Seasonal variability of interception evaporation from the canopy of a mixed deciduous forest
Introduction
The rainfall interception loss (Iloss) of forested areas is an important process in the catchment water balance and has been the subject of many studies. Expressed as a percentage of gross rainfall (PG) the magnitude of Iloss varies strongly with forest structure, forest size and climate. Many of the earlier studies were carried out in forests in temperate regions, often in monospecific plantations, and reported Iloss totals between 25 and 50% in coniferous forests (Rutter et al., 1975, Gash et al., 1980, Johnson, 1990) and between 10 and 35% in broadleaved forests (Rutter et al., 1975, Rowe, 1983, Nizinski and Saugier, 1989). Due to temporal variations in the duration, frequency and magnitude of rainfall events, as well as in the evaporation rate, there are large seasonal and interannual variations in the fraction of intercepted rainfall within a site. Leyton et al. (1967) were among the first authors to point out the need to distinguish between climatic and structural factors when interpreting the interception loss. This motivated the development of various types of models which replaced the use of simple percentage values with physically more consistent estimation procedures of the interception loss from the gross rainfall (e.g. Rutter et al., 1971, Gash, 1979, Calder, 1986, Gash et al., 1995). Most widely and successfully used is the analytical model of Gash (1979) which requires daily rainfall data and site specific parameters such as the ratio of the average evaporation rate () to the average rainfall rate () for wet canopy conditions, the interception storage capacities of the canopy (S) and the tree trunks (St) and the fractions of the rainfall falling through gaps in the canopy (p) and being diverted to the trunks (pt). Once derived these parameters allow a reliable estimation of Iloss from PG and remove the need for continuous monitoring of the net rainfall. A small modification of the model made by Gash et al. (1995) also allows it to be used for sparse canopies.
Since the interception loss is calculated as the difference between two large numbers (gross and net rainfall), even a small fractional uncertainty in either of them can generate large errors in Iloss. To reduce such errors, attempts have been made to correct PG collected with elevated raingauges for potential losses in the catch due to locally increased turbulence (Rodda and Smith, 1986, McJannet et al., 2007) or to achieve a higher spatial representativeness of the net rainfall measurements by increasing the number of raingauges, moving them to new positions at regular intervals (Aussenac, 1968, Gash and Stewart, 1977, Lloyd and Marques, 1988, Holwerda et al., 2006) or using very large collectors (Calder and Rosier, 1976, Cuartas et al., 2007). These corrections are particularly relevant for deciduous forests in temperate regions that face large seasonal changes in canopy structure and weather patterns and often grow on sites which are not ideal for micrometeorological and hydrological measurements. For example, many of those woodlands are managed as continuous cover forestry and have no clearings where PG could be measured in a sheltered environment. All this causes large uncertainties in the determination of both PG and Pnet and the few attempts to derive structural parameters from the measured interception loss of temperate deciduous forests have been inconclusive (Halldin et al., 1984, Dolman, 1987, Hörmann et al., 1996) or restricted to the growing season only (Carlyle-Moses and Price, 1999, Price and Carlyle-Moses, 2003). Another problem is the notoriously difficult determination of the evaporation rate from the wet canopy (Stewart, 1977, Gash et al., 1999, Van der Tol et al., 2003, Herbst et al., 2006) which is often unrelated to the net radiation and has a strong influence on the interception loss (Gash et al., 1999). In mixed deciduous forests the spatial heterogeneity in the evaporation rate and the seasonal change in the canopy structure complicates the situation. This also means, in a more general sense, that a better description of the hydrological processes in such forests is needed.
The aim of this case study was to parameterise the Gash et al. (1995) sparse canopy rainfall interception model for a typical mixed deciduous forest throughout the complete annual cycle which to our knowledge has never been done before. A comparison with other components of the stand water balance which were measured independently serves as a validation of the results.
Section snippets
Research site
The study was carried out in Grimsbury Wood, near the town of Newbury in Berkshire, U.K., at 51°27′N, 1°16′W at an altitude of 115 m above sea level. The woodland covers about 350 ha and is made up of a mosaic of different plantations which include both broadleaved and coniferous stands. The research site is part of the hydrological monitoring network of the LOCAR research programme (Wheater et al., 2006). Here the forest canopy consists of an overstorey dominated by oak (Quercus robur L.) and
Meteorological conditions and evaporation rate
Some micrometeorological characteristics for the periods when the canopy was completely wet are shown in Table 1. Notable are the low net radiation and the high downward flux of sensible heat (H = −88 W m−2) in the leafless period. This shows that the energy input from irradiance was much less important for the evaporation of intercepted rain than the energy provided by cooling of the air above the canopy. An additional small gain in available energy resulted from a net release of stored energy due
PG and Pnet
The largest uncertainties in this study came from the recordings of PG in the leafless period, due to higher wind speeds in the winter in connection with a lower position of the zero plane. It seems plausible that both effects added to the observed catch loss, albeit an empirical regression with the monthly average of the wind speed during rainfall, regardless of the presence of leaves, was the only practical correction procedure. The recording of the throughfall turned out to be more
Acknowledgements
The work was part of the Lowland Catchment Research (LOCAR) programme funded by the National Environmental Research Council (NERC) through grant no. NER/T/S/2001/00939. It was proposed and initiated by the late John Roberts of CEH Wallingford.
References (48)
Estimating biomass in northern lower Michigan forest stands
For. Ecol. Manage.
(1982)A stochastic model of rainfall interception
J. Hydrol.
(1986)- et al.
The design of large plastic-sheet net rainfall gauges
J. Hydrol.
(1976) - et al.
A note on the dynamic calibration of tipping-bucket gauges
J. Hydrol.
(1978) - et al.
An evaluation of the Gash interception model in a northern hardwood stand
J. Hydrol.
(1999) - et al.
Interception water-partitioning dynamics for a pristine rainforest in Central Amazonia: marked differences between normal and dry years
Agric. For. Meteorol.
(2007) Estimates of roughness length and zero plane displacement for a foliated and non-foliated oak canopy
Agric. For. Meteorol.
(1986)Summer and winter rainfall interception in an oak forest. Predictions with an analytical and a numerical simulation model
J. Hydrol.
(1987)- et al.
The evaporation from Thetford forest during 1975
J. Hydrol.
(1977) - et al.
An application of the Rutter model to the estimation of the interception loss from Thetford forest
J. Hydrol.
(1978)