ABSTRACT
Nocturnal water temperature (Tw) affects the behaviour of aquatic biota and metabolism of whole rivers. However, night-time water temperature (nTw) is poorly understood because spot samples are typically taken during daylight hours, or Tw series are aggregated in ways that mask sub-daily properties. This paper examines 15-minute measurements of Tw and air temperature (Ta) collected at 36 sites in the Rivers Dove and Manifold, English Peak District. Data were stratified by day and night then analysed using hysteresis, auto-correlation and logistic regression techniques. Daily hysteresis loops show lagged responses between nTw and previous daylight air temperatures (dTa), plus the influence of groundwater and discharge variations. Logistic regression models were modified using a seasonal factor and explained between 80 and 94% of the variance in daily maximum nTw; minimum nTw were predicted with less skill, particularly for headwater sites in summer. Downstream variations in model parameters also reflect the influence of groundwater and/or riparian shade, and prevailing weather conditions. A case is presented where an intense summer storm resulted in the propagation of a thermal wave that produced maximum Tw at some sites during hours of darkness. Hence, our findings show that Tw management by riparian shade has to be seen in a catchment wide context, with anticipated benefits normalised for weather variability, extreme rainfall events, local influence of groundwater, and channel structures.