Estimating regional forest productivity and water yield using an ecosystem model linked to a GIS

We used the PnET-II model of forest carbon and water balances to estimate regional forest productivity and runoff for the northeastern United States. The model was run at 30 arc sec resolution (approximately 1 km) in conjunction with a Geographic Information System that contained monthly climate data and a satellite-derived land cover map. Predicted net primary production (NPP) ranged from 700 to 1450 g m2 yr1 with a regional mean of 1084 g m2 yr1. Validation at a number of locations within the region showed close agreement between predicted and observed values. Disagreement at two sites was proportional to differences between measured foliar N concentrations and values used in the model. Predicted runoff ranged from 24 to 150 cm yr1with a regional mean of 63 cm yr1. Predictions agreed well with observed values from U.S. Geologic Survey watersheds across the region although there was a slight bias towards overprediction at high elevations and underprediction at lower elevations.