Spatially explicit analysis of land use change: a case study for Ecuador

Introduction and objectivesWithin agricultural research increasing attention is paid to the integrated study of agro-ecosystems in order to address issues related to sustainable food production at the eco-regional level. This has been stimulated by the awareness that the world-wide demand for food will continue to increase while at the same there is high pressure on natural resources needed for food production, such as suitable soils and available water. Human-driven land use change is also relevant for global change because of its influence on greenhouse gas emissions, water and energy balances, and biodiversity.These issues have confronted research with substantive methodological challenges, such as the integration of biophysical and socio-economic disciplines over various spatio-temporal scales, and the development of modelling approaches for the exploration of future changes in land use and their effects.The general objective of this thesis is the analysis of spatial variation and temporal dynamics of agricultural land use systems in order to quantitatively assess the interaction between land use and the natural resource base. This is addressed through three derived objectives. Firstly, the spatial analysis of land use systems with the aim to detect the main biophysical and socio-economic drivers of actual land use at different spatial scales. Secondly, the spatially explicit modelling of near-future land use change dynamics, taking into account the multi-scale structure of actual land use and its drivers. Thirdly, the quantification of possible effects of future land use change on the natural resource base and agricultural production.The study areaThe study area is the South-American country of Ecuador. Agriculture is important in Ecuador, both for the production of subsistence crops as well as export crops. The country is characterised by a high agro-ecological diversity. A wide variety of climate and soil conditions exist and land use is diverse, with respect to the crops that are grown as well as the technology levels that are used. Agricultural land use is dynamic due to resource degradation, changes in demand for agricultural products, migration, export and import developments, economic developments and sector policies. A number of land use developments are causing serious threats to the natural resource base.Statistical multi-scale analysis of actual land use systemsLand use in Ecuador was investigated for the year 1991 by means of statistical analysis with the purpose of deriving quantitative estimates of the relative areas of land use types on the basis of biogeophysical, socio-economic and infrastructural conditions (Chapter 2). The smallest spatial units of investigation were 5 by 5 minute (9.25 x 9.25 km) cells of a geographical grid covering the whole country. Through aggregation of these cells, a total of six artificial aggregation levels was obtained with the aim of analysing spatial scale dependence of land use structure. For all aggregation levels independent multiple regression models were constructed for the estimation of areas within cells of permanent crops, temporary crops, grassland and natural vegetation.The variables used in the regression models were selected from a set of potential land use drivers or their proxies. A spatial stratification was applied by dividing the country into three main eco-regions. The results showed that at higher aggregation levels, the independent variables explained more of the variance in areas of land use types. In most cases, biogeophysical, socio-economic as well as infrastructural variables were important for the explanation of land use. The variables included in the models and their relative importance varied between land use types and eco-regions. Also within one eco-region the model variables varied with aggregation level. The results demonstrated spatial scale dependence of land use drivers.Spatially explicit modelling of land use change dynamicsA spatially explicit multi-scale land use change model was explained and demonstrated in Chapter 3. Important inputs for the model were the results of the multi-scale system analysis of Chapter 2. The model consists of two main modules: the demand module and the allocation module. Changes in the national demand for agricultural commodities are estimated with the demand module. The sub-national changes in land use following changes in demand are calculated in a multi-scale allocation module. The finest resolution for which changes were calculated for Ecuador, were the 5 by 5 minute cells. The allocation module takes into account the non-linear complexities of land use systems by dealing with competition between land use types and interconnectivity of landscape elements. Phenomena that operate over large distances like the influence of urban centres, the proximity to infrastructure and migration fluxes are addressed.The allocation of land use changes was validated by modelling backwards from the year 1991 to the year 1974, a year for which an independent data set was available. The validation results showed a rather good agreement with actual land use data at the level of administrative units called cantons. A hypothetical future base-line scenario of increasing demands for agricultural commodities was used to demonstrate how dynamics of land use are modelled. The results indicated "hot-spots", areas with potentially highly dynamic land use change where impacts of land use change on the natural resource base can be expected.Nutrient balances as