Title: | Estimation of fresh water inflow to bays from gaged and ungaged watersheds |
Authors: | Lee, T., R. Srinivasan, J. Moon and N. Omani |
Year: | 2011 |
Journal: | Applied Engineering in Agriculture |
Volume (Issue): | 27(6) |
Pages: | 917-923 |
Article ID: | |
DOI: | 10.13031/2013.40631 |
URL (non-DOI journals): | |
Model: | SWAT |
Broad Application Category: | hydrologic only |
Primary Application Category: | model and/or data comparison |
Secondary Application Category: | hydrologic assessment |
Watershed Description: | 16,100 km^2 Galveston Bay and 11,600 Matagorda Bay watersheds in southern Texas, U.S. |
Calibration Summary: | |
Validation Summary: | |
General Comments: | This paper is part of a special SWAT issue that includes 12 SWAT-related papers in Trans. ASABE 2011 Vol. 54(5) and four additional papers in Applied Engr. in Agricultural 2011 Vol. 27(6) |
Abstract: | The long‐term estimation of fresh water inflow to coastal bays is important for understanding and managing
estuarine coastal ecosystems. The Texas Water Development Board (TWDB) has estimated the total fresh water inflow to bays in Texas using the TxRR (Texas Rainfall‐Runoff) model, which is a simple rainfall‐runoff relation model. Recently, TWDB requested to develop and apply the SWAT model using up‐to‐date technologies for estimating inflow to the bays. Two watersheds were selected for a pilot study; one represents an urbanized watershed draining into Galveston Bay (Galveston watershed) and the other represents a rural watershed draining into Matagorda Bay (Matagorda watershed). Two separate SWAT models were developed, one for each watershed. Weather data from weather stations were enhanced and
adjusted using NEXRAD (Next Generation Radar) precipitation data. Model calibration and validation was conducted using daily flow observations from USGS stream gage stations (gaged) and the same parameter settings were applied to the rest of the watersheds (ungaged). The total fresh water inflow to the bays by the SWAT model was compared to the estimation by the TxRR model. The daily streamflow calibration at each gage station showed an acceptable coefficient of determination (r2) ranging from 0.496 to 0.736 with Nash‐Sutcliffe coefficient (NS) ranging from 0.372 to 0.643. The correlation and NS for model validation, however, did not show a good agreement and the possible explanation can be applying recent landuse data for model runs for earlier years. The monthly streamflow estimation showed much better agreement between observed and modeled flows; r2 for calibration ranged from 0.647 to 0.916 and NS ranged from 0.613 to 0.941. The correlation for validation ranged from 0.485 to 0.694 for r2 and from 0.461 to 0.772 for NS. The comparison of the SWAT and TxRR models' estimation showed a good agreement in monthly total inflow to the bays. The coefficient of determination between the monthly estimations in the Galveston and Matagorda watersheds by the two models was 0.948 and 0.900, respectively. |
Language: | English |
Keywords: | Inflow to bay, Large watershed, SWAT, TxRR, Ungaged watershed |