, diameters were estimated for every 10 cm from the base to the tip of the tree. As with the calculation of true volume, the first section was assumed to be a cylinder, the last section a cone, and every other section a paraboloid frustum. 3.2.4. Model evaluation In order to evaluate the accuracy of the taper models for estimating diameter and volume, eight related measures of fit were used: 1) Pseudo coefficient of determination (pseudo R2): Equation 5 Where ∑  ̂  Equation 6  ∑  ̅̅̅  Equation 7  and yk = the kth volume or diameter observation; y = mean diameter or mean volume; ̂  the kth estimate of diameter or volume; and n = the number of observations.  2) Standard error of the estimate (SEE): √  Equation 8  3) Relative standard error of the estimate (% SEE): ̅  Equation 9  4) Average absolute bias (AAB):  33  ∑  |  ̂|  Equation 10  5) Relative average absolute bias (% AAB): ∑  ̂|  |  Equation 11  ̅  6) Average bias (AB): ∑  ̂  Equation 12  7) Relative average bias (% AB): ∑  ̂  Equation 13  ̅  8) Root mean square error (RMSE): √  ∑  ̂  Equation 14  These eight measures of fit were calculated for diameter measures using all the sample data by species, and then repeated using data separated by different heights and relative heights above ground by species to indicate the accuracy of estimation along the main bole. For volume, the eight measures of fit were calculated using all trees of a species, and then also by DBH and height class per species. 3.2.5. Fuelwood volumes by tree size The ultimate purpose of the taper functions in this chapter was to segregate the merchantable volume of a tree with various combinations of DBH and height for each species into product segments, which allows for a much more precise economic valuation of forest resources. Based on the results from Chapter 2, it was found that fuelwood markets in Nicaragua were segmented into different products based on small-end diameters (Figure 10). Small-end diameters < 0.4 cm were considered non-merchantable in this chapter although they can theoretically be employed as general biomass for the production of bricks and tiles. Logs with small-end diameters that range between 0.4 cm and 2.5 cm are used as unprocessed fuelwood in commercial bakeries under the smaller diameter product segment. Unprocessed logs with diameters that range between 2.5 cm 34  and 7 cm are sold as fuelwood logs under the larger diameter product segment. Finally, logs with small-end diameters > 7 cm can be split and sold in the domestic fuelwood market whereas smaller logs cannot.  Small-end diameter 0.4 cm and smaller nonmerchantable  Small-end diameters between >0.4 and 2.5 cm  Small-end diameters between >2.5 and 7 cm  Small-end diameters > 7 cm  Bottom 10 cm non-merchantable  Figure 10 - Tree volume separated into product segments.  35  Using the developed taper models, the merchantable volume, in this case defined as the volume above a 0.1 m stump height to a 0.4 cm minimum small-end diameter, for a particular DBH and height combination was separated into logs by the aforementioned fuelwood product segments for C. velutina and G. Sepium. The fuelwood logs were set to 1 m in length, which is typical for the fuelwood industry in Nicaragua so that they can be transported in standard sized carts. After the main bole was sectioned into 1 m logs, leftover sections > 0.9 m were considered merchantable whereas smaller sections were not. 3.3.  Results and discussion  3.3.1. The taper models Taper models using a nonlinear mixed-effects model were developed using the Kozak 2001 (cited in Kozak 2004) taper model for Caesalpinia velutina and Gliricidia sepium trees. The statistical procedure used accounted for autocorrelation of the errors as a function of distance between segments and for heteroscedasticity. For both species, b1 was not statistically significant from 0 using  0.05; the associated variable was  dropped from both models. Remaining parameters were all different from zero (see Table 8. For both equations, rho1, the correlation parameter, was significantly different from zero. As such, the use of a first-order continuous autoregressive error structure (CAR (x)) was effective. Visual inspections of normality plots of “white noise errors” revealed that there were slight departures from normality, but these were expected to have little effect on p-values (Kutner et al., 2005). Residual plots showed no lack of fit (Figure 11) indicating that the taper models fitted the sample data. Furthermore, the same plot indicated an even distribution of these errors, suggesting that the variance function used was effective in stabilizing error variance. A similar method was effectively used by Özçelik et al. (2010) using a nonlinear mixed-effects taper model for coniferous tree species in Turkey.  36  Table 8 - Taper model estimates for Caesalpinia velutina and Gliricidia sepium. C. velutina Parameter  Approximate Standard Deviation 0.0446 0.0278 0.0163 0.0197 0.0727 0.0726 0.0534 0.4755  Estimate  a0 a1 b0 b2 b3 rho1 sigma pow  0.7856 1.2045 0.3577 0.1578 -0.7855 0.6134 -0.6997 -3.2417  t-value  Approximate p-value  17.62 43.27 21.96 8.21 -10.80 8.45 -13.10 -6.82  <.0001 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001 <.0001  t-value  Approximate p-value  11.55 23.93 16.44 2.57 -4.76 8.07 -10.85 -4.17  <.0001 <.0001 <.0001 0.0110 <.0001 <.0001 <.0001 <.0001  G. sepium Parameter a0 a1 b0 b2 b3 rho1 sigma pow  1.0935 1.0352 0.3491 0.0794 -0.4823 0.5654 -0.6572 -3.3674  6  6  4  4  White Noise Residual  White Noise Residual  Approximate Standard Deviation 0.0947 0.0433 0.0212 0.0309 0.1013 0.0700 0.0606 0.8069  Estimate  2 0 -2  0 -2 -4  -4  -6  -6 0  A)  2  5  10  15  20  Predicted Diameter (cm)  25  0  30  B)  2  4  6  8  10  12  14  16  Predicted Diameter (cm)  Figure 11 - White noise residuals plotted against predicted diameters for A) C. velutina and B) G. sepium.  37  To evaluate the taper models, a variety of measures of fit were used by species as a whole but also separated into different sections of the trees as suggested by Kozak and Smith (1993) (Table 9). The Pseudo R2 (an index of fit analogous to the coefficient of multiple determination (R2) for multiple linear regression) values were 0.981 and 0.978 for Caesalpinia velutina and Gliricidia sepium diameters respectively. Moreover, average biases for both species were close to 0 for most positions along the stem. This suggests that a very large portion of the taper variation was explained by DBH and total height for both species. However, when the diameter estimates were broken down for different sections of the trees, some lack of fit was noted with high RMSE values for the upper sections of the trees where the diameters were small (< 5 cm). This is likely due to the fact that the main stems of deciduous trees tend to become increasingly less defined towards the top of the tree. Furthermore, fewer field measurements were taken at higher relative heights (hij/Hti). Figure 12 displays the predicted and actual relative diameters (dij/DBHi) plotted over relative height (hij/Hti) using all sample trees for each species. In terms of total tree volume for all sample trees, there were slight negative average biases implying that the taper equations slightly over-predicted volume of both species (Table 10 and Table 11). For C. velutina, this over-prediction was in the magnitude of 0.68% and 1.42% for G. sepium, but this relative measure is affected by the average tree size difference in the sample of each species. The overall Pseudo R 2 for all trees of each species was 0.993 and 0.953 for Caesalpinia velutina and Gliricidia sepium respectively.  38  Table 9 - Fit statistics for diameters using taper models for C. velutina and G. sepium. C. velutina Height from ground 0.3 m 0.3 to < 1.3 m 1.3 m 20%a 40% 60% 80% 100% Total G. sepium  n 29 13 32 17 33 31 26 33 214  Height from ground  n  0.3 m 0.3 to < 1.3 m 1.3 m 20% 40% 60% 80% 100% Total  32 11 33 13 33 27 24 33 206  Average diameter (cm) 9.63 9.79 8.93 9.28 7.82 6.43 4.68 0.15 6.70  Average bias (cm) 0.03 -0.01 0.16 0.08 0.17 -0.04 -0.17 -0.02 0.03  Average absolute bias (cm) 0.42 0.32 0.30 0.46 0.54 0.57 0.56 0.02 0.39  Average bias (%) 0.29% -0.09% 1.76% 0.82% 2.12% -0.64% -3.60% -15.43% 0.42%  Average diameter (cm) 9.68 9.25 8.57 9.18 7.00 5.61 4.48 0.00 6.33  Average bias (cm) 0.12 -0.16 0.05 -0.01 0.17 -0.02 -0.01 0.00 0.04  Average absolute bias (cm) 0.45 0.31 0.15 0.46 0.48 0.54 0.61 0.00 0.36  Average bias (%) 1.25% -1.68% 0.57% -0.14% 2.49% -0.40% -0.24% 0.00% 0.65%  Absolute bias (%)  SEE (cm)  SEE (%)  RMSE  Pseudo R2  4.34% 3.25% 3.32% 4.93% 6.85% 8.82% 12.00% 15.43% 5.87%  0.60 0.41 0.41 0.64 0.70 0.80 0.82 0.13 0.61  6.22% 4.16% 4.54% 6.89% 8.93% 12.51% 17.58% 88.62% 9.07%  0.021 0.046 0.019 0.035 0.018 0.019 0.023 0.018 0.003  0.982 0.978 0.986 0.967 0.961 0.890 0.746 0.975 0.981  Absolute bias (%)  SEE (cm)  SEE (%)  RMSE  Pseudo R2  4.68% 3.37% 1.78% 4.99% 6.89% 9.55% 13.56% 0.00% 5.67%  0.61 0.35 0.20 0.58 0.61 0.70 0.76 0.00 0.53  6.26% 3.84% 2.31% 6.27% 8.68% 12.55% 17.06% 0.00% 8.41%  0.019 0.055 0.018 0.047 0.018 0.022 0.025 0.018 0.003  0.930 0.935 0.990 0.724 0.842 0.654 0.417 1.000 0.978  n = number of observations  39  Table 10 - Volume bias tables for C. velutina and G. sepium for different height class. C velutina Height Class (m)  n  0 to 10 >10 to 15 >15 to 20 Total  12 17 3 32  Average true volume (cm3) 0.02270 0.04102 0.19956 0.04901  Average bias (cm3) 0.00052 -0.00034 -0.00368 -0.00033  Average absolute bias (cm3) 0.00216 0.00255 0.01459 0.00353  Average bias (%)  Absolute bias (%)  SEE (cm)  SEE (%)  RMSE  Pseudo R2  2.28% -0.84% -1.84% -0.68%  9.49% 6.21% 7.31% 7.20%  0.00 0.00 0.01 0.01  14.88% 7.31% 7.37% 11.05%  0.000 0.000 0.001 0.000  0.957 0.968 0.985 0.993  Average bias (%)  Absolute bias (%)  SEE (cm)  SEE (%)  RMSE  Pseudo R2  0.00% 5.37% -1.08% -6.28% -1.42%  0.00% 9.98% 5.61% 8.20% 7.04%  0.00 0.00 0.00 0.01 0.00  0.00% 12.61% 8.14% 13.68% 12.04%  0.000 0.000 0.000 0.000 0.000  0.00 0.857 0.958 0.481 0.953  G. sepium Average true n volume (cm3) 0 to 6 0 0.00000 >6 to 9 11 0.01781 >9 to 12 17 0.03889 12+ 5 0.06382 Total 33 0.03564 n = number of observations Height Class (m)  Average bias (cm3) 0.00000 0.00096 -0.00042 -0.00401 -0.00050  Average absolute bias (cm3) 0.00000 0.00178 0.00218 0.00523 0.00251  40  Table 11 - Volume bias tables for C. velutina and G. sepium for different DBH classes. C velutina DBH Class (cm)  n  0 to 7 >7 to 9 >9 to 12 12+ Total  10 9 9 4 32  Average true volume (m3) 0.01505 0.02981 0.05781 0.15733 0.04901  Average bias (m3)  Average absolute bias (m3)  Average bias (%)  Absolute bias %  SEE (cm)  SEE (%)  RMSE  Pseudo R2  -0.00004 0.00068 -0.00102 -0.00179 -0.00033  0.00114 0.00251 0.00425 0.01017 0.00353  -0.30% 2.30% -1.77% -1.14% -0.68%  7.56% 8.43% 7.36% 6.46% 7.20%  0.00 0.00 0.01 0.01 0.01  9.69% 9.77% 9.34% 7.65% 11.05%  0.000 0.000 0.000 0.000 0.000  0.959 0.894 0.763 0.991 0.993  Average bias (m3)  Average absolute bias (m3)  Average bias (%)  Absolute bias (%)  SEE (cm)  SEE (%)  RMSE  Pseudo R2  -0.00081 -0.00007 -0.00093 -0.00050  0.00157 0.00204 0.00422 0.00251  -5.23% -0.21% -1.48% -1.42%  10.21% 6.52% 6.69% 7.04%  0.00 0.00 0.01 0.00  11.85% 8.07% 11.59% 12.04%  0.000 0.000 0.000 0.000  0.806 0.905 0.470 0.953  G. sepium Average true n volume (m3) 0 to 7 9 0.01540 >7 to 10 15 0.03129 10+ 9 0.06312 Total 33 0.03564 n = number of observations DBH Class (cm)  41  A)  B)  C)  D)  Figure 12 - Relative diameter (dij/DBHi) versus relative height (hij/Hti) for: A) measured diameters of C. velutina; B) predicted diameters for C. velutina; C) measured diameters for G. sepium; and D) predicted diameters for G. sepium. Although only results of the Kozak 2001 taper model applied to these two species are shown here, other models from the Kozak (2004) were also fitted. For this data set, the Kozak 2001 model showed good results and was less complex than the Kozak 2002 model. Although taper equations have been widely used for several decades, this study presents the only published work that I am aware of on taper equations for Central American deciduous tree species. Of the scarce published taper research done in Latin America as a whole, most of it has been done on either conifers or exotic species. However, Gezan and Ortega (2009) fitted several taper functions, notably a version of the Kozak function, on a few native hardwoods in Chile. In terms of volume prediction, Hurtarte (1990) developed a log-linear model for C. velutina assuming a single geometric shape for the entire tree and using sample data with a smaller range of tree sizes. 42  However, this does not provide the information needed to access plantation values. For G. sepium, a lot of research has been done on biomass equations (e.g., Otárola and Ugalde, (1983)), and also for height and diameter growth, but no publications for volume or taper were found. 3.3.2.  