(A2)
Cities across the United States are focusing stormwater management efforts on control of nonpoint source pollution and flooding. Development in upstream portions of watersheds is increasing flooding hazard to established downstream communities. Urban stormwater runoff is the second most common source of water pollution for lakes and estuaries and the third most common source for rivers nationwide (EPA 1994). During normal rainfall, pollutants are washed from impervious surfaces, lawns, and other sources into streams and storm sewerage systems (Claytor and Schueler 1996). During heavy rainfall, excessive runoff can outstrip the storage capacity of storm sewerage systems and streams. Localized flooding is a frequent result, and pollutant loading can exceed desirable levels at receiving water bodies and treatment plants. Also, heavy runoff increases soil erosion, as well as the transport and downstream deposition of pollutant-laden sediment.
A healthy urban forest can mitigate stormwater impacts of urban development
(Sanders 1986; Lormand 1988). Trees intercept and store rainfall on leaves
and branch surfaces, thereby reducing runoff volumes and delaying the onset
of peak flows. Root growth and decomposition increase the capacity and
rate of soils to infiltrate rainfall and reduce overland flow. Urban forest
canopy cover reduces soil erosion by diminishing the impact of raindrops
on barren surfaces. This study focuses on interception of rainfall by Sacramento's
urban forest. Our objectives are to 1) quantify annual rainfall interception,
2) describe relations between interception and rainfall seasonality, duration,
and volume for typical storm events, and 3) identify important structural
traits of urban forests that can be manipulated to increase rainfall interception.
The simulation results reported above relied on application of models derived from TR-55 (Soil Conservation Service 1975). The TR-55 model and its adaptations are widely used to evaluate effects of land use change on runoff. However, they are limited in their capabilities to accurately estimate effects of urban forest management on runoff volume and peak rate. Some important limitations include the following.
2. Curve numbers were originally developed from 24-hour storm data and are assumed to be constant for a large range of rainfall events. Thus, TR-55 is better at predicting longer, larger storm events than smaller, shorter events (Pitt 1994). Because small storms are responsible for most annual urban runoff and pollutant washoff, accurate simulation of shorter events is important for water quality resource protection.
3. It is limited in computing the time of concentration and peak rates of flow for small catchments. This limits use of the model for flooding analysis.
4. Interception is held constant regardless of storm characteristics. Interception and depression storage (stormwater held in surface depressions) are modeled as storage capacities that are filled before overland flow begins. In fact, interception is a dynamic process, with canopy storage changing as water evaporates from the crown, drips from leaves, and flows down branches (Carder 1996).
Forest canopy interception has been studied in both laboratory and field experiments (Rutter et al. 1971; Aston 1979; Gash et al. 1995). In rural forests, Zinke (1967) found that 15% to 40% of annual gross precipitation can be lost by interception in conifer-dominated forests and 10% to 20% in hardwood-dominated forests. Interception may exceed 59% for old growth forest trees (Baldwin 1938). However, information on interception by open-grown urban trees is lacking.
Statistical models estimate interception as a linear proportion of gross precipitation (Horton 1919; Zinke 1967). Regression coefficients for statistical methods are difficult to obtain because they are site specific and a long historical data record is needed to derive these coefficients. ln contrast to the statistical approach, Rutter et al. (1975,1977) developed a physically based canopy interception model that computes the water balance of canopy and trunk components. This approach was successfully tested (Gash and Morton 1978; Lloyd et al. 1988) with data from a coniferous plantation in Great Britain. Based on the assumption that the time lag between rainfall events was long enough for the canopy surface to dry, an analytical model was developed by Gash (1979) that has a simple form and is easier to apply than Rutter's model. Some other physically based interception models (Calder 1977; Gash et al. 1980; Massman 1983) have been developed and applied to natural forests and found to produce results in agreement with field observed interception.
Forest-derived interception models may not be applicable to urban forests because both the microclimate and tree architecture of urban forests are different from those of rural forests. The gradient of microclimate can vary more quickly in urban forests than in rural forests. Microclimate differences affect evaporation rates, leaf drip, and other hydrologic processes in the tree crown. Compared with most rural forests, urban forests have fewer trees per unit area, tree size (dbh, diameter at breast height) that is larger on average, a more diverse mix of species with different phonological patterns, and greater spatial variation in canopy cover (McPherson 1998). Gash et al. (1995) found that existing interception models need to be reformulated for sparse forests.
