Modifying a dynamic global vegetation model for simulating large spatial scale land surface water balances

摘要: Satellite-based data, such as vegetation type and fractional cover, are widely used in hydrologic models to prescribe the state a study region. Dynamic global (DGVM) simulate land surface hydrology. Incorporation of satellite-based data into DGVM may enhance model's ability hydrology by reducing task model parameterization providing distributed information on characteristics. The objectives this (i) modify for simulating water balances; (ii) evaluate modified actual evapotranspiration (ET), soil moisture, runoff at regional or watershed scales; (iii) gain insight both original large spatial scale To achieve these objectives, we introduce "LPJ-hydrology" (LH) which incorporates Lund-Potsdam-Jena (LPJ) DGVM. ran LH using historical (1981–2006) climate covers 2.5 arc-min grid cells conterminous US entire world coarser cover data. We evaluated simulated ET, set observed different scales. Our results demonstrate that patterns LH-simulated annual ET accordance with previously published US; LH-modeled monthly stream flow 12 major rivers was consistent values respectively during years 1981–2006 ( R 2 > 0.46, p 0.52). modeled mean discharges 10 worldwide also agreed well (differences = 0.79, < 0.01) 1984–2001. Overall, justifies feasibility incorporating reliability land-surface balances. better estimate surface/river mid-to-high latitudes, recommended LPJ-DGVM considers effects solar radiation snowmelt.