Volume by product segment Table 12 presents the distribution of estimated volumes by fuelwood product  segments using the developed taper models for different DBHs and associated total height. The DBH height relationship was developed in a height prediction model in Chapter 4. However, in the analysis of economic viability, DBH and total height over time would be obtained using a growth and yield model. As expected, larger trees have a larger portion of the merchantable volume in the fuelwood product segments with larger small-end diameter limits. Since these fuelwood product segments have different values per unit of volume, this has important implications from a plantation management perspective. The concept of product segments defined by diameter limits having different values is well integrated into numerous decision support tools. For example, in British Columbia, Canada, a software application package called TIPSY was developed to provide growth and yield estimates over time for a number of coniferous species. Included in the software are options to cut (i.e., buck) the trees from a growth forecast to any future time according to diameter and length specifications to optimize values assigned to different log sizes (Di Lucca, 1998). As another example, the US Forest Service uses the Forest Vegetation Simulator as the national standard for forest growth and yield modelling. This software has an economic extension that assesses revenue by diameter segment (Martin, 2012). In Latin America, the SILVIA Forest Management System, a computer application that simulates growth and yield, contains a module on financial analysis (Vallejo et al., 2006); however, this financial analysis does not segment volume by log size classes.  43  Table 12 - Distribution of merchantable volume (m3) by product segments for DBH height combinations of C. velutina and G. sepium trees. Diameter class (cm) >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7 >1 to 3 >3 to 7 >7  DBH (cm) 5  7  9  11  13  15  17  19  C. velutina Height Volume (m) (m3) 0.0009 5.88 0.0042 0.0000 0.0004 8.89 0.0166 0.0000 0.0009 11.07 0.0123 0.0232 0.0006 12.19 0.0121 0.0462 0.0015 12.62 0.0072 0.0732 0.0007 12.74 0.0105 0.0935 0.0012 12.76 0.0078 0.1204 0.0018 12.77 0.0049 0.1495  G. sepium Height Volume (m) (m3) 0.0008 4.11 0.0030 0.0000 0.0006 6.71 0.0122 0.0000 0.0007 9.20 0.0119 0.0174 0.0009 10.97 0.0091 0.0427 0.0010 11.93 0.0110 0.0651  The recognition that markets commonly value units of volume depending upon product segments defined by small-end diameters must be integrated into existing decision support tools if plantations are to be managed for value yield. Questions that can be addressed include: When should a plantation be harvested to maximize value? When harvesting, what is the optimal bucking for each tree? How does plantation management vary by species? As markets change, how does this alter management for value? Taper models, along with growth and yield models and market information, provide the critical information needed to respond to these and other questions regarding plantation management. However, taper models are lacking for most Latin American species. As demonstrated in this study, new taper equations could be developed for plantation tree species in particular, thereby providing important tools for assessing tree and plantation values. In addition to the tree’s main bole, branches can potentially provide an important component to a tree’s total sellable volume provided that they meet market dimension 44  requirements. To provide a more complete analysis, further research is needed to develop taper equation for tree species branches. As well as providing information for improved plantation management, more precise economic valuation can decrease economic uncertainty. This reduced uncertainty may provide increased investments in plantations for fuelwood production in Nicaragua, thereby reducing forest degradation due to non-sustainable harvest for fuelwood. The taper models presented in this chapter are, therefore, important components to improving overall management of Nicaragua’s forests. 3.4.  Conclusions Valuation of forest plantations to improve management and to stimulate  investment is important to the overall sustainability of forests in Nicaragua. In this research, taper models based on those developed by Kozak (2004) were developed for Caesalpinia velutina and Gliricidia sepium, two tree species commonly used by smallholders in the dry regions of Nicaragua. Although this model was initially developed for Canadian tree species, predominantly conifers, it performed well for the two deciduous tropical tree species used in this study. These taper models provide a critical tool in valuating forest plantations, as demonstrated in this chapter for fuelwood product segments commonly employed by the fuelwood market in Nicaragua. As such, the equations presented in this study can allow industry, forest plantation managers and government to improve product flow forecasts. Decreased uncertainty around the economic viability of C. velutina and G. sepium plantations can contribute towards a more sustainable management of the resource.  45  4. Economic viability and yield 4.1.  Introduction To test the viability of fuelwood plantations from the smallholder perspective, I  selected north-western Nicaragua as a case study. As part of an initiative called the Limay Community Carbon Project (LCCP) (FAO, 2010c), 135 smallholder farmer families are growing Caesalpinia velutina or Gliricidia sepium trees on the underutilized portions of their land. Taking Root, a not-for-profit organisation that coordinates the initiative, is exploring the option of setting up a cooperative that would connect the farmers’ trees to existing fuelwood markets, hereafter called the buyer. Building upon previous work by van den Broek et al. (2000a), the primary objective of this study was to evaluate the economic viability of fuelwood plantations within the smallholder context. Lack of accurate growth and yield information is commonly one of the largest sources of risk in evaluating the economic viability of tree plantations (Taylor, 1991). Forecasting growth and yield is commonly oversimplified and integrated from external research unintended for economic considerations (Niskanen, 1998; Siregar et al., 2007; Griess and Knoke, 2010). Multiplying merchantable volume by a single price regardless of market requirements is an oversimplified but common practice (Felker and Gevera, 2003; Piotto and al., 2010; Griess and Knoke, 2010). Therefore, using an approach similar to the one used by Avohou et al. (2011) who directly linked yield modelling with an economic analysis, a secondary objective to this study was to develop growth and yield models for the tree species in question. One of the innovative components of this research is that it integrates the effects of market-specific fuelwood log product segments defined by their small-end diameters on price as well as transportation and harvest costs. Since intermediate buyers are not currently purchasing smallholders’ fuelwood, no reference stumpage price is available. The stumpage price is defined as the price received by the landowner in exchange for the right to harvest the tree.Therefore, in this study, the situation is analyzed from the viewpoint of the hypothetical buyer and the smallholder. From a buyer’s perspective, I consider the maximum stumpage price that  46  could be paid without making a loss. From a farmer’s perspective, I consider at the minimum stumpage price required in order to cover investment costs. 4.2.  Case description The planting methods used in the LCCP commence with the establishment of low  technology nurseries within the various communities using locally sourced seeds, earth, sand and manure. The seedlings are grown in plastic nursery bags and transported to the planting sites by hand or in carts pulled by oxen. The plantation establishment phase consists of setting up a barbed wire fence around the planting site to prevent grazing animals from destroying the young plants, manually clearing the land of bushy vegetation, demarking planting sites according to prescribed distances between trees, digging small holes approximately 15 cm deep using planting spades or shovels and manually planting the trees. Maintenance consists of intensively removing competing vegetation around the young trees for the first four years using a machete. In order to evaluate the micro-economic feasibility of fuelwood plantations, the assessment in this study was made using two different scenarios. The first scenario was designed to evaluate if fuelwood plantations could be an efficient use of underutilized land. As such, the analysis was conducted assuming that the smallholder plants an entire single hectare lot with one of the two fuelwood species. In this scenario, the fuelwood species were assumed to be planted at a starting density of 2,500 trees per hectare (STPH) and the smallholder was responsible for all establishment and maintenance costs. In the second scenario, the analysis was conducted to evaluate whether fuelwood plantations could be an efficient addition to another land use activity such as a mixture of other trees or food crops. In this second scenario, the costs of removing the bushy vegetation and the costs of the fencing were omitted since they were assumed to have already taken place for the initial land-use activity. However, all other costs were included, notably, seedling cost, planting cost, and maintenance. In this second scenario, one of the fuelwood species was planted closely but with wide spacing between rows of trees with 1,111 STPH.  47  4.3.  Methods  4.3.1. Growth and yield models 4.3.1.1.  Data description  In 1979, the Tropical Agricultural Research and Higher Education Centre funded by the United States Agency for International Development along with the different forestry departments of the various countries of Central America developed a research project to identify tree species with promising potential for sustainable fuelwood production in the region (Cannon and Galloway, 1995). The project established numerous permanent sample plots (PSPs) of Caesalpinia velutina and Gliricidia sepium trees throughout Central America. Prior to PSP establishment, the land use varied from case to case but was predominantly abandoned agriculture fields or cattle pasture. The stand average values of the PSP data were made available to the general public as part of the CATIE technical series (CATIE, 1986b). For my study, 68 PSPs established between 1980 and 1983 were used for Caesalpinia velutina trees originating from Guatemala, Honduras, El Salvador, Nicaragua, Costa Rica and Panama. These data were augmented by measures made at a later date that were published in a graduate thesis (Hurtarte, 1990). For Gliricidia sepium, stand-level temporary sample plot (TSP) data from the MADELENA project (Díaz et al., 1997; Gonzales et al., 1997) were utilized and supplemented with PSP data published by Maravilla and Vázquez (1995). The stand-level data available for both species represent a wide range of topographic and climatic growing conditions. In addition to these plot-level data, 32 C. velutina trees and 33 G. sepium trees were purposively sampled in San Juan de Limay in order to acquire data on a wider variety of DBH and heights for modelling DBH/height relationships. Once the trees were located, DBH was measured using diameter tape. Thereafter, the trees were felled at ground level and total height was obtained using a measuring tape (see Chapter 3 for more details). Table 7 (page 31) and Table 13 present descriptive statistics for the tree-level and plot-level data respectively used in this study.  48  Table 13 – Plot-level descriptive statistics for all permanent and temporary sample plots of Caesalpinia velutina and Gliricidia sepium. Number of observations  Minimum  Maximum  Mean  Standard deviation  105  10  97  36.8  19.2  105  1.5  10.0  4.5  2.0  105  1.2  12.3  4.9  2.5  105 105  750 30  6875 820  3001 205  1638 178  Mean annual temperature (oC)  105  22.3  28.6  27.0  0.9  Mean annual Precipitation (mm) Months dryb  105  512  2835  1328  593.52  105 105  4 0  9 60  7.0 19.3  1.5 15.8  Age (months) Stand average DBH (cm) Stand average height (m) TPHa Elevation (m.a.s.l.)  38  16  104  44.3  19.5  38  3.7  13.3  6.7  2.1  38  2.6  18.3  5.0  2.7  38 38  625 20  4311 975  2217 318  597 298  Mean annual temperature (oC)  38  22.3  28.7  25.9  1.9  Mean annual Precipitation (mm) Months dryb Slope (%)  38  1110  2865  1589  494  26 9  6 3  7 60  6.5 12.2  0.5 18.3  Caesalpinia velutina Age (months) Stand average DBH (cm) Stand average height (m) TPHa Elevation (m.a.s.l.)  Slope (%) Gliricidia sepium  a  Trees per hectare at the time of the first measurement (after seedling mortality); bMonths per year with mean precipitation <100mm. Sources: (CATIE, 1986b; Hurtarte, 1990; Maravilla and Vázquez, 1995; Díaz et al., 1997; Gonzales et al., 1997).  