In this study, a one-dimensional numerical model of rainfall interception was developed based on the previous work of Rutter et al. (1971) and Gash (1979). Rutter and Gash's model is physically based, and their parameters are easy to obtain. We used drying power of the air to estimate potential evaporation (Pruitt and Doorenbos 1977a, 1977b). Remotely sensed data and GIS techniques were used to characterize the land surface and link the model to specific local conditions.
Study site. Sacramento County is located in the lower Sacramento
Valley of California and falls within the coordinates between longitudes
W 121° 51' 43" and W 121°01' 20". For a more complete description
of the study area and sampling units, see McPherson 1998 (pages 175-177
of this issue).
|
(1)
|
Model parameterization and scale up. We assumed that total rainfall interception is the summation of interception for all trees. Further, we assumed that leaf surface temperature is in equilibrium with air temperature and that leaf surface area is constant throughout the leaf-on (mid-March to mid-November) and leaf-off periods. At the smallest scale, interception was calculated for each cell in a grid system of length dx (100 m [330 ft]) and dy (100 m) that was superimposed on the study area. Interception was analyzed at 2 spatial scales: SubRADs (Sub-Regional Assessment Districts) and sectors. Interception values were aggregated for each of the 71 SubRADs and for each of the 3 sectors. Three groups of parameters were estimated.
Tree canopy characterization. Aerial photos and ground surveys were used to estimate tree species composition, tree dimensions, crown projection area (area enclosed by the dripline), and leaf surface area by SubRAD (see McPherson 1998, beginning on page 175 of this issue, for a detailed explanation of methods). Vegetation was divided into 3 categories: tree, shrub, and grass. Trees were further divided into broadleaf evergreen, broadleaf deciduous, conifer, and palm. Tree canopy parameters included species, leaf area (McPherson 1998), shade coefficient (visual density of the crown from McPherson 1984), and tree height. Three tree height classes were established: large (> 15 m [50 ft]), intermediate (5 to 15 m [16.5 to 50 ft.]), and small (< 5 m). Tree height data were used to estimate wind speed at different heights above the ground and the resulting rates of evaporation (Jetten 1996). The volume of water stored in the tree crown was calculated from crown projection area (area under tree dripline), leaf area indexes (LAI, the ratio of leaf surface area to crown projection area), and water depth on the canopy surface. Species-specific shade coefficients influenced the amount of projected throughfall. Although rainfall is intercepted by trees, shrubs, and buildings, in this study we focused on rainfall interception by trees only.
Precipitation and potential evaporation. Scaling-up meteorological data from a limited number of stations to a region has been widely applied in hydrological and climatic studies (Hungerford et al. 1989; Ustin et al. 1996; Xiao 1997). A Kriging method (Edward and Srivastava 1989) was used to extrapolate precipitation and evaporation data from a meteorological base station to the entire study area. Precipitation and evaporation coefficients of each grid cell were estimated as the ratio of the value at the cell to the value at the base station based on the spatial data extrapolation results from Kriging. The Stonemead base station (38° 30' 31" N, 121° 17' 36" W, elevation 37 m [122 ft]) is located near the center of the study area and has been operated since 1982 by the California Department of Water Resources.
Meteorological parameters were derived based on data obtained from NOAA (National Oceanic and Atmospheric Administration), CIMIS (California Irrigation Management Information System), and CDEC (California Data Exchange Center) meteorological stations located in or near the study area. Mean precipitation (from 57 stations) and evaporation (40 stations) data from stations with more than 20 years of meteorological records were used to create long-term averages for their respective grid cells. These data, in conjunction with Kriging, allowed us to conduct simulations for a variety of time intervals and weather conditions.
Numerical simulation. This study focused on the spatial and temporal distribution of canopy interception in Sacramento County. Three sets of simulations were conducted.
