- Title
- Modelling changing catchment under the climate variability: a case study from a semi-arid catchment in the upper basin of the Goulburn River
- Creator
- Binesh, A.; Yeo, I. Y.
- Relation
- 22nd International Congress on Modelling and Simulation (MODSIM2017). MODSIM2017, 22nd International Congress on Modelling and Simulation. Modelling and Simulation Society of Australia and New Zealand (Hobart, Tas. 3-8 December, 2017) p. 1656-1662
- Relation
- https://www.mssanz.org.au/modsim2017/index.html
- Publisher
- Modelling and Simulation Society of Australia and New Zealand (MODSIM)
- Resource Type
- conference paper
- Date
- 2017
- Description
- Hydrological response of arid and semi-arid regions to climate fluctuations are highly variable and predicting induced hydrological variability is extremely challenging. This study investigated the temporal variability of streamflow (SF), and its relationship with soil moisture (SM) in a semi-arid region at the catchment scale, using SWAT (Soil and Water Assessment Tool). SWAT is a continuous semi-distributed model, which simulates the spatial variability of SM, considering the spatial heterogeneity of a catchment such as land cover, soils, and slope, and predicts SF after the routing process. SWAT has been widely used to evaluate the long term impacts of land management practices on water resources in agricultural landscapes. Despite its wide application, the capability of SWAT has not been fully realized in the (semi-) arid region. It is largely because the underpinning concept of SWAT in runoff generation lies on the USDA Soil Conservation Service - curve number (SCS-CN) method. The CN method is limited by its empirical origin. Based on the infiltration loss model, it does not consider long term water losses such as evapotranspiration (ET) and evaporation, the most crucial hydrological term in a water limited environment. Several approaches have been made to improve SM estimation for a continuous modeling (William et al., 2012). The main focus was given to improve estimates of the water retention term (S) based on the soil characteristics (known as Direct Soil Moisture Index, DSMI) or relating it to the potential ET (hence varying it with accumulated plant ET) (referred to as Revised Soil Moisture Index, RSMI). Considering the potential of improved CN methods on estimating SM, we apply SWAT to predict and understand hydrologic behaviors of a semi-arid catchment, located within the Goulburn River catchment in the Upper Hunter Region of NSW. The catchment has experienced an extreme climate variability and shown noticeable changes in hydrologic behaviors, in particular SF pattern over last two decades. We report improved hydrological prediction of a changing catchment by: (1) better accounting for the spatial variability of precipitation, soil characteristics, and antecedent SM condition (using DSMI and RSMI methods), and (2) calibrating model parameters over the period of climate fluctuations (both dry and wet periods) for SF. SWAT was set up at a monthly time scale to simulate SM and SF over the period of 2008-2014 using DSMI and RSMI methods. We used SWAT embedded sensitivity analysis and SUFI-2 to calibrate the model over 2008-2012 and validated it over 2013-2014. Calibration contained both wet and dry climatic conditions, however validation only contained the wet period. Sensitivity analysis revealed that both surface runoff and ET were sensitive processes for both the RSMI and DSMI methods to generate SF. However, soil characteristics became sensitive only for DSMI method. The calibrated model performance for SF estimation was reasonable based on both methods, but RSMI provided improved overall SM prediction. The calibrated models provided NSE (Nash-Sutcliff efficiency) values of 0.60-0.76 over the calibration and validation periods, with ~ 11 % of RMSE (root mean squared error). Overall, DSMI method predicted SF slightly better during both of wet and dry period, while RSMI provided better estimation of SM for the study period. In addition, we noted the improvement on the spatial variability of precipitation enhanced SM prediction. Most inconsistent results between simulated and observed SM were from those areas with mismatching soil characteristics between available soil information and actual site conditions. This study shows the potential of using a simple, semi-distributed catchment scale model (calibrated based on SF) to predict SM and SF, and to investigate the threshold for SF generation for the catchment with a changing behaviour in a semi-arid region. Accurate representation of precipitation, climate variations and soil characteristics were crucial to reduce the prediction errors.
- Subject
- SWAT; streamflow; soil moisture; semi-arid region; hydrological variability
- Identifier
- http://hdl.handle.net/1959.13/1386747
- Identifier
- uon:32456
- Identifier
- ISBN:9780987214379
- Language
- eng
- Full Text
- Reviewed
- Hits: 1019
- Visitors: 1271
- Downloads: 134
| Thumbnail | File | Description | Size | Format | |||
|---|---|---|---|---|---|---|---|
| View Details Download | ATTACHMENT01 | Publisher version (open access) | 1 MB | Adobe Acrobat PDF | View Details Download |