#L15 Runoff partitioning and its impact on water and energy budgets in the ACME land model

Abstract

Soil moisture plays an important role in the coupled water, energy, and carbon cycles. Surface and subsurface runoff are important boundary fluxes that influence soil moisture directly. These fluxes are parameterized in land surface models (LSMs) based on conceptual theories or physical laws, which describe the subgrid heterogeneities of land surface and subsurface runoff using probability distributions derived from subgrid topography or using analytical functional forms with parameters. To date, two runoff parameterizations are widely used in LSMs: the TOPMODEL and Variable Infiltration Capacity (VIC) formulations. The TOPMODEL formulation makes the following assumptions: (1) the saturated zone dynamics could be approximated by successive steady states; (2) the recharge rate to the water table over a catchment is homogeneous; (3) the hydraulic gradient of the saturated zone could be approximated by the local surface slope; and (4) the distribution of downslope transmissivity is assumed to be an exponential function of storage deficit or depth to the water table. In the VIC formulation, surface runoff generation is a function of the mean soil moisture capacity as well as its spatial heterogeneity in the shallow soil layers over a grid cell, and subsurface runoff generation can be approximated by a nonlinear relationship as a function of deep-layer soil moisture Therefore, theoretically TOPMODEL is more suitable under humid climate and mountainous areas with surface saturation and large relief, while VIC is more general but requires calibration of model parameters. However, performance of a LSM in simulating runoff and its partitioning is highly dependent not only on the implementation details of the runoff scheme, but also on its non-linear interactions with other LSM parameterizations.  In this study, numerical experiments were conducted using the ACME land model (ALM) with the TOPMODEL and VIC formulations. The model simulations are benchmarked against runoff fields from Global Runoff Data Centre, evapotranspiration datasets derived from FLUXNET and the Moderate Resolution Imaging Spectroradiometer, the terrestrial water storage data from the Gravity Recovery and Climate Experiment, and gridded mean annual base flow indices estimated from streamflow observations globally. Results from this study will be used to guide sensitivity analysis and calibration of ALM runoff parameters in the near future.