干旱区生态环境知识资源中心

Forecasts of the states and fluxes of terrestrial ecosystems are an increasingly important tool for a large fire, famine, irrigation, energy, recreation, and agriculture community. A detailed understanding of the relative importance of vegetation phenology and meteorology, two of the main forcings of ecosystem forecasts, and the likely impact of errors in phenological and (or) meteorological forecasts are required prior to management implementation. Using the terrestrial observation and prediction system (TOPS) and 1982-1997 leaf area index (LAI) and meteorology for the conterminous United States, we tested the relative importance of interannual variability in meteorology and LAI for soil water simulations. In nearly all cases, meteorological variability influenced simulations far more than did LAI; the effects of ignoring realistic variability in either variable were most pronounced in arid, low-LAI regions. We then identified the critical meteorological forecast errors in temperature and precipitation that were required to generate statistically significant differences in 1-week soil water forecasts. Temperature critical errors approached 10 degreesC in winter but were only about 2-3 degreesC in summer. Precipitation critical errors were much more constant throughout the year and were usually less than 1 cm (error in weekly total precipitation).