REGIONAL ESTIMATION OF NITROGEN MINERALIZATION IN FOREST ECOSYSTEMS USING GEOGRAPHIC INFORMATION-SYSTEMS

A knowledge-based geographic information system (GIS) model was develo ped and used to predict net nitrogen mineralization within forest ecos ystems of the Midwestern Great Lakes region, USA (Illinois, Indiana, M ichigan, Ohio, and Wisconsin). Climate, soil, and forest inventory dat a were used in conjunction with data relating initial N and lignin con centrations of leaf litter. Net N mineralization (N-net) from leaf lit ter of forest ecosystems of the entire region was predicted as a funct ion of litter quality (N:C ratio), annual actual evapotranspiration, s oil texture, and litter production. Regional variation was evident in the model results: N-net decreased from rates exceeding 120 kg.ha(-1). yr(-1) in the deciduous forest soils of southern Illinois, Indiana, an d Ohio to 20 kg.ha(-1).yr(-1) in the coniferous forest soils of northe rn and eastern Michigan. Wisconsin's forest soils had intermediate N-n et rates, predominantly 40-90 kg.ha(-1)yr(-1). The model was most sens itive to N concentration of litter, which indicated that litter qualit y of plant species is the most important factor controlling spatial di stribution of N mineralization of ecosystems, even at the regional sca le. The GIS model in this research provides a means for scaling betwee n stand and regional scales. The model was a practical approach for es timating net N mineralization in forest ecosystems at the regional sca le. Such models will become increasingly important in simulating the i mpacts of global environmental change on forest ecosystems.