Work

Finding the Optimum Unsaturated Zone Storage Change Estimation Method to Improve Water Accounting in Vietnam

Year: 2016
Type: dissertation
Abstract: Water Accounting Plus or WA+ was developed as an answer to the ever growing need of a reliable, basinscale, moisture accounting.The main attraction of WA+ is its ability to draw an explicit relationship between water users and available water resources.By doing this, it is able to improve the understanding of human and nature interactions in the field of water resource management.Similar to other moisture accounting methods, the computation in WA+ starts with the tracking of various hydrological stocks that are present in the area of interest.The accuracy and reliability of these inputs will significantly influence the performance of WA+.The terrestrial storage change or ∆S is one of those required inputs.The term itself is a combination of a surface component (∆S SW ), a groundwater component (∆S GW ), and an unsaturated zone component (∆S U Z ).Since it is very complicated to directly measure ∆S U Z , many moisture accountings implement a certain estimation method to get an approximation of the real value.Among the more popular ∆S estimation methods are hydrological models (i.e.WaterPix, STREAM, and PCRGlobwb) and land surface model (i.e.GLDAS-Noah).In addition to these, there are also empirical relationships that are used to derive the unsaturated zone soil moisture from the surface condition.A more uncommon method is to use the Budyko Curve approach on to solve an unsteady water balance.This thesis aims to find the optimal (∆S U Z ) estimation method for the application within WA+.The Ca River Basin and the Downstream Red River Basin were chosen as the study areas.A period of 4 years, from December 2006 to 2010 was chosen as the simulation period.Each method was run on a monthly time step to facilitate the need of a monthly application of WA+.The methods mentioned above were evaluated based on their ability to regenerate the observed streamflow at available measuring stations in both basins.An additional check was also done by comparing the estimated ∆S U Z from each method to the streamflow-derived ∆S U Z .A second evaluation, based on the spatial distribution of land use and land cover classes, was done to see which of the methods is able to reflect the variation of ∆S U Z based on the land use and land cover types.The results of both evaluation processes suggest that the downscaled products of the PCRGlobwb model are the most accurate method to estimate ∆S U Z .However, it is not the most practical since the original spatial resolution of the outcomes of this model is relatively coarse (i.e. 10 km) which will be a problem if the study area is classified as a small to medium basin.On the other hand, although the statistical performance of the WaterPix model and the empirical relationship are quite low during the evaluation process, these two methods show promising potential for application within WA+, considering their flexible computing structure and their ability to produce outcomes at a fine spatial resolution (i.e.250m).Further investigation should test the accuracy of these methods in regions with other hydrological characteristics.Using the same evaluation processes, the results from the Budyko Curve method are found to be the least accurate.
Language: English
Open Access status: green
OpenAlex ID: W3119027688
Cites: 37
Cited by:
Related to: 10