indicators of sustainabilityUsing a nutrient balance model, inputs and outputs of nitrogen, phosphorus and potassium were estimated for each cell for the main land use types for the year 1991 (Chapter 4). Inputs considered were mineral fertiliser, organic fertiliser, atmospheric deposition, biological N-fixation and sedimentation; outputs considered were harvested product, removed crop residues, leaching, gaseous losses and erosion. The outcomes for cells were aggregated to sub-national and national level.In general, the estimates showed a depletion of the soil nutrient stock in Ecuadorian agro-ecosystems. Nationally, for temporary crops there was mainly a deficit of nitrogen (42 kg ha <sup>-1</SUP>yr <sup>-1</SUP>), while for permanent crops both nitrogen and potassium balances were clearly negative (40 and 25 kg ha <sup>-1</SUP>yr <sup>-1</SUP>, respectively). For grassland overall, losses are smaller. Erosion is a major cause of nitrogen loss but leaching and denitrification also contribute significantly. In permanent crops relatively large amounts of potassium leave the agro-ecosystem through harvested products, due to high potassium concentrations in these products and high yields.At sub-national scale, nutrient depletion under current land use is more severe in the Andean region than the coastal region, mainly as a result of higher erosion losses. The Amazon region is still largely unexploited but this study suggests that the current conversion of forest to agricultural land may cause serious nutrient balance problems at a local level.Exploration through scenariosThe land use change model for Ecuador was used for the dynamic and spatially explicit exploration of near future agricultural land use changes (Chapter 5). A number of plausible scenarios were formulated for the period until the year 2010. At the national level, different developments in national food demand were defined on the basis of assumptions for population growth, consumption patterns and export developments. At the sub-national level the protection of nature parks and land use restrictions due to land degradation were evaluated with respect to their possible spatial impacts on the land use change dynamics within the country. Under the assumptions of the demand scenarios, the area agricultural land expanded significantly, resulting in more use of land in existing agricultural areas and frontier-type expansion into rather undisturbed natural areas. The patterns of change depended on the increase in demand, competition between land use types, changes in the drivers of land use, and the area of land that was excluded from agricultural use.Soil fertility impacts were considered by linking the results of the nutrient balances calculated in Chapter 4 with the results of the land use change model. The results indicated potential negative effects of land use changes on the soil nutrient balance. It was argued that quantification of land use dynamics at the landscape level can support research and policy aimed at understanding the drivers of land use change and the behaviour of complex agro-ecosystems under changing conditions at different spatial scales. This way, issues dealing with sustainable food production and the management of the natural resource base can be addressed in a more integrated and quantitative manner.Spatial and temporal characterisation of Andean potato production systemsChapter 6 described how a statistical analysis of the spatial structure of cropping systems can be combined with the land use change model in order to characterise a specific cropping system. The method was applied for potato production systems in the Ecuadorian Andes. The variation of areas and yields of potato was analysed on the basis of the data for the grid-cells. The results showed a large spatial variability in conditions of potato production systems. The combinations of biophysical and socio-economic variables that best explained the spatial structure of potato areas as well as their yields depended on the geographic area analysed. The whole Andean region was analysed, as well as the provinces Carchi and Chimborazo. Output of the land use change model was used to evaluate the position of potato production systems within the general dynamics of the main land use types. In this way, areas were located where pressure on potato production systems is probable, due to expansion of other agricultural land uses in response to increasing national food demands. Especially competition with grasslands is expected in the near future.When potato-growing areas are pushed into marginal areas, this is likely to have negative consequences for potato productivity. With the results a typology can be given of production systems in relation to their environment. Such information can be used to direct agro-ecological research at specific interactions between environment and crop management.Conclusions<UL>A multi-scale system analysis demonstrated scale dependence of Ecuadorian land use.The results of the multi-scale system analysis could be integrated in a spatially explicit, dynamic, land use change model.Scenarios studies with the model allowed for the exploration of possible future land use changes.The quantitative, spatially explicit, information on land use changes generated by the model offers possibilities for the assessment of the effects of land use change on sustainability of food production, biodiversity and global climate change.Priorities in further methodology development are linkage with geophysical landscape processes and a better incorporation of crop specific management.