1.1 Forest plot studies of wet-canopy evaporation: losses during large and extreme rainfall events
Forest plot studies estimate Ewc , using a canopy water balance (CWB), as the difference between the gross rainfall (𝑃𝑔) incident upon a vegetation canopy and the fraction of 𝑃𝑔 that reaches the ground asnet rainfall (Pn). Net rainfall comprises rainfall that bypasses or drips from the canopy (throughfall : TF ) and that which flows via stems and trunks (stem flow : SF ). As noted above, very few studies have focused on CWB estimatedEwc during large (> 50 mm d-1 of 𝑃𝑔) or extreme (> 150 mm d-1 of 𝑃𝑔: Collier, Fox & Hand, 2002) rainfall events. A notable exception is the work of Keim, Skaugset, Link and IroumΓ© (2004) who report Ewc losses above 30% of 𝑃𝑔 at temperate sites in Chile and Northwest USA. Equally high Ewc losses during large magnitude rainfall events at other locations with a temperate climate have been reported (e.g. see Deguchiet al ., 2006 and Hashino et al ., 2002). Taken at β€œface value”, these Ewc losses appear to be potentially significant in the context of flooding: removal of such large fractions of event rainfall from a catchment system are likely to have a significant effect on a flood hydrograph where tree planting covers a large proportion of a catchment (Hankin et al ., 2017). Consequently, there is an apparent disparity between the publications which conclude that forest effects on flood peaks are likely to be small or insignificant for large and extreme events and the CWB observations from forest plot studies.
The significance of forest Ewc for flood mitigation depends upon the difference in Ewc between a given forest canopy and another land cover. For example, moorland vegetation species such as Heather (Calluna vulgaris ) tend to have long-termevapotranspiration losses approximately 30% to 60% of that for tall forests and short semi-natural grasses lose around 10 to 40% of that for forests (e.g. Calder, 1976, Calder & Newson, 1979; Calderet al ., 1981). However, as most comparative studies derive estimates from catchment or lysimeter water balances over relatively long periods, and do not separate Ewc and transpiration losses, these estimates are of limited use when considering individual events. Additionally, evaporation from a forest understorey or soil litter layer can be significant (Bulcock & Jewitt 2012; Carlisle et al ., 1967; Gerrits et al ., 2007, 2010) and effectively increase the difference in Ewc loss (e.g. between forest and grass). We therefore assume that Ewc losses from tall canopies are likely to be significantly higher than for short vegetation under meteorological conditions favourable for wet-canopy evaporation and, consequently, that the absolute magnitude of Ewc from tall forest canopies is of primary relevance here. Thus it is important to determine the full extent of evidence for significant Ewc from forest canopies during large and extreme rainfall events as well as an understanding of the meteorological conditions under which significant losses might be supported.