Sensitivity of pan-Arctic terrestrial net primary productivity simulations to daily surface meteorology from NCEP-NCAR and ERA-40 reanalyses

Ke Zhang, The University of Montana
John S. Kimball, University of Montana - Missoula
Maosheng Zhao
Walter C. Oechel
John Cassano
Steven W. Running, University of Montana - Missoula


We applied a terrestrial net primary production (NPP) model driven by satellite remote sensing observations of vegetation properties and daily surface meteorology from the 45-year ECMWF Re-Analysis (ERA-40) and the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP-NCAR) reanalysis (NNR) to assess NPP spatial and temporal variability for the pan-Arctic basin and Alaska from 1982 to 2000. Sensitivity analysis of the production efficiency model (PEM) to uncertainties in surface meteorological inputs indicate that ERA-40 solar radiation and NNR solar radiation and surface temperatures are the primary sources of PEM-based NPP uncertainty for the region. Considerable positive bias in solar radiation inputs relative to surface observation networks resulted in overprediction of annual NPP by approximately 35.2 and 61.6% using ERA-40 and NNR inputs, respectively. Despite these uncertainties, both reanalysis products captured the major annual anomalies and trends in surface meteorology for the domain. The two reanalysis products also produced similar NPP spatial patterns for 74.7% of the domain, and similar annual anomalies and temporal trends, though there were significant regional differences particularly for Eurasia. A simple correction method based on a sensitivity experiment between reanalysis and surface station meteorological measurements produced generally consistent NPP results that were considerably smaller than PEM simulations derived from uncorrected reanalysis drivers. The results of this study identify major sources of uncertainty in reanalysis-based surface meteorology, and associated impacts on regional NPP simulations of the northern high latitudes.