4.3.1.2.  Model components  The growth and yield model for each species was based on the integration of three separate sub-models. Simplifying assumptions incorporated into the models were: 1) the average tree represents all trees in the plantation; 2) the number of stems per ha over time is equal to the starting planting density, reduced by 10% for seedling failure. The main 49  reason for making these assumptions was that the number of PSPs available for this study was low. Also, since the models are for plantations forecasted for relatively short periods of time, variation in DBH, mortality of planted trees, and ingrowth of other trees are all expected to be low. Therefore, these assumptions were considered reasonable for this study. Using these models based on these simplifying assumptions, yields by product segment, defined by small-end diameter and log length as presented in Chapter 2, were obtained by the following progression. First, mean DBH was estimated from plantation age, and other variables. Second, mean height was estimated from mean DBH. Using the mean DBH and mean height, taper functions developed for these species in Chapter 3 were applied and used to obtain fuelwood volumes by product segment. Then, since the mean DBH and height were considered representative of all trees, the distribution of fuelwood volumes by product segment were then expanded to a per hectare basis using the planted trees per hectare (STPH), less 10% for seedling mortality (TPH). The development of the mean DBH and mean height sub-models are described in the sections below. As noted, details on developed taper models are available in Chapter 3. 4.3.1.3.  Diameter at breast height (DBH)  The base model for DBH over time was the commonly used Chapman-Richard’s yield model, specifically: (  )  Equation 15  where DBHj,t is mean DBH for plantation j at time t; t = time in months; e is the base of the natural logarithm, which is a constant = 2.71828; ß1, ß2 and ß3 are fixed-effects parameters to be estimated; and  e j,t = error term of the equation.  Initially a random coefficients modelling (i.e., parameter prediction) approach (Clutter et al., 1983; Littell et al., 2006; Schabenberger and Pierce, 2002) was used to introduce and test other important explanatory variables based on the method employed by Rathbun et al. (2011). These additional variables were used to modify the fixed-effects parameters of Equation 15 so that:  50  Equation 16 where f1, f2 and f3 are linear functions of other explanatory variables. The other explanatory variables were: mean annual precipitation; months per year with mean monthly precipitation < 100 mm; elevation above sea level; trees per hectare; slope; and mean annual temperature. However, since there was a limited PSP database available for each species, generally with only a few remeasurements for each PSP, iterative modelling using different subsets of possible explanatory variables was ultimately used instead of random coefficients modelling. Since mean DBH was measured at irregular intervals in time for each plantation, a continuous autoregressive model (i.e., CAR(x) (Pinherio and Bates, 2000)) was used to model the correlation of errors over time within a PSP. After accounting for autocorrelation, a variance model was added to account for possible heteroscedasticity by modelling variance as a function of DBH. Each possible model was fitted using PROC MODEL of SAS version 9.2. PROC NLIN of SAS version 9.2 was used to get starting values for fixed-effects parameters. The validity of the model including the modelled error covariance structure was verified visually using residuals plots of “white-noise” errors (i.e., errors remaining after accounting for autocorrelation and heteroscedasticity). These errors were also used to visually check for any model lackof-fit, and for normality. Candidate models were then compared using the Pseudo R2 where y = mean DBH. Since there were only a few PSPs for G. sepium, the models were fitted using data for both species combined. Then, a subset of candidate models that indicated good fit for the pooled data were modified using a dummy variable to alter each fixed-effect parameter for G. sepium. A likelihood ratio test was used to test the null hypothesis of no differences between the two species (i.e., fixed-effects parameters with all dummy variable modified are equal to 0). 4.3.1.4.  Height prediction  The data from the PSPs, TSPs and supplementary tree-level data (i.e., tree height rather than mean height) were combined for each species to obtain a wider range of DBH and height values for fitting height prediction models. A number of mean height prediction models were considered based on those proposed by Staudhammer and LeMay 51  (2000), along with a model proposed by Yang et al. (1978). Initially, only mean DBH was used as a predictor variable, but then the models were augmented using measures of competition (i.e., basal area per hectare and trees per hectare) and the same topographic and climatic variables considered for the mean DBH models. Each candidate nonlinear mixed-effects model was fitted per species using full information maximum likelihood with PROC MODEL in SAS version 9.2. To account for possible heteroscedasticity, variance models were included, where error variance was a function of mean DBH. Unlike the models for mean DBH, correlation among errors was considered to be nearly zero since: 1) the dataset for each species included TSPs and tree-level data measured at one time; and 2) PSPs had only a few remeasurements. As a result, the error covariance matrix was sparse with most of the off-diagonal elements (i.e., covariances) equal to zero. Using the goodness of fit measure as for the mean DBH models, a species-specific model for mean height was selected for each species. 4.3.2. Volume by product segment Using the developed growth and yield model, volume yield by product segment was forecasted over time for each of the two scenarios. As noted, the first scenario was based on 2,500 STPH, whereas the second scenario used 1,111 STPH, where the second scenario was used to simulate a tree plantation with another land-use activity, such as another crop species planted in alternative rows. For this second scenario, it was assumed that the additional land-use did not affect tree growth. A modelled plantation was grown under each scenario resulting in the mean DBH and mean height over time. These were then input to the existing species-specific taper models from Chapter 3 to obtain the volume of the mean tree bole. In Nicaragua, fuelwood is commonly sold by volume and priced according to product segments defined by log length and small-end diameters (see Chapter 2). Therefore, the tree bole was then segmented into 1 m log lengths and the taper model was used to obtain the large and small-end diameter of each log. Logs with small-end diameters ranging from > 0.4 to 2.5 cm locally known as burusca were grouped together as the smaller fuelwood product segment and logs with small-end diameters > 2.5 cm locally known as rolliza were grouped together as the larger product segment. For both scenarios, the bottom 10 cm of  52  the tree was considered non-merchantable as it was considered to be a stump left behind at harvest; also, the top of the tree bole with a diameter less than 0.4 cm was also considered non-merchantable. The 1 m length used for logs is typical for the fuelwood industry in Nicaragua since these logs are commonly transported in standard-sized carts (see Chapter 2). However, the log length was varied by up to 10% when this shorter log length resulted in a more valuable product segment. Then, for each log, the volumes were estimated. Since the taper functions cannot be integrated, the taper function was used to obtain the small and large-end diameters of each 10 cm sections within each log. Using these diameter estimates, volume was estimated for each 10 cm section assuming a paraboloid frustum shape and using Smalian’s equation (Husch et al., 1972) given in Equation 1 in Chapter 2. These volumes were summed for each log and then by product segment. Since the mean tree is assumed to represent all trees in the plantation, the volume per ha by product segment was obtained at each age and for each scenario. 4.3.3. Methods for economic viability assessment Economic viability assessments are commonly made using a number of different metrics, notably the net present value (NPV) of future cash transactions or the internal rate of return (IRR) made on the investment. However, both of these measures require a reference stumpage price for the fuelwood, which was not available. The only price information available was the purchase prices of fuelwood by product segment delivered to the doorstep of the buyer located in an urban municipality 53 km away from San Juan de Limay (see Chapter 2). Since it is unlikely that smallholder farmers will be able to identify, locate and negotiate favourable markets in a different region and arrange transportation logistics without the help of a third party organization, a part of the economic viability assessment was made by analyzing upper and lower feasibility price thresholds. To do so, a method similar to the one employed by van den Broek et al. (2000a) was used but segmented for different fuelwood log products since they are valued differently and have different transportation costs per unit of volume (see Chapter 2).  53  From the smallholder’s perspective, the minimum price required (Stumpage MIN) represents the average stumpage price ($/m3) that the farmer would need to receive in order to cover his or her labour and capital investment (land was assumed to be underutilized and thus without an opportunity cost) given a necessary minimum rate of return, which was calculated as follows: Equation 17 where Vtot = the combined merchantable volume of all product segments at harvest (m3/ha); and TCSMALLHOLDER = total costs incurred by the smallholder defined as: ∑  {[∑  ]  }  Equation 18  where ci = cost per hectare of item i ($); fi(t) = frequency of cost item i in year t; r = real inflation rate of capital or labour (%); T is the number of years until harvest; and M is the number of cost items in year t. From the buyer’s perspective, the maximum stumpage price (StumpageMax) ($/m3) was based of the Rothery method (Rothery, 1945) and represents the highest price that could be offered for the fuelwood without making a loss given operating costs (harvest, transport and a profit and risk allowance). It was calculated as follows: Equation 19 where (  )  Equation 20  where NI = net income received by the buyer for selling the fuelwood expressed in future value ($); VB = volume per hectare of the smaller fuelwood product segment (m3); VR = volume per hectare of the larger fuelwood product segment (m3); PR = price of the larger fuelwood product segment in year t ($/m3); PB = price of the smaller fuelwood product segment in year T ($/m3); T = harvest time when the fuelwood was sold (year); rinf = inflation rate of fuelwood prices in real terms (%); [  ] Equation 21 54  where TCbuyer = total costs per hectare for bringing the fuelwood to market in year T ($); HB = the cost of harvesting and sectioning the smaller fuelwood log product segment expressed in future value ($/m3); HR = the cost of harvesting and sectioning the larger fuelwood log product segment expressed in future value ($/m3); TDB = the average cost of loading and transporting the fuelwood logs of the smaller product segment in carts pulled by oxen to a centralized depot expressed in future value ($/m3); TDR = the average cost of loading and transporting the fuelwood logs of the larger product segment in carts pulled by oxen to a centralized depot expressed in future value ($/m3); TMB = the average cost of loading and transporting the fuelwood logs of the smaller product segment from the depot to the market in a truck with a load dimension of 2.3 x 2 x 7 m expressed in future value ($/m3); TMR = the average cost of loading and transporting the fuelwood logs of the larger product segment from the depot to the market in a truck with a load dimension of 2.3 x 2 x 7 m expressed in future value ($); PRA is a profit and risk allowance set at 12%, which is a common rate used in forestry (Davis, 1966), and all the other variables are as previously defined. Using the maximum stumpage price that the buyer could pay, the IRRs and NPVs per hectare from the smallholder’s perspective for both species were calculated for harvest years ranging from 5 to 14. A reference real discount rate of 5% was used based on a similar study done in Nicaragua (van den Broek et al., 2000a). Using the maximum stumpage price for the economic viability calculations assumes that the buyer is competitive. In other words, the maximum stumpage price is equal to the revenue minus harvest and transportation costs and a small allowance for risk and profit. Since many of the costs and revenues associated with the fuelwood plantations occur throughout time, uniform inflation was assumed (i.e., costs and prices increase at the same rate) and the results are presented in real terms (i.e., net of inflation) (Karathanssis, 1980). The calculations in both scenarios were made assuming a single rotation. 