Premium
Regression models for daily stream temperature simulation: case studies for the river Elbe, Germany
Author(s) -
Koch Hagen,
Grünewald Uwe
Publication year - 2010
Publication title -
hydrological processes
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.222
H-Index - 161
eISSN - 1099-1085
pISSN - 0885-6087
DOI - 10.1002/hyp.7814
Subject(s) - mean squared error , environmental science , streams , hydrology (agriculture) , coefficient of determination , calibration , regression , linear regression , regression analysis , scale (ratio) , statistics , mathematics , computer science , geology , geography , computer network , geotechnical engineering , cartography
Daily stream temperatures are needed in a number of analyses. Such analyses might focus on aquatic organisms or industrial activities. To protect aquatic systems, industrial activities, for example, water withdrawals or discharges, are sometimes restricted. To evaluate where new industrial settings should be placed or if climate change will affect already existing industrial settings, the simulation of stream temperature is needed. Stream temperature models with weekly or monthly time scale might not be sufficient for this kind of analysis. Different regression models to simulate daily stream temperature for the river Elbe (Germany) are developed and their performance is estimated. For the calibration period the Nash–Sutcliffe coefficient (NSC) for the simplest model is 0·97, and the root mean squared error (RMSE) is 1·48 °C. For the most sophisticated model the NSC also is 0·97. However, the RMSE is 1·32 °C. For the validation period the NSC for the simplest model is 0·96, and the RMSE is 1·45 °C. The NSC for the most sophisticated model is 0·97, and the RMSE is 1·25 °C. Copyright © 2010 John Wiley & Sons, Ltd.