Climate variability and implications for keeping rivers cool in England

• Published in: Climate Risk Management
• Main Authors: R.L. Wilby, M.F. Johnson
• Format: Article in Journal/Newspaper
• Language: English
• Published: Elsevier 2020
• Subjects: Water temperature; North Atlantic Oscillation; Climate variability; England; Logistic regression; Riparian shade; Meteorology. Climatology; QC851-999; Rho; North Atlantic; North Atlantic oscillation
• Online Access: https://doi.org/10.1016/j.crm.2020.100259; https://doaj.org/article/4ca085dfb06b41cabebe3896571585bd

Water temperature (Tw) is a primary determinant of river ecosystem health and function that is strongly controlled by climate variability and change but mediated by catchment properties. We apply a nested analysis to: (1) evaluate how annual and seasonal mean Tw varied across England during the period 2000–2018; (2) assess the extent to which these regional-temporal dynamics correlate with the North Atlantic Oscillation (NAO); and (3) quantify the impact of local climate variability on modelled daily maximum Tw for open, shaded and spring-fed river reaches. Such information is used to identify sentinel locations for long-term monitoring and reporting, to evaluate the true benefit of riparian shade management, and to assess the impacts of climate change on Tw. We draw on a national archive of nearly 1 million Tw values and data from a high-resolution field experiment in central England. Nationally, annual mean Tw changed by −0.4 °C/decade over the period 2000 to 2018, broadly in line with Central England Temperatures, although summer Tw changed by +0.6 to +1.1 °C/decade in parts of central and northern England. There were significant associations between summer Tw and NAO (rho = 0.64, p < 0.05), especially at sites above 300 m altitude (rho = 0.70, p < 0.01). The regional analysis reveals strongest links between summer Tw and NAO in northeast England and weakest associations in lowland regions of southern and east England with major aquifers. Hence, places with significant groundwater flows offer the greatest chance of detecting long-term signals in Tw that are not being driven by the NAO. Site-specific, logistic regression models of daily maximum Tw are found to be sensitive to the prevailing NAO phase during calibration periods. Such models show a thermal benefit for shaded sites compared with open sites that is one average 0.2 °C under negative NAO but 2.8 °C under positive NAO. Based on the findings from our nested analysis we suggest ways of optimising monitoring networks plus improving the appraisal ...