Climate interpolation for land resource and land use studies in mountainous regions

Researchers in the field of land resources and land use are increasingly faced with a serious data constraint. New techniques like simulation models require detailed and quantitative data on climate and soils. Large mapping units with representative weather stations or representative soil profiles ignore an important part of the inherent spatial variability of the landscape.New techniques involving geographical information systems (GIS), geostatistics, and remote sensing open new opportunities.A good example in the field of soil science arethe rapid advances in digital soil mapping. The main objective of this research is to resolve the data constraint related to the meteorological data through the interpolation of meteorological data from weather stations and to explore the value of the newly created datasets in the field of land resource and land use studies. Research focused on two areas in the Peruvian and Ecuadorian Andean highlands: (i) La Encañada and Tambomayo watersheds in Peru (2950 to 4000 meters above sea level) where agriculture is performed in marginal areas; and (ii) Chitan and San Gabriel watersheds in Ecuador (2700 to 3840 meters above sea level) where agriculture is commercially oriented to the production of potato and milk. During the first step, empiric and physical process-based models for climate interpolation were developed, calibrated, evaluated and validated to be applied under topography-complex terrains. The use of seasonal-climate forecast from Global Circulation Models' (GCM) was also analyzed to explore the feasibility to support decision makers. Effects of detailed-spatial climate information as a soil-forming factor were analyzed as a basis for digital soil mapping to disaggregate soil mapping units, and the question of how much information do we really need for land resources and land use studies was assessed. As results, two physical process-based models to interpolate maximum and minimum temperatures, and rainfall were developed and four empirical models to estimate incident solar radiation were evaluated and calibrated. The feasibility to produce crop-yield's forecast by downscaling GCM's outputs was assessed by linking statistical and crop models aggregated at watershed level. Theory of soil forming factors applicable at large scales was demonstrated to be applicable also at small scales by using the detailed interpolated climate information. Finally, an analysis with the Tradeoff Analysis system to assess the impact of different resolutions of input data gave as a result that there is not a thumb rule to establish the resolution's threshold. To each study case a sensitivity analysis must be performed to specifically establish the threshold resolution according to the project objectives, the biophysical and economic model to be used, and to the necessities of the different stakeholders among other main factors.