4.3.3.1.  Input data for economic assessment  In order to compute the minimum and maximum stumpages using Equations 17 through 21, information on costs, prices, frequencies and the moment in time for which they occurred were needed. Table 14 presents all cost data for establishment and  55  maintenance in both scenarios including the sources of the data. Table 15 presents the harvest costs and fuelwood prices including the sources of the data. Table 16 presents the frequencies and timing of costs including the sources of the data. Table 14 – Establishment and maintenance costs and other input variables for the fuelwood plantations under the two different scenarios. Parameter Trees per hectare planted (STPH)  Scenario 1 Value Unit 2500  Trees / ha  10  %  Trees per hectare after mortality (TPH)  2250  Trees / ha  Trees cleared of grass competition  170  Trees planted per day of labour  Source  Scenario 2 Value Unit  Source  1111  Trees / ha  10  %  N/A  1000  Trees / ha  N/A  Trees / day  (Baumann, 2012)  170  Trees / day  (Baumann, 2012)  100  Trees / day  (Baumann, 2012)  100  Trees / day  (Baumann, 2012)  Removal of bushy vegetation prior to planting  72.00  $ /ha  (Baumann, 2012)  0.00  $ /ha  (Baumann, 2012)  Average fencing costs  140.81  $ /ha  (Baumann, 2012)  0.00  $/ha  (Baumann, 2012)  Labour day rate  3.48  $ /day  (Baumann, 2012)  3.48  $ /day  (Baumann, 2012)  Cost per seedling  0.05  $  (Baumann, 2012)  0.05  $  (Baumann, 2012)  4.00  $  (Asamblea Nacional de la Republica de Nicaragua, 2003)  Establishment mortality  N/A (Baumann, 2012)  N/A (Baumann, 2012)  Plantation registration cost (INAFOR)  4.00  $  (Asamblea Nacional de la Republica de Nicaragua, 2003)  Plantation inspection cost (INAFOR)  5.00  $  (Castellon, 2012)  5.00  $  (Castellon, 2012)  Legal fees for plantation registration  6.50  $  (Castellon, 2012)  6.50  $  (Castellon, 2012)  5  %  (van den Broek et al., 2000a)  5  %  (van den Broek et al., 2000a)  0.00  $ / ha  (Baumann, 2012)  0.00  $ /ha  (Baumann, 2012)  Minimum required discount rate Opportunity cost of the land  N/A = non-applicable; INAFOR = Nicaraguan forestry institute  56  Table 15 - Prices and harvest-related costs for scenarios 1 and 2. Parameter  Value  Unit  Source  Harvest, bucking and loading costs  1.9  $/m3  (Castellon, 2012)  Cartload transport cost from farm to depot  3.48  $ / journey  (Castellon, 2012)  Price of small diameter fuelwood  54.93  $  Chapter 2  Price of large diameter fuelwood  35.89  $  Chapter 2  Truck transport cost from depot to market  150  $ / journey  (Castellon, 2012)  Volume of small diameter fuelwood per cartload  0.33  m3 / load  Chapter 2  Volume of large diameter fuelwood per cartload  0.76  m3 / load  Chapter 2  Volume of small diameter fuelwood per truck  3.96  m3 / load  Chapter 2  Volume of large diameter fuelwood per truck  9.12  m3 / load  Chapter 2  Table 16 – Frequency and timing of fuelwood plantation costs for scenarios 1 and 2. Cost Items Fencing Fencing set-up Land clearing for planting Seedlings Grass removal Planting the trees INAFOR registration INAFOR inspection INAFOR Legal fees  0 1 1 1 1 1 1 0 0 0  1 0 0 0 0 2 0 0 0 0  Year 2 0 0 0 0 3 0 1 1 1  3 0 0 0 0 2 0 0 0 0  4 0 0 0 0 1 0 0 0 0  INAFOR = Nicaraguan forestry institute  For both scenarios, it was assumed that the farmers own their own tools (i.e., shovel and machete) and, therefore, these do not present additional costs. In terms of transportation costs, it was assumed that the fuelwood is transported from the plantation to various landings beside public roads using oxen pulled carts at a constant average cost. From there, the fuelwood was assumed to be picked up by a truck with a load size 57  measuring 2 x 2.3 x 7 m (the largest truck available in the region) and transported directly to the market. When the fuelwood was sold at the market, different product segments receive a different price per unit of volume; however, the stumpage price paid to the farmers was based on the proportional volume of the different product segments. 4.3.3.2.  Sensitivity analysis  The base analyses for both scenarios were calculated assuming that all costs and prices stay constant over time (i.e. 0% real increase). Using the present as the base year, a sensitivity analysis was conducted to observe the change in the NPV per hectare with a harvest in year 10 under the following conditions: +/- 10% change in fuelwood prices in real terms; +/- 10% change in costs in real terms; +/- 1% change in the discount rate; and +/- 10% DBH yield at 10 years of growth. 4.4.  Results  4.4.1. Mean DBH and height models The nonlinear mixed-effects model used in DBH prediction included an error covariance structure that accounted for autocorrelation using a continuous-time autoregressive error structure (CAR (x)). The correlation parameter was significantly different from 0 using  . Visual inspections of residual plots of errors once  autocorrelation was removed and error variance was stabilized indicated an even distribution of errors. For C. velutina, there was some lack of fit for mean DBH values < 2.5 cm (Figure 13) whereas the model fit the G. sepium data well (Figure 14). This slight lack of fit for small trees would only have an impact at very young ages and would not have any great impact on the economic assessment.  58  3  White noise residuals  2 1 0  -1 -2 -3  -4 -5 0  2  4 6 Predicted DBH (cm)  8  10  Figure 13 – White noise residuals plotted for DBH yield model for C. velutina. 3  White noise residuals  2 1 0 -1 -2  -3 0  2  4  6 8 Predicted DBH (cm)  10  12  Figure 14 – White noise residuals plotted for DBH yield model for G. sepium. Using the likelihood ratio test and  , differences between species were  detected. The species-specific models (i.e., using the dummy variables to alter the equation for species) were used in the yield model. For mean height, the selected model was: (  )  Equation 22  59  where j = plantation, t = time. No other possible explanatory variables provided improvements over this model using DBH only. The final species-specific models showed no lack of fit (see Figure 15 and Figure 16 below). 4  White noise residual  3 2 1  0 -1 -2 -3 -4 0  2  4  6 8 Predicted height (m)  10  12  14  Figure 15 – White noise residuals plotted against predicted heights for C. velutina. 4  White noise residual  3 2  1 0 -1 -2 -3 -4 0  2  4  6 8 Predicted height (m)  10  12  14  Figure 16 – White noise residuals plotted against predicted heights for G. sepium. Table 17 presents the results for the mean DBH and height models for Caesalpinia velutina and Gliricidia sepium trees.  60  Table 17 – DBH and height yield models for C. velutina and G. sepium. DBH Pseudo R2  Model  Caesalpinia velutina 0.71 DBH = f1 x (1-EXP(-0.0048 x age))0.713526 where f1 = 12.8028 - 0.00143 x TPH Gliricidia sepium 0.80 DBH = f1 x (1-EXP(-0.017184 x age))1.0753 where f1 = 15.43347 - 0.000191 x TPH Height Model Pseudo R2 Caesalpinia velutina Height = 1.3 + 11.4673 x (1-EXP(-0.0137 x DBH2.2468)) 0.83 Gliricidia sepium Height = 1.3+11.1716 x (1-EXP(-0.0056 x DBH2.4532)) 0.73 TPH = trees per hectare after seedling failure; age = age of the stand in months  4.4.2. Merchantable yield by product segment Using the yield model, mean DBH and mean height were predicted over time for each of the two scenarios. Using these two variables and the previously developed species-specific taper equations (Chapter 3), fuelwood volume yield by product segment was predicted over time for both species and both scenarios (Figure 17).  A)  B)  C)  D)  Figure 17 – Fuelwood yield by product segment for C. velutina and G. sepium: A) small product segment and B) large product segment in scenario 1; C) small product segment and D) large product segment in scenario 2. 61  In scenario 1, when the trees were planted at full density, the sellable volume of small diameter fuelwood (i.e., small-end log diameters ranging from > 0.4 cm to 2.5 cm) for both species peaked within the first three years but then actually declined in absolute terms as the young shoots transitioned into the larger product segment. After the initial peak, small diameter fuelwood log volumes from the main bole varied drastically from year to year but usually staying below 2 m3/ha. Contrarily, the large diameter fuelwood log volumes from the main bole (i.e., small-end log diameters > 2.5 cm) increased progressively over time. For the smaller product segments, Caesalpinia velutina produced more volume per hectare than Gliricidia sepium but the opposite was true for the larger product segment. Figure 18 presents the mean annual increment (MAI) of both fuelwood products of C. velutina and Gliricidia sepium over time. The MAI of C. velutina increased over time and only started reaching a maximum around the 14th year at a rate of 10.39 m3/ha/yr. On the other hand, Gliricidia sepium reached its maximum MAI just after year 7 at a rate of 13.30 m3/ha/yr of merchantable fuelwood. 14  MAI (m3/ha/yr)  12 10 8 6 4 2 0 0  1  2  3  4  5  6  7  8  9  10  11  12  13  10  11  12  13  Age (year)  A)  C. velutina  G. sepium  8  MAI (m3/ha/yr)  7 6 5 4 3 2 1 0 0  1  2  3  4  5  6  7  8  9  Age (year)  B)  C. velutina  G. sepium  Figure 18 – Mean annual increment (MAI) of fuelwood for both products in A) scenario 1 and B) scenario 2.  62  In scenario 2 where the trees were planted at a density of 1,111 STPH, the trend was very similar. The maximum MAI of merchantable fuelwood was reached just after year 13 for C. velutina at a rate of 5.50 m3/ha/yr. For G. sepium, the maximum MAI was reached just before year 8 at a rate of 7.35 m3/ha/yr. 4.4.3. Economic viability 4.4.3.1.  Stumpage prices  Table 18 presents the range of bargaining prices for different harvest ages of the high-density fuelwood plantations under scenario 1. These results assumed a 5% discount rate, a 0% increase in fuelwood prices and a 0% increase in costs over time. As such, the results are said to be real (i.e., net of inflation). Furthermore, this table indicates the present value of the investment required to establish and manage the plantations. Table 18 – Range of potential harvest prices for scenario 1. C. velutina G. sepium Max. Min. Max. Min. Year PV/haa of offering required offering required of TCsmallholder stumpage stumpage stumpage stumpage harvest price ($/m3) price ($/m3) price ($/m3) price ($/m3) 5 $824 $11.41 $36.78 $11.53 $20.01 6 $824 $11.50 $28.37 $11.58 $15.72 7 $824 $11.56 $21.91 $11.62 $12.52 8 $824 $11.43 $17.15 $11.39 $11.54 9 $824 $11.50 $14.94 $11.73 $10.78 10 $824 $11.65 $14.05 $11.63 $10.08 11 $824 $11.62 $13.03 $11.61 $9.97 12 $824 $11.60 $12.25 $11.73 $10.08 13 $824 $11.68 $11.59 $11.73 $10.02 14 $824 $11.73 $10.16 $11.67 $11.10 a PV/ha is the present value per hectare for establishing and maintaining the plantation for the first four years using a 5% discount rate; Min=Minimum; Max=Maximum. Results in bold indicate the largest spread between the maximum potential and minimum required stumpage prices.  Table 19 presents the same concept but for scenario 2 where the minimum required stumpage price from the smallholder’s perspective only represents the additional costs of planting and maintaining the fuelwood species within the existing land-use system.  63  Table 19 - Range of potential harvest prices for scenario 2. C. velutina G. sepium Max. Min. Max. Min. Year PV/haa of offering required offering required of TCsmallholder stumpage stumpage stumpage stumpage harvest price ($/m3) price ($/m3) price ($/m3) price ($/m3) 5 $281 $11.50 $20.06 $11.60 $11.49 6 $281 $11.56 $15.50 $11.62 $8.86 7 $281 $11.60 $11.75 $11.73 $7.59 8 $281 $11.48 $9.98 $11.45 $6.79 9 $281 $11.63 $8.99 $11.73 $6.47 10 $281 $11.60 $8.25 $11.73 $6.21 11 $281 $11.68 $7.80 $11.73 $6.10 12 $281 $11.67 $7.47 $11.73 $6.21 13 $281 $11.66 $7.22 $11.67 $6.19 14 $281 $11.66 $6.41 $11.65 $7.01 a PV/ha is the present value per hectare for establishing and maintaining the plantation for the first four years using a 5% discount rate. Min=Minimum; Max=Maximum. Results in bold indicate the largest spread between the maximum potential and minimum required stumpage prices.  