Uncertainty of Coupled Soil-Vegetation-Atmosphere Modelling Methods for Estimating Groundwater Recharge
Abstract
Coupled soil-vegetation-atmosphere models have become increasingly popular for estimating groundwater recharge, because of the integration of carbon, energy and water balances. The carbon and energy balances act to constrain the water balance and as a result should reduce the uncertainty of groundwater recharge estimates. However, the addition of carbon and energy balances also introduces a large number of plant physiological parameters which complicates the estimation of groundwater recharge. Moreover, this method often relies on existing pedotransfer functions to derive soil water retention curve parameters and saturated hydraulic conductivity from soil attribute data. The choice of a pedotransfer function is usually subjective and several pedotransfer functions may be fit for the purpose. These different pedotransfer functions (and thus the uncertainty of soil water retention curve parameters and saturated hydraulic conductivity) are likely to increase the prediction uncertainty of recharge estimates. In this study, we aim to assess the potential uncertainty of groundwater recharge when using a coupled soil-vegetation-atmosphere modelling method. The widely used WAter Vegetation Energy and Solute (WAVES) modelling code was used to perform simulations of different water balances in order to estimate groundwater recharge in the Campaspe catchment in southeast Australia. We carefully determined the ranges of the vegetation parameters based upon a literature review. We also assessed a number of existing pedotransfer functions and selected the four most appropriate. Then the Monte Carlo analysis approach was employed to examine potential uncertainties introduced by different types of errors. Preliminary results suggest that for a mean rainfall of about 500 mm/y and annual pasture vegetation, the estimated recharge may range from 10 to 150 mm/y due to the uncertainty in vegetation parameters. This upper bound of the recharge range may double to 300 mm/y if different pedotransfer functions are used. However, this uncertainty range can be narrowed when field measurements such as soil moisture, leaf area index or evapotranspiration are used to constrain the model. This study calls for careful consideration of parameter uncertainties and the need to obtain field measurements to calibrate the model.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2016
- Bibcode:
- 2016AGUFM.H23K..04X
- Keywords:
-
- 1829 Groundwater hydrology;
- HYDROLOGYDE: 1834 Human impacts;
- HYDROLOGYDE: 1910 Data assimilation;
- integration and fusion;
- INFORMATICS