Annual interception. Data for a typical meteorological year (determined to be 1992 based on analysis of 10 years meteorological data at Stonemead station) were used to simulate annual canopy rainfall interception over the entire study area. Among the total 30 storms in 1992 at Stonemead, 7 storms had precipitation greater than 25.4 mm (1 in.), and these events accounted for 77% of total annual precipitation. Seven storms had precipitation between 6.2 and 25.4 mm (0.25 to 1 in.), accounting for 17% of total annual precipitation. The remaining 16 events were each less than 6.2 mm and accounted for 6% of annual precipitation. We assumed that individual storms were separated by intervals of at least 24 hours without precipitation (Hamilton and Rowe 1949).
Summer and winter storm events. We simulated rainfall events occurring during summer (May 31,1993) and winter (December 3, 1994) to examine effects of tree species composition and size on interception. The summer event depicted interception when deciduous trees were in-leaf, while the winter event occurred during the leaf-off season. Our analysis was limited to 2 adjacent SubRADs in the northern part of the county with very different forest structures. The rural SubRAD (Rio Linda-Elverta, 12.3 km2 [4.7 mi2]) was dominated by relict native oaks and conifers (68% of the trees were broadleaf evergreen and 17% coniferous). The city SubRAD (North Sacramento, 15.7 km2 [6 mi2]) contained a diverse mix of introduced shade trees and conifers characteristic of established neighborhoods near downtown Sacramento (50% of the trees were broadleaf deciduous and 41% coniferous).
Flood events. Five additional storm events were selected to study interception for rainfall of different amounts and durations. Using the same two SubRADs as described above, we simulated precipitation events with return frequencies of 2, 5, 25,100, and 200 years to better understand the extent to which Sacramento's urban forest can mitigate flooding. Rainfall events were selected from Stonemead's 1990 to 1997 records based on depth-duration-frequency relationships developed by the local flood control agency (City/County of Sacramento 1996). We simulated interception assuming both leaf-on and leaf-off conditions for deciduous trees for both the rural and city SubRAD sites.
Simulation results (annual, seasonal, and flood events) are presented at the urban forest canopy level and landscape level. Interception at the urban forest canopy level is the percentage of total precipitation falling on the urban forest canopy that is intercepted by the canopy (mm3 interception per mm gross precipitation per mm2 crown projection area). Interception at the landscape level is the percentage of total precipitation falling on the entire study site that is intercepted by the urban forest canopy (mm3 interception per mm gross precipitation per mm2 total land area).
To reduce numerical estimation errors, interception processes were simulated
with an hourly time-step for analysis of annual interception and a 1-minute
timestep for seasonal and flooding events. Due to the relatively small
amount of stem surface area compared to leaf surface area (Vertessy et
al.1995) and low evaporation rate (Rutter and Morton 1977; Gash 1979),
evaporation from stem surfaces was ignored by forcing the stem surface
water storage capacity to zero.
At the urban forest canopy level, interception was strongly influenced by the mix of tree species and their phenology. Interception was lowest in the city sector, where broadleaf deciduous trees dominated and were leafless during the winter rainy season (Table 1). In the suburban sector, broadleaf evergreens and conifer trees accounted for 67% of total leaf area. In addition to maintaining foliage year-round, evergreens generally have higher LAIs than deciduous trees, thereby increasing canopy storage per unit crown projection area. Annual interception was as high as 22% for suburban SubRADs.
Summer and winter storm events. From the outset of the 6-hour, 12 mm (0.48 in.) summer storm (May 31, 1993), canopy storage increased until saturated after about 2 hours in the city SubRAD and about 2.5 hours in the rural SubRAD (4 mm [0.16 in.]) (Figure 2). Maximum canopy storage in the rural SubRAD was nearly twice that of the city SubRAD (4.5 and 2.3 mm [0.18 and 0.09 in.). For the next 2 hours of relatively heavy rainfall, most precipitation reached the ground as leaf drip, throughfall, and stem flow. From hours 4 to 6, the rainfall rate decreased and canopy storage gradually increased. After a continuously high leaf drip rate during hours 3 to 3.5 in the rural SubRAD, and hours 3.5 to 4 in the city SubRAD, canopy water storage was less than the maximum storage capacity. The small amount of rainfall added was not enough to fill canopy water storage to capacity. Once the rainfall stopped, canopy storage dropped and evaporation of intercepted rainfall began.
At the urban forest canopy level for the summer storm, interception loss was 36% and 18% for the rural and city SubRADs, respectively (Figure 2). Taller trees and more tree species with relatively high LAls in the rural than city SubRAD