For C. velutina plantations in scenario 1, mutually beneficial buying and selling prices could only be found starting in year 13. The highest stumpage price that could be offered by the buyer was in year 14 for $11.67 per m3 whereas the minimum required stumpage price for the smallholder to break even was $11.10. This represents a mutually beneficial bargaining range of $0.57 per m3. For G. sepium, a mutually beneficial price could be found starting in year 9 whereas the largest spread between the two prices took place is year 13 with a bargaining range of $1.71 per m3. However, the investment required to establish such a plantation was in the magnitude of USD $824 per hectare, which could limit its accessibility for many smallholders given a gross national income (GNI) per capita in Nicaragua of USD $1,1003. In scenario 2 with 1,111 STPH and the omittance of the costs associated with clearing the bushy vegetation prior to planting and fencing, the results were superior. For C. velutina, the maximum potential buying price was greater than the minimum required price starting in year 8. The largest spread of prices were reached in year 14 where the maximum potential buying price was USD  3  Source: World Bank data for 2010 using the Atlas method. Retrieved from http://data.worldbank.org/country/nicaragua  64  $11.65/m3 and the minimum price required by the smallholder was USD $7.01/m3. For G. sepium, the largest spread took place in year 11 where the maximum potential buying price was USD $11.73 m3 and the minimum price required was USD $6.10 m3. In addition to both species being economically viable, the required investment of USD $281/ha makes it a much more economically accessible endeavour. 4.4.3.2.  Internal rate of return (IRR) and net present value (NPV)  Figure 19 presents the IRRs for both species as a function of harvest age in scenario 1 using the maximum potential offering as the stumpage price. Figure 20 presents the same results for scenario 2. For C. velutina, the internal rate of return was positive in scenario 1 starting in year 9 and reached its maximum value in year 14 with an IRR of 5.26% in real terms. Under scenario 2, C. velutina was economically viable starting in year 7 and reached its maximum IRR in year 11 at a rate of 8.91% in real terms. For G. sepium in scenario 1, the IRR reached its maximum in year 10 at a rate of 6.46% and started declining thereafter. In scenario 2 where it was planted at a much lower density, the IRR reached its maximum at 12.57% in year 9 and progressively declined thereafter.  IRR (%)  Scenario 1: Maximum IRR at harvest in real terms 10% 5% 0% -5% -10% -15% -20% -25% 5  6  7  8  G. sepium  9 10 Harvest year  11  12  13  14  C. velutina  Figure 19 – Maximum IRR at harvest net of inflation for scenario 1.  65  Scenario 2: Maximum IRR at harvest in real terms 15% IRR (%)  10% 5% 0% -5% -10% -15% 5  6  7  8  G. sepium  9 10 Harvest year  11  12  13  14  C. velutina  Figure 20 - Maximum IRR at harvest net of inflation for scenario 2. Figure 21 presents the NPV of both species in scenario 1 and Figure 22 presents the results for scenario 2. For scenario 1, G. sepium had a positive value as of year 9 and peaked in year 13. Contrarily, the NPV C. velutina only reached a positive value in year 14 but showed a steady increase over time. In scenario 2, the trend was similar to scenario 1 for both species except with higher values. G. sepium reached its optimal NPV of $221/ha in year 11 and started declining thereafter. On the other hand, although inferior to G. sepium, the NPV of C. velutina increased over time and the NPVs approached convergence. $200 $100 NPV ($/ha)  $$(100)  $(200) $(300) $(400) $(500) $(600) 5  6  7  8  G. sepium  9 10 11 Harvest year  12  13  14  C. velutina  66  Figure 21 – Maximum NPV at harvest in real prices in scenario 1. $250 $200 NPV ($/ha)  $150 $100 $50 $$(50) $(100)  $(150) 5  6  7  8  G. sepium  9 10 11 Harvest year  12  13  14  C. velutina  Figure 22 - Maximum NPV at harvest in real prices in scenario 2. 4.4.4. Sensitivity analysis Table 20 presents the results of the sensitivity analysis using the present as the base year. The results showed that the viability of G. sepium was susceptible to negative changes in prices, costs and yield in scenario 1. However, in scenario 2, it remained an economically viable endeavour after 10 years despite decreases in prices and yield or increases in costs and the discount rate. As for C. velutina, the base analysis with a harvest at year 10 was not an economically viable endeavour; however, increases in price and yield tipped the balance favourably. As for scenario 2, the results were all positive.  67  Table 20 – Sensitivity analysis: effects of change in the economic parameters on the NPV with a 10-year harvest.  Parameter Base analysis + 10% price inflation - 10% price inflation +10% cost inflation -10% cost inflation DBH yield = +10% DBH yield = -10% Discount rate= +1% Discount rate = -1%  Gliricidia sepium  Caesalpinia velutina  NPV  NPV  Scenario 1 $105 0%  Scenario 2 $213 0%  Scenario 1 -$145 0%  $393  274%  $367  72%  $62  143%  $206  129%  -$159  -251%  $72  -66%  -$334  -130%  -$17  -119%  -$98  -193%  $106  -50%  -$292  -101%  $7  -92%  $291  177%  $311  46%  -$10  93%  $166  84%  $297  183%  $320  50%  $20  114%  $170  89%  -$117  -211%  $111  -48%  -$296  -104%  $16  -82%  $31  -70%  $171  -20%  -$194  -34%  $60  -33%  $189  80%  $260  22%  -$89  39%  $124  38%  Scenario 2 $90 0%  The NPV was highly sensitive to changes in price, cost, DBH yield and the discount rate for both species in both scenarios. Overall, scenario 1 was more sensitive to changes in the various parameters than scenario 2 and C. velutina was more sensitive to the same changes than G. sepium under both scenarios. Given the magnitude of the change in the variables used in this sensitivity analysis, the parameter with the smallest effect on the NPV was the discount rate whereas the parameters with the greatest effect were the price of fuelwood and its yield. 4.5.  Discussion  4.5.1. Growth and yield models Over 20 years ago, Hughell (1990) developed a growth model for G. sepium trees using three site indices (low, medium and high). At a similar planting density, the yields found in this study fall within the range predicted by that model using the medium and high site indices. A number of other papers (e.g. Herrera, 1990; Wishnie et al., 2007) refer to the high growth potential of G. sepium; however the studies were based on very short time periods. The results from this research suggest that although the species does 68  exhibit high productivity in its early years, the MAI declines rapidly thereafter, especially at higher planting densities. Hurtarte (1990) developed growth curves based on site indices for Caesalpinia velutina trees. He also attempted to build a site index equation using soil mineral content and the number of months with mean precipitation < 100 mm but found no statistically significant relationships. In this research, neither climate nor topographic variables improved the mean DBH and mean height models, despite the wider range of the data. This is potentially due to the limited number of PSPs along with few repeated measurements for each plot, which is a common problem given the costs of remeasurement. Overall, the maximum MAI of the main bole for both species at both planting densities was generally low compared to the potential of other species in neighbouring regions. At a density of 2,500 STPH, C. velutina reached a MAI of 10.39 m3/ha/yr of sellable fuelwood after 14 years and G. sepium reached 13.30 m3/ha/yr after 7 years. Although the productivity of G. sepium was higher, this was over shorter time periods. If a longer rotation was used C. velutina might have a higher yields with time. Regardless, studies on the productivity rates of native tree species in tropical dry forests of Costa Rica have reported MAI more than three times higher (Piotto et al., 2004b). 4.5.2. Economic viability of fuelwood plantations This study estimated the merchantable volume of fuelwood by product segments typically used within a Nicaraguan fuelwood market for the main bole of Caesalpinia velutina and Gliricidia sepium under two different scenarios. It should be noted that this analysis did not consider the yield potential of the branches, which would likely yield a greater volume of merchantable fuelwood, particularly for the smaller diameter product segment. Unlike the yield of merchantable volume over time for all product segments combined, which commonly follows a sigmoidal-shaped curve, sellable volume of the smaller product segment had a rapid initial production but then decreased in absolute terms with heavy fluctuations. This transition of volume from one product segment to another can have important economic implications. For example, in the scenarios analysed in this study, the cost per m3 to transport the smaller fuelwood log product segment was more than double that of the larger fuelwood log product segment resulting  69  in very little net revenue per m3 for the smaller diameter product segment. As such, its high productivity in the early years of the plantation has little economic benefit. Therefore, the omitted branches in this study are unlikely to have a large influence on the results. In this study I evaluated the economic viability of fuelwood plantations under two distinct scenarios from the economic perspective of a smallholder farmer. In the first scenario, trees were planted at a starting density of 2,500 trees per hectare and it was assumed that the land-use was entirely dedicated to fuelwood production. Under such conditions using a 5% real discount rate while ignoring the opportunity cost of the land, fuelwood plantations were found to be economically viable but only over longer rotation cycles, particularly for C. velutina. However, as demonstrated by the sensitivity analysis, small increases in fuelwood prices over time can have large effects on the economic viability of such endeavours. Although the market price for fuelwood is largely beyond the control of the seller, transforming it into alternative products such as charcoal could potentially allow the buyer to further increase the price offered to the farmer. When the trees are planted at a lower density in addition to another land-use activity, fuelwood plantations were found to be an even more economically viable activity. The main contributors to these differences were the reduced start-up costs in scenario 2 and the higher productivity of the larger fuelwood log product segment per tree, especially in the case of G. sepium. In terms of NPV, G. sepium outperformed C. velutina over the time period analyzed. This suggest that from a purely economic perspective, G. sepium would be a preferred species over C. velutina, especially given smallholders’ potential unexpected need for money at any given year which could result in the need for a pre-mature harvest. However, fuelwood price data used in this study was not species specific and findings from Chapter 2 found that consumer preferences do not favour small diameter G. sepium fuelwood logs, presumably due to the lower specific gravity of the lower diameter wood (FAO, 1994). As such, it is possible that returns on investment would not be able to be realized until the logs reached a larger size. Although this analysis was conducted within the context of the LCCP, the climate conditions in the region are similar to many parts of the country and fuelwood remains the country’s predominant energy source (FAO, 2004). As such, this study suggests that 70  planting fuelwood species, particularly when part of a complimentary land-use system, can be economically sustainable. Furthermore, due to low costs in Nicaragua, the levels of return on investment in the second scenario were found to be comparable to study results found in neighbouring Costa Rica with much faster growth rates and much higher timber prices (Piotto et al., 2004b). 