TY - JOUR
T1 - On the Robustness of Conceptual Rainfall-Runoff Models to Calibration and Evaluation Data Set Splits Selection
T2 - A Large Sample Investigation
AU - Guo, Danlu
AU - Zheng, Feifei
AU - Gupta, Hoshin
AU - Maier, Holger R.
N1 - Funding Information:
Professor Zheng is funded by the National Natural Science Foundation of China (Grant 51922096). Professor Gupta acknowledges partial support from the Australian Research Council through the Centre of Excellence for Climate System Science (Grant CE110001028). The historical streamflow data for all study catchments are available from the Australian Bureau of Meteorology (BoM) Hydrological Reference Stations, available online (http://www.bom.gov.au/water/hrs/). The historical rainfall and PET data were from the BoM Australian Water Availability Project (AWAP) (http://www.csiro.au/awap/). The authors would also like to thank Ms. Jie Jian for her great help on data input and Mr. Luis de la Fuente for his kind support to investigate how model robustness might be related to catchment characteristics.
Funding Information:
Professor Zheng is funded by the National Natural Science Foundation of China (Grant 51922096). Professor Gupta acknowledges partial support from the Australian Research Council through the Centre of Excellence for Climate System Science (Grant CE110001028). The historical streamflow data for all study catchments are available from the Australian Bureau of Meteorology (BoM) Hydrological Reference Stations, available online ( http://www.bom.gov.au/water/hrs/ ). The historical rainfall and PET data were from the BoM Australian Water Availability Project (AWAP) ( http://www.csiro.au/awap/ ). The authors would also like to thank Ms. Jie Jian for her great help on data input and Mr. Luis de la Fuente for his kind support to investigate how model robustness might be related to catchment characteristics.
Publisher Copyright:
© 2020. American Geophysical Union. All Rights Reserved.
PY - 2020/3/1
Y1 - 2020/3/1
N2 - Conceptual rainfall-runoff (CRR) models are widely used for runoff simulation and for prediction under a changing climate. The models are often calibrated with only a portion of all available data at a location and then evaluated independently with another part of the data for reliability assessment. Previous studies report a persistent decrease in CRR model performance when applying the calibrated model to the evaluation data. However, there remains a lack of comprehensive understanding about the nature of this “low transferability” problem and why it occurs. In this study we employ a large sample approach to investigate the robustness of CRR models across calibration/validation data splits. Specially, we investigate (1) how robust is CRR model performance across calibration/evaluation data splits, at catchments with a wide range of hydroclimatic conditions; and (2) is the robustness of model performance somehow related to the hydroclimatic characteristics of a catchment? We apply three widely used CRR models, GR4J, AWBM and IHACRE_CMD, to 163 Australian catchments having long-term historical data. Each model was calibrated and evaluated at each catchment, using a large number of data splits, resulting in a total of 929,160 calibrated models. Results show that (1) model performance generally exhibits poor robustness across calibration/evaluation data splits and (2) lower model robustness is correlated with specific catchment characteristics, such as higher runoff skewness and aridity, highly variable baseflow contribution, and lower rainfall-runoff ratio. These results provide a valuable benchmark for future model robustness assessments and useful guidance for model calibration and evaluation.
AB - Conceptual rainfall-runoff (CRR) models are widely used for runoff simulation and for prediction under a changing climate. The models are often calibrated with only a portion of all available data at a location and then evaluated independently with another part of the data for reliability assessment. Previous studies report a persistent decrease in CRR model performance when applying the calibrated model to the evaluation data. However, there remains a lack of comprehensive understanding about the nature of this “low transferability” problem and why it occurs. In this study we employ a large sample approach to investigate the robustness of CRR models across calibration/validation data splits. Specially, we investigate (1) how robust is CRR model performance across calibration/evaluation data splits, at catchments with a wide range of hydroclimatic conditions; and (2) is the robustness of model performance somehow related to the hydroclimatic characteristics of a catchment? We apply three widely used CRR models, GR4J, AWBM and IHACRE_CMD, to 163 Australian catchments having long-term historical data. Each model was calibrated and evaluated at each catchment, using a large number of data splits, resulting in a total of 929,160 calibrated models. Results show that (1) model performance generally exhibits poor robustness across calibration/evaluation data splits and (2) lower model robustness is correlated with specific catchment characteristics, such as higher runoff skewness and aridity, highly variable baseflow contribution, and lower rainfall-runoff ratio. These results provide a valuable benchmark for future model robustness assessments and useful guidance for model calibration and evaluation.
KW - calibration
KW - conceptual rainfall-runoff models
KW - data split
KW - evaluation
KW - robustness
KW - transferability
UR - http://www.scopus.com/inward/record.url?scp=85083072617&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85083072617&partnerID=8YFLogxK
U2 - 10.1029/2019WR026752
DO - 10.1029/2019WR026752
M3 - Article
AN - SCOPUS:85083072617
SN - 0043-1397
VL - 56
JO - Water Resources Research
JF - Water Resources Research
IS - 3
M1 - e2019WR026752
ER -