Predicting Latent and Sensible Heat Fluxes in Stream Temperature Models: Current Challenges and Potential Solutions

Abstract

Stream temperature studies typically use either the Penman combination equation (PCE) or empirical wind functions to compute the latent heat flux, which can be an important control on daily maximum water temperature. The sensible heat flux is usually computed from the latent heat flux via the Bowen ratio or a bulk transfer model. Unfortunately, both the PCE and empirical wind functions involve challenges. The PCE is inappropriate for application to subdaily stream evaporation because, as implemented in many stream temperature studies, it does not account for energy associated with warming or cooling of the water column or bed heat fluxes. The PCE thus tends to overestimate evaporation under conditions typical of warm summer days during the diurnal warming phase, which is when evaporation is likely to be an important heat loss mechanism. Empirical wind functions involve substantial uncertainty given the broad range of coefficient values that have been reported in the literature and the lack of above‐stream meteorological data to use in their application. We provide suggestions for improving current practice and also identify research directions to support development of robust approaches to computing the latent and sensible heat fluxes.

A primer on stream temperature processes

“…When applying an Eulerian framework (Box 2), longitudinal advection can occur due to water entering the stream segment of interest at the upstream boundary and leaving the stream segment at the downstream boundary. The importance of longitudinal advection will be proportional to the longitudinal temperature gradient (i.e., the rate at which temperature changes with downstream distance at a given time, Moore & Leach, 2021). High thermal gradient conditions tend to occur where there are significant and abrupt longitudinal changes in surface energy exchanges and discharge—for example, where riparian vegetation structure changes rapidly, such as downstream of a clearcut‐forest transition (Moore, Sutherland, et al, 2005), or downstream of a tributary.…”

Simulation and prediction of water temperature in a water transfer channel during winter periods using a new approach based on the wavelet noise reduction-deep learning method

“…By using the general physical equation to study the change of water temperature in rivers, Moore and Leach (2021) found that the heat flux was an important factor in controlling the change of water temperature. Generally, due to the complexity of boundary condition and environment aspects, the parameters in equations vary widely and become obvious uncertainty.…”