4.5.3. Economic viability in the smallholder context and the role of the buyer According to ESMAP (2010), smallholders in Nicaragua in general have shown little interest in establishing fuelwood plantations without institutional support. ESMAP states that the most important factors are the low sale price of fuelwood, the lack of financial incentives, the absence of a local reforestation tradition, the low level of forest productivity, and the lack of available technical expertise. The NPV and the IRR used in this study were calculated based on the highest potential offering price that the hypothetical buyer could offer under current market conditions without making a loss. This is not necessarily intended to reflect the level of return that a smallholder would make if he or she undertook this investment but rather to highlight the potential opportunity for reforestation initiatives. The economic viability of planting trees in this context, even as a complimentary land-use activity, is dependent on off-farm support. One of the largest constraints is the back-loaded cash flow inherent of plantation forestry, particularly to those who do not have access to credit. However, various additional constraints exist. For example, the registration of the plantations with the Nicaraguan national forestry institute (INAFOR), a prerequisite for legally harvesting and selling forest products, requires writing a forestry plan, geo-referencing the plantation and delivering this information to the capital city hundreds of kilometers away (Castellon, 2012). Furthermore, in the absence of a buyer willing to pay a reasonable stumpage price, the coordination of the transportation logistics of the harvest, the negotiation of price, and the start-up investment required to establish a tree plantation is likely beyond the reach of most smallholders. As such, the results from this study help explain the findings of McCrary et al. (2005) that there is a general perception among smallholders in Nicaragua that growing fuelwood is not economically viable. As discussed by Felker and Guevara (2003), forestry is more than a science or a business; it is a culture that commonly does  71  not exist in arid lands. For forestry to be successful in such regions, it is necessary for forest enterprises to bring the science, business and culture from outside. Successful examples of such arrangements, although structured somewhat differently, can be found in Brazil where biomass consuming industries supply basic inputs to smallholders to set up their plantations based on a contract to purchase the wood in the future for a fixed price (Larson et al., 1994). Due to limited economic opportunities, access to credit, education and technical knowledge, smallholders are unlikely to evaluate the economic viability of an endeavour solely based on the conventional measures of IRR and NPV used in this study. In the context of the LCCP, many smallholders own relatively large properties; however, the area of land cultivated is commonly constrained by limits of family labour and credit. Furthermore, due to the prolonged dry season, the window of opportunity for planting is short resulting in high seasonality of labour availability. Tree crops, unlike traditional agricultural crops, have the added benefit of not needing to be planted annually. Although they require labour for maintenance, this activity can be shifted to non-peak moments of labour demand. The harvesting of trees can also help bridge seasonal gaps of income since trees can be harvested at any time of the year so its seasonality can be countercyclical to agriculture (Falconer, 1990). It can also provide a safety net in times of economic hardship (Townson, 1995) or act as a buffer against income shocks (Arnold et al., 2003; Hedge and Bull, 2011). Furthermore, in addition to the direct measures of economic viability of fuelwood plantations, there are a number of other economic benefits not included in the analysis presented in this study. In Nicaragua, and in a number of other countries, landowners receive an exoneration of property taxes on the portion of land registered as a tree plantation (Asamblea Nacional de la Republica de Nicaragua, 2003). Additionally, 50% of the money invested into the plantation, as well as 50% of the income from the sale of plantation products, is deductible of income tax (Asamblea Nacional de la Republica de Nicaragua, 2003), which was not factored into the analysis in this study. 4.6.  Conclusions In this study, I set out to determine if tree plantations could be an economically  viable endeavour for smallholders if sold in a Nicaraguan fuelwood market according to 72  market specific product segmentation. As part of a more accurate analysis, I developed growth and yield models for Caesalpinia velutina and Gliricidia sepium trees along with taper models that enabled me to estimate yield by product segment utilized within the fuelwood market. The results were analyzed under two different scenarios. In the first scenario, the trees were planted at a high density and represented the only land-use activity. In the second scenario, the trees were planted at a lower density as part of a secondary land-use activity. From this study, the following could be concluded:   The recognition that market-specific fuelwood log segments defined by small-end diameter are priced differently and have different transport and harvest costs can have important economic implications.    Growth and yield by product segment has important implications for the analysis of economic viability. When yield by product segment is observed, there is a decoupling in the trend between merchantable yield and yield of certain product segments. The yield of the small diameter fuelwood logs for both species is expected to increase very quickly in the first few years but then fluctuate widely thereafter.    Despite the high initial productivity of smaller diameter fuelwood products, transportation costs per unit of volume are very high thus greatly diminishing its net value.    Under current market conditions, the establishment of fuelwood plantations of either species can be an economically efficient use of underutilized land over longer time periods and given a number of conditions. However, under scenario 2 when planting the trees within another land-use system, the start-up costs and viability of the endeavour are much better for both species.    Given the need for reaching economies of scale, the legal requirements for registering a tree plantation for commercial use, the complexity of identifying potential markets, and organizing harvest and transportation logistics, it is unlikely that smallholders will be able to receive favourable prices for their fuelwood without institutional support.  73  Although the establishment of tree plantations can play an important role in helping to meet energy demand while providing a productive use of smallholder’s marginal lands, in the context of this study, it is only economically efficient to do so over long enough time periods. From a micro-economic viewpoint and under such circumstances, I can conclude that growing trees for fuelwood could present a good opportunity for smallholder farmers only if there is institutional support.  74  5. Final conclusions In this thesis, I set out to determine if Caesalpinia velutina and Gliricidia sepium tree plantations in San Juan de Limay, Nicaragua could be an economically viable endeavour in the smallholder context if sold as fuelwood. To improve the accuracy of the economic viability assessment, I used a novel approach that forecasted fuelwood yields by product-specific segments, thereby accounting for the effects of market requirements on differential revenues and costs. In Chapter 2, I assessed the market potential for plantation-based fuelwood by examining product segmentation, prices and merchantability requirements of fuelwood sold within the nearby vibrant rosquilla market in Madríz, Nicaragua. This fuelwood market was segmented into different product segments based largely on small-end diameter and sold in traditional purchasing units. Smaller diameter fuelwood logs were sold at lower prices per cartload but when converted to cubic metres, the opposite was true. In Chapter 3, taper models based on those developed by Kozak (2004) were developed for Caesalpinia velutina and Gliricidia sepium to estimate volume for product segments identified in Chapter 2. Although the taper model form was initially developed for Canadian tree species, predominantly conifers, it provided highly accurate volume by merchantable product segment estimates for the two deciduous tropical tree species used in this study. In Chapter 4, the economic viability of Caesalpinia velutina and Gliricidia sepium fuelwood plantations were evaluated in the smallholder context. To do so, the highest potential stumpage price a hypothetical buyer could offer without making a loss and the lowest price a smallholder would require in order to cover his or her capital and labour investment was analyzed. To conduct the analysis as a function of harvest age, growth and yield models for the tree species were developed and combined with product information and taper models to obtain yield by product segment. It was concluded that over long-enough rotations, both species could be economically viable endeavours. Major conclusions from this thesis are that the recognition and consideration of market-specific merchantability requirements and product segmentation, even within the fuelwood market, are very important for a thorough economic viability assessment. Furthermore, given that fuelwood value is affected by small-end log diameter, taper models provide a critical tool in valuing fuelwood resources. Furthermore, although 75  conventional measures of economic viability such as the IRR and NPV are considered an important pre-condition for smallholders to convert underutilized portions of their farms into fuelwood plantations, they are not a sufficient condition. Barriers to entry such as access to capital and technical knowledge combined with: the need for reaching economies of scale; the legal requirements for registering a tree plantation for commercial use; the complexity of identifying potential markets, organizing harvest and transportation logistics. Based on these constraints, it is unlikely that fuelwood plantations could be an economically viable endeavour for smallholders without institutional support. 5.1.  Future research Further research on evaluating the economic viability of establishing tree  plantations on marginal or underutilized portions of smallholders’ land is needed. In this thesis, a number of simplifying assumptions were made due to the limitations of the available data. If tree within plot level data was available, integrating the market effects of diameter distributions (i.e., parameter recovery) as opposed to simply using average values could further refine economic viability analyses. Furthermore, to improve accuracy: better growth and yield models are needed for other species used by smallholders; growth interaction effects between species should be taken into account, and tree branches and branch taper questions should be considered. From the smallholder’s perspective, little is known regarding the opportunity costs of the various competing land-use types and how this affects the food, fuel or fibre debate. Other major challenges in evaluating the economic viability of fuelwood or tree plantations in general is that market effects are unlikely to stay constant over the tree rotation period. As such, the prevailing market conditions that catalyzed the establishment of a tree plantation are unlikely to be the same at the time of harvest. As such, scenario analyses that evaluate the various potential market outcomes (energy, fibre or ecosystem services) could be used to further refine the analysis. These analyses should be further integrated within a broader landscape model to recognize competing land-use options. Measures of economic viability like the NPV use a discount rate to reflect the opportunity cost of capital. Given smallholders lack of access to capital, new measures of 76  economic viability for plantations should be developed with a greater emphasis on cash flow availability. With these considerations, carbon finance could be structured as a catalyst for business development. Further research is needed to identify the carbon price and the frequency of payments required so that sufficient cash be available to smallholders for fuelwood plantations to be viable.  77  Works cited Acuna, M. A., and Murphy, G. E. (2007). Estimating relative log prices of Douglas-fir through a financial analysis of the effects of wood density on lumber recovery and pulp yield. Forest Products Journal, 57(3), 60-65. Alcaldia de Limay. (2009). Programa de desarrollo de área Limay. Alcaldia de Limay, San Juan de Limay, Nicaragua. Arabatzis, G., and Klonaris, S. (2009). An analysis of Greek wood and wood product imports: Evidence from the linear quadratic aids. Forest Policy and Economics, 11(4), 266-270. doi:10.1016/j.forpol.2009.04.002 Arnold, M., Köhlin, G., Persson, R., and Shepherd, G. (2003). Fuelwood revisited: What has changed in the last decade? CIFOR, Jakarta. Asamblea Nacional de la Republica de Nicaragua. (2003) Ley No. 462, Ley de conservacion, Formento y Desarrollo Sostenible del Sector Forestal. Article 38. Retrieved June 8, 2012 from http://www.asamblea.gob.ni/index.php?option=com_wrapper&view=wrapper&Item id=360 Avohou, T. H., Houehounha, R., Glele-Kakai, R., Assogbadjo, A. E., and Sinsin, B. (2011). Firewood yield and profitability of a traditional Daniellia oliveri shortrotation coppice on fallow lands in Benin. Biomass and Bioenergy, 35(1), 562–571. doi:10.1016/j.biombioe.2010.10.030 Baker, K., Baumann, D., and Michaud, J. S. (2011). Mixed forest plantation: the Limay Community Carbon Project. Plan Vivo Technical Specification. Edinbourgh, Scotland. Baumann, D. (2012). Field data from the Limay Community Carbon Project. Taking Root, Montreal. 78  Brouwer, R., and Falcão, M. P. (2004). Wood fuel consumption in Maputo, Mozambique. Biomass and Bioenergy, 27(3), 233-245. doi:10.1016/j.biombioe.2004.01.005 Brown, A. G., Nambier, E. K. S., and Cossalter, C. (1997). Plantations for the tropics: their role extent and nature. In: Management of soil, water and nutrients. In: Tropical Plantation Forest. Australian Centre for International Agricultural Research. Canberra. ISBN: 1 86320 198 X Cannon, P. G., and Galloway, G. E. (1995). Fostering regional unity: The Madeleña Project of Central America. Journal of Forestry, 93(10), 32-35. CATIE. (1986a). Silvicultura de especies promisorias para producción de leña en América Central. Serie técnica 86 del Centro Agronomo Tropical de Investigacion y Enseñanza, Turrialba, Costa Rica. CATIE. (1986b). Crecimiento y rendimiento de especies para leña en areas secas y humedas de América Central. In: Serie técnica 79 del Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica. ISBN: 9977-951-86-1. Castellon, E. (2012). Personal Communication. APRODEIN, Somoto, Nicaragua. Chavarria, M. R. (2002). Estado de la Informacion Forestal en Nicaragua. In: Estado Actual de la Informacion Sobre Madera para Energia. Food and Agriculture Organization of the United Nations, Santiago Clutter, J. L., Fortson, J. C., Pienaar, L. V., Brister, G. H., and Bailey, R. L. (1983). Timber management: a quantitative approach. John Wiley and Sons, New York. Cordero, J., Boshier, D.H., Barrance, A., Beer. J., Boshier, D.H., Chamberlain, J., Detlefsen, G., Finegan, B., Galloway, G., Gómez, M., Gordon, J., Hands, M., Hellin, J., Hughes, C., Ibrahim, M., Kass, D., Leakey, R., Mesén, F., Monter, T. (2003a). Caesalpinia velutina. In: Arboles de Centroamérica: un manual para extensionistas. Oxford Forestry Institute and CATIE. Turrialba, Costa Rica. ISBN: 0 85074 161 0  79  Cordero, J., Boshier, D.H., Barrance, A., Beer. J., Boshier, D.H., Chamberlain, J., Detlefsen, G., Finegan, B., Galloway, G., Gómez, M., Gordon, J., Hands, M., Hellin, J., Hughes, C., Ibrahim, M., Kass, D., Leakey, R., Mesén, F., Monter, T. (2003b). Gliricidia sepium. In: Arboles de Centroamérica: un manual para extensionistas. Oxford Forestry Institute and CATIE. Turrialba, Costa Rica. ISBN: 0 85074 161 0 Creswell, J. W. (2008). Research design: qualitative, quantitative, and mixed methods approaches (3rd ed.). Sage, London. Davis, K. P. (1966). Forest management regulation and valuation (2nd ed.). McGrawHill, New York. Del Lungo, A., Ball, J., and Carle, J. (2006). Global planted forests thematic study: results and analysis. In: Planted Forests and Trees. FAO Working Paper 38, Rome. de Montalembert, M. R., and Clement, J. (1983). Fuelwood supplies in the developing countries. FAO Forestry Paper 42, Rome. Di Lucca, C. M. (1998). TASS / SYLVER / TIPSY: systems for predicting the impact of silvicultural practices on yield, lumber value, economic return and other benefits. In: Stand Density Management Conference: Using the Planning Tools. November 23-24, 1998, Colin R. Bamsey. Clear Lake Ltd., Edmonton, pp. 7-16. Díaz, H., García, C., Salinas, J., Valle, J., and Guillén, J. (1997). Resultados de 10 años de investigacion silvicultural del proyecto madeleña en El Salvador. Informe Técnico del Centro Agronómico Tropical de Investigación y Enseñanza. Turrialba, Costa Rica. ESMAP. (2010). Sustainable production of commercial woodfuel: lessons and guidance from two strategies. World Bank, Washington, DC.  80  Falconer, J. (1990). The major significance of “minor” forest products: The local use and value of forests in the West African humid forest zone. Food and Agriculture Organization of the United Nations, Rome. FAO. (1994). Origins and characteristics of the principal forage tree legume species. In: Forage tree legumes in tropical agriculture. Tropical grassland society of Australia Inc., Queensland. ISBN: 0 9585677 1 9 FAO. (2003). Case study: Identifying hot spots for fuelwood consumption in Mexico. In: Woodfuels Integrated Supply/Demand Review Mapping. Food and Agriculture Organization of the United Nations, Rome. FAO. (2004). Nicaragua: forests and the forestry sector. Food and Agriculture Organization of the United Nations. Retrieved May 12, 2012 from http://www.fao.org/forestry/country/57478/en/nic/ FAO. (2005). Economic viability, critical requirement for sustainable forest management. FAO Newsroom. Rome. Retrieved July 15, 2012 from http://www.fao.org/newsroom/en/news/2005/100241/index.html FAO. (2007). Approaches to linking producers to markets. Agricultural Management, Marketing and Finance Occasional Paper. Food and Agriculture Organization of the United Nations, Rome. FAO. (2010a). What woodfuels can do to mitigate climate change. FAO Forestry Paper 162, Rome. FAO. (2010b). Making the most of agricultural investment: A survey of business models that provide opportunities for smallholders: Survey of Business. Food and Agriculture Organization of the United Nations, Rome. FAO. (2010c). Global survey of agricultural mitigation projects. Mitigation of Climate Change in Agriculture Series. Food and Agriculture Organization of the United Nations, Rome. 81  FAO. (2010d). FAOSTAT Online Database. Food and Agriculture Organization of the United Nations. Retrieved March 15, 2012 from http://faostat.fao.org/ Felker, P., and Guevara, J. C. (2003). Potential of commercial hardwood forestry plantations in arid lands - an economic analyses of Prosopis lumber production in Argentina and the United States. Forest Ecology and Management, 186(1-3), 271– 286. doi:10.1016/S0378-1127(03)00280-9 GERSON R.L. (2010). Directorio de empresas de rosquillas de los municipios de Yalaguina y Somoto en el departamento de Madriz. Cooperativa Multisectorial GERSON R.L., Somoto, Nicaragua. Gezan, S. A., and Ortega, A. (2009). Modelos fustales para renovales de roble, raulí y coigüe en Chile. Bosque, 30(2), 61-69. doi: 10.4067/S0717-92002009000200001 Gonzales, J., Sandoval, C., and Zavala, M. (1997). Resultados de 10 años de investigacion silvicultural del proyecto madeleña en Honduras. Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica. Griess, V. C., and Knoke, T. (2010). Can native tree species plantations in Panama compete with Teak plantations? An economic estimation. New Forests, 41(1), 13– 39. doi:10.1007/s11056-010-9207-y Guevara, M. (2004). Informe nacional nicaragua. Estudio de tendencias y perspectivas del Sector Forestal en América Latina al año 2020. Food and Agriculture Organization of the United Nations, Rome. Hall, J. S., Ashton, M. S., Garen, E. J., and Jose, S. (2011a). The ecology and ecosystem services of native trees: Implications for reforestation and land restoration in Mesoamerica. Forest Ecology and Management, 261(10), 1553-1557. doi:10.1016/j.foreco.2010.12.011 Hall, J. S., Love, B. E., Garen, E. J., Slusser, J. L., Saltonstall, K., Mathias, S., van Breugel, M., et al. (2011b). Tree plantations on farms: Evaluating growth and 82  potential for success. Forest Ecology and Management, 261(10), 1675-1683. doi:10.1016/j.foreco.2010.09.042 Hedge, R., and Bull, G. Q. (2011). Socio-economics of Miombo woodland resource use: A household level study in Mozambique. In: Managing the Miombo woodlands of southern Africa: Policies, incentives and options for the rural poor. World Bank, Technical Annex 2, Washington, D.C. Heidarsson, L. (2011). Taper functions for lodgepole pine (Pinus contorta) and Siberian larch (Larix sibirica) in Iceland. Icelandic agricultural sciences, 24, 3-11. doi:1670567X Herrera, M. E. (1990). Analisis del comportamiento de 12 especies arboreas de uso multiple en Guapiles de Costa Rica. Master's thesis. Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica. Hughell, D. A. (1990). Modelos para la Prediccion del Crecimiento y Rendimiento de Eucalyptus camaldulensis, Gliricidium sepium, Guazuma ulmifolia y Leucaena leucocephala en America Central. Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica. Hurtarte, E. O. (1990). Comportamiento en Plantacion de Mangium (Acacia mangium willd) y Aripin (Caesalpinia velutina (ByR) Standl) En America Central. Master's thesis. Centro Agronómico Tropical de Investigación y Enseñanza, Turrialba, Costa Rica. Husch, B., Miller, C. I., and Beers, T. W. (1972). Forest mensuration (2nd ed.). The Ronald Press Co., New York. IEA. (2009). Bioenergy – a sustainable and reliable energy source: a review of status and prospects. International Energy Agency, Paris.  83  IMF. (2011). World Economic Outlook Database. World Economic Outlook. Retrieved March 14 2012 from http://www.imf.org/external/pubs/ft/weo/2011/01/weodata/index.aspx INIFOM. (2002a). Ficha técnica San Juan de Limay. Alcaldia de Limay. Retrieved February 8, 2012 from http://www.inifom.gob.ni/municipios/documentos/ESTELI/sn_juan_limay.pdf INIFOM. (2002b). Ficha Municipal - Somoto. Managua. Retrieved February 16, 2012 from http://www.inifom.gob.ni/municipios/municipios_Madriz.html INIFOM. (2002c). Ficha Municipal - Yalagüina. Managua. Retrieved February 16, 2012 from http://www.inifom.gob.ni/municipios/municipios_Madriz.html Jones, J. R., and Otarola, A. (1981). Diagnostico socio-economico sobre el consumo y produccion de leña en fincas pequeñas de Nicaragua. Informe Técnico No. 21 del CATIE, Turrialba, Costa Rica. Karathanssis, G. (1980). The net present value method under conditions of anticipated inflation. Aston University Management Centre, Birmingham, U.K. Khamzina, A., Lamers, J. P. A., and Vlek, P. L. G. (2012). Conversion of degraded cropland to tree plantations for ecosystem and livelihood benefits. Springer Netherlands. doi:10.1007/978-94-007-1963-7_15 Kozak, A. (2004). My last words on taper equations. Forestry Chronicle, 80(4), 507–515. doi: 10.5558/tfc80507-4 Kozak, A., Munro, D. D., and Smith, J. H. G. (1969). Taper Functions and their application in forest inventory. The Forestry Chronicle, 45(4), 278-283. doi:10.5558 Kozak, A., and Smith, J. (1993). Standards for evaluating taper estimating systems. The Forestry Chronicle, 69(4), 438–444. doi: 10.5558/tfc69438-4  84  Kozak, A. (1988). A variable-exponent taper equation. Canadian Journal of Forest Research, 18(11), 1363–1368. Kutner, M. H., Nachtsheim, C. J., Neter, J., and Li, W. (2005). Applied linear statistical models (5th Ed.). McGraw-Hill, New York. Larson, E., Rodriguez, L., and Azevedo de Rezende, T. (1994). Farm forestry in Brazil. BioResources ’94: Biomass resources: a means to sustainable development. Conference proceddings, Bangalore, India, October 3-7. Li, C., Barclay, H. J., Hans, H., Liu, J., Klos, R., and Carlson, G. (2010). Accuracy in population estimation: A methodological consideration. Ecological Complexity, 7(2), 208-211. doi:10.1016/j.ecocom.2010.02.002 Li, R., and Weiskittel, A. (2010). Comparison of model forms for estimating stem taper and volume in the primary conifer species of the North American Acadian region. Annals of Forest Science, 67(3), 302-302. doi:10.1051/forest/2009109 Littell, R. C., Milliken, G. A., Stroup, W. W., Wolfinger, R. D., and Schabenberger, O. (2006). SAS for mixed models (2nd Ed.). SAS Institute Inc, Cary, NC. MARENA. (2006). Flora Útil: Etnobotánica de Nicaragua. Ministerio del Ambiente y los Recursos Naturales (MARENA), Managua. Maravilla, E., and Vázquez, W. (1995). Crecimiento y rendimiento de Gliricidia sepium bajo 4 densidades iniciales de plantación en Tipitapa, Nicaragua. Silvoenergia, (60), 26. Markelova, H., Meinzen-Dick, R., Hellin, J., and Dohrn, S. (2009). Collective action for smallholder market access. Food Policy, 34(1), 1–7. doi:10.1016/j.foodpol.2008.10.001  85  Martin, F. (2012). User guide to the economic extension (ECON) of the forest vegetation simulator. Internal Report. U.S. Department of Agriculture, Forest Services, Forest Management Service Center. Fort Collins, CO. Max, T. A., and Burkhart, H. E. (1976). Segmented polynomial regression applied to taper equations. Forest Science, 22(3), 283–289. Mayers, J. (2006). Poverty reduction through commercial forestry: What evidence ? What prospects? The Forest Dialogue no. 2. Yale University, New Haven. Mayorga, G., and Urbina, M. (1993). Extraccion, comercializacion y consumo de leña en la subcuenca D - Cuenca Sur del Lago de Managua. Master's Thesis. Faculdad de Recursos Naturales y del Ambiente. Universidad Nacional Agraria, Managua, Nicaragua. McCrary, J. K., Walsh, B., and Hammett, A. L. (2005). Species, sources, seasonality, and sustainability of fuelwood commercialization in Masaya, Nicaragua. Forest Ecology and Management, 205(1-3), 299-309. doi:10.1016/j.foreco.2004.10.012 MITRAB. (2012). Acta Acuerdo Salario Mínimo 2012. Ministerio del Trabajo (Nicaragua), Managua. Niskanen, A. (1998). Financial and economic profitability of reforestation in Thailand. Forest Ecology and Management, 104, 57-68. OAS. (1997). Manejo Ambiental y Desarrollo Sostenible de la Cuenca del Río San Juan. Secretaría General de la Organización de Estados Americanos. Washigton, DC. Otárola, T., and Ugalde, L. (1983). Productividad y tablas de biomasa de Gliricidia sepium (Jacq.) Steud en bosque naturales de Nicaragua. Centro Agronomo Tropical de Investigacion y Enseñanza, Turrialba, Costa Rica. Özçelik, R., Brooks, J. R., and Jiang, L. (2010). Modeling stem profile of Lebanon cedar, Brutian pine, and Cilicica fir in Southern Turkey using nonlinear mixed-effects 86  models. European Journal of Forest Research, 130(4), 613–621. doi:10.1007/s10342-010-0453-5 PROLEÑA. (2000). Alternativas viables para solucionar el problema de demanda de leña en la región de Las Segovias (ciudades: Estelí, Condega, Somoto y Ocotal). PROLEÑA, Managua. Pachauri, S., and Spreng, D. (2003). Energy use and energy access in relation to poverty. CEPE Working Paper #25, Zürich. Pinherio, J. C., and Bates, D. M. (2000). Mixed-effects models in S and S-Plus. Spinger, New York. Piotto, D., Montagnini, F., Kanninen, M., Ugalde, L., and Viquez, E. (2004a). Forest plantations in Costa Rica and Nicaragua: performance of species and preferences of farmers. Journal of sustainable forestry, 18(4), 59–59. doi: 10.1300/J091v18n04_04 Piotto, D., Víqez, E., Montagnini, F., Kanninen, M. (2004b). Pure and mixed forest plantations with native species of the dry tropics of Costa Rica: a comparison of growth and productivity. Forest Ecology and Management, 190(2-3), 359–372. doi:10.1016/j.foreco.2003.11.005 Piotto, D., Craven, D., Montagnini, F., and Alice, F. (2010) Silvicultural and economic  aspects of pure and mixed native tree species plantations on degraded pasturelands in humid Costa Rica. New Forests, 39, 369-385. doi:10.1007/s11056-009-9177-0 Poulton, C., Dorward, A., and Kydd, J. (2010). The future of small farms: New directions for services, institutions, and intermediation. World Development, 38(10), 1413– 1428. doi:10.1016/j.worlddev.2009.06.009 Ramadhani, T., Otsyina, R., and Franzel, S. (2002). Improving household incomes and reducing deforestation using rotational woodlots in Tabora district, Tanzania.  87  Agriculture, Ecosystems and Environment, 89(3), 229-239. doi:10.1016/S01678809(01)00165-7 Ramos, M. A., and Paulino De Albuquerque, U. (2012). The domestic use of firewood in rural communities of the Caatinga: How seasonality interferes with patterns of firewood collection. Biomass and Bioenergy, 39, 147-158. doi:10.1016/j.biombioe.2012.01.003 Rathbun, L., LeMay, V., and Smith, N. (2011). Diameter growth under varying silvicultural treatments for mixed-species stands of coastal British Columbia using a random coefficients modelling approach. Ecological Modelling, 222(14), 2234– 2248. doi:10.1016/j.ecolmodel.2011.04.004 Regmi, B. N. (2003). Contribution of agroforestry for rural livelihoods: A case of Dhading District, Nepal. Rural Livelihoods, Forests and Biodiversity. CIFOR, Bonn, Germany. Roberts, D. (2008). Convergence of the fuel, food and fibre markets: a forest sector perspective. International Forestry Review, 10(1), 81-94. doi:10.1505/ifor.10.1.81 Rothery, Julian E. (1945). Some aspects of appraising standing timber. Journal of Forestry, 43(7), 490–498. SAS Institute Inc. (2008). SAS/STAT 9.2 User’s Guide. SAS Institute Inc., Cary, N.C. Sanchez-Azofeifa, A., Portillo, C., Fernandes, G. W., Stoner, K., and Shimizu, T. (2011). The Policy Process for Land Use/Cover Change and Forest Degradation in the SemiArid Latin American/Caribbean Region: Perspectives and opportunities. A literature review prepared for the Inter-American Development Bank. The University of Alberta, Edmonton. Schabenberger, O., and Pierce, F. J. (2002). Contemporary statistical models for the plant and soil sciences (p. 139). Taylor and Francis, New York.  88  Siregar, U. J., Rachmi, a., Massijaya, M. Y., Ishibashi, N., and Ando, K. (2007). Economic analysis of sengon (Paraserianthes falcataria) community forest plantation, a fast growing species in East Java, Indonesia. Forest Policy and Economics, 9(7), 822–829. doi:10.1016/j.forpol.2006.03.014 Staudhammer, C., and LeMay, V. (2000). Height prediction equations using diameter and stand density measures. The Forestry Chronicle, 76(2), 303–309. doi:10.5558/tfc76303-2 Stille, L., Smeets, E., Wicke, B., Singh, R., and Singh, G. (2011). The economic performance of four agro-forestry systems on alkaline soils in the state of Haryana in India. Energy for Sustainable Development, 15(4), 388-397. doi:10.1016/j.esd.2011.08.002 Street, W. V., and Price, S. (2009). Advancing poverty reduction and rural livelihoods through sustainable commercial forestry. The Forest Dialogue no 4. Yale University, New Haven. Taking Root. (2012). Taking Root database. Accessed March 2012. Montreal, Canada. Taylor, R. G. (1991). Optimum plantation planting density and rotation age based on financial risk and return. Forest science, 37(3), 886–902. Townson, I. M. (1995). Forest products and household incomes: A review and annotated bibliography. Tropical Forestry Paper 31. Oxford Forestry Institute, Oxford. Urbina, R. (2005). Inventario de politicas agroambientales en Nicaragua. Instituto Interamericano de Cooperacion para la Agricultura (IICA), San Jose, Costa Rica. Vallejo, A., Kanninen, M., Montero, M., and Viquez, E. (2006). Silvia: a flexible system for integrated plantation management. In: Multipurpose trees in the tropics: management and improvement strategies. Scientific Publishers. Jodhpur, India. van den Broek, R., van Wijk, A., and Turkenburg, W. (2000a). Farm-based versus 89  industrial eucalyptus plantations for electricity generation in Nicaragua. Biomass and Bioenergy, 19(5), 295-310. doi:10.1016/S0961-9534(00)00035-0 van den Broek, R., van den Burg, T., van Wijk, A., and Turkenburg, W. (2000b). Electricity generation from eucalyptus and bagasse by sugar mills in Nicaragua: A comparison with fuel oil electricity generation on the basis of costs, macroeconomic impacts and environmental emissions. Biomass and Bioenergy, 19(5), 311–335. doi:10.1016/S0961-9534(00)00034-9 Wicke, B., Smeets, E., Watson, H., and Faaij, A. (2011). The current bioenergy production potential of semi-arid and arid regions in sub-Saharan Africa. Biomass and Bioenergy, 35(7), 2773-2786. doi:10.1016/j.biombioe.2011.03.010 Wishnie, M., Dent, D., Mariscal, E., Deago, J., Cedeno, N., Ibarra, D., Condit, R., et al. (2007). Initial performance and reforestation potential of 24 tropical tree species planted across a precipitation gradient in the Republic of Panama. Forest Ecology and Management, 243(1), 39-49. doi:10.1016/j.foreco.2007.02.001 World Energy Council. (2007). 2007 Survey of Energy Resources. World Energy Council, London. Yang, R. C., Kozak, A., and Smith, J. H. G. (1978). The potential of Weibull-type functions as flexible growth curves. Canadian Journal of Forest Research, 8(4), 424– 431. doi:10.1139/x78-062 Zwcic, Z., Stankic, I., Vusic, D., Bosner, A., and Jaksic, D. (2009). Volume Utilization and Value of Timber Assortments of Dried Silver Fir (Abies alba Mill.) Trees. Šumarski list, 133(1), 27-37. ISSN: 0373-1332 Yin, R. K. (2002). Case Study Research: Design and Methods (3rd Ed.). Sage, London.  90  Appendix Appendix A: Interview  Identification Key:  1. Basic Company Information 1.1. 1.2. 1.3. 1.4.  Date of interview Average weekly production Type of oven Name of interviewer  2. Firewood Requirements and Preferences 2.1. Please list all the different tree species that you purchase for firewood _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________  _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________ _______________________  2.2. Of the species that you listed, are there species that you prefer to purchase? If yes, which ones? _______________________ _______________________ _______________________ _______________________ _______________________  _______________________ _______________________ _______________________ _______________________ _______________________  2.3. Are there species that you will not purchase? If yes, which ones? 91  _______________________  _______________________  _______________________ _______________________ _______________________ _______________________  _______________________ _______________________ _______________________ _______________________  2.4. If you listed any species in 2.3, why don’t you purchase the species listed? _______________________________________________________________ _______________________________________________________________ _______________________________________________________________ _______________________________________________________________ _______________________________________________________________  2.5. What is the minimum length and base end diameter a piece of wood has to have for you to purchase it? Minimum length:________________________________________ Minimum base end diameter: _____________________________  2.6. Are there other features of the log (e.g., rot, external damage) that would preclude you from purchasing it? _________________________________________________________ _________________________________________________________ _________________________________________________________ _________________________________________________________ 2.7. Given that different sizes of wood have different values, what are the diameter class (base end of log) categories that command different prices? Example: 92  Category name  Minimum length  a) Sawmill scraps b) Small sticks c) Medium logs d) Large logs  50 cm N/A 1 metre 1 metre  Category name  Base End Diameter range > 2cm 2 to 5 cm 5 to 10 cm 10 cm +  Minimum length  Diameter range  a) b) c) d) e) f) g)  3. Price and Price Variation 3.1. On average, how much did you pay last month for the wood categories that you just listed above per purchasing unit? Category a) b) c) d) e) f) g)  Cartload C$ C$ C$ C$ C$ C$ C$  Donkey load C$ C$ C$ C$ C$ C$ C$  Truck Load C$ C$ C$ C$ C$ C$ C$  Other (____________) C$ C$ C$ C$ C$ C$ C$  3.2. Did the price you paid vary much from one supplier to the other for the same wood category? (Yes or no) _______________________ If yes, what factors affected the price you paid? _______________________  _______________________  _______________________ _______________________ _______________________  _______________________ _______________________ _______________________  93  3.3. For exactly the same type of wood of the same wood categories, how much would you have paid if it were the rainy season for each wood category per purchasing unit?  Category a) b) c) d) e) f) g)  Cartload C$ C$ C$ C$ C$ C$ C$  Donkey load C$ C$ C$ C$ C$ C$ C$  Truck Load C$ C$ C$ C$ C$ C$ C$  Other (____________) C$ C$ C$ C$ C$ C$ C$  Comments:____________________________________________  3.4. Which months of the year do you pay the higher price? _______________________  _______________________  _______________________ _______________________  _______________________ _______________________  _______________________  _______________________  Comments:____________________________________________ ______________________________________________________ ______________________________________________________ 3.5. In your experience, have wood prices gone up, down or remained constant over the last 5 years? ______________________________________________________ Comments:____________________________________________ ______________________________________________________ ______________________________________________________ 3.6. For exactly the same type of wood of the same wood categories, to 94  the best of your memory, how much did you pay per purchasing unit in 2007 (5 years ago)?  Category a) b) c) d) e) f) g)  Cartload C$ C$ C$ C$ C$ C$ C$  Donkey load C$ C$ C$ C$ C$ C$ C$  Truck Load C$ C$ C$ C$ C$ C$ C$  Other (____________) C$ C$ C$ C$ C$ C$ C$  3.7. Do you expect wood prices to go up, down or remain constant in the next 5 years? Why? ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________  95  4. Wood Demand 4.1. How many units of firewood do you purchase per month on average for each wood category in the rainy season and in the dry season? Dry season: Category  Cartload  Donkey load  Truck Load  Other (____________)  Cartload  Donkey load  Truck Load  Other (____________)  a) b) c) d) e) f) g) Wet season: Category a) b) c) d) e) f) g)  Comments:____________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________ ______________________________________________________  96  5. Wood Supply 5.1. Of the wood that you purchased last year, to the best of your knowledge, what were the different sources of the wood you purchased? Example: Source name a) Forest b) Tree plantation c) Sawmill  Source name a) b) c) d) e) f) g)  5.2. Of these sources given in 5.1, approximately what percentage of your supply comes from the different sources during the dry season and during the wet season? Dry season: Source a) b) c) d) e) f) g)  Estimated percentage (%)  Wet season: Source a) b) c) d) e) f) g)  Estimated percentage (%)  97  5.3. What type of person / entity sells you your wood and what percentage of your wood comes from the different types of suppliers? Example:  Supplier type  Percentage of supply (%)  a) Private companies b) Independent individuals working full time c) Independent individuals working seasonally  Supplier type  Percentage of supply (%)  a) b) c) d) e) f) g)  Thank you!  98  

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.24.1-0073182/manifest

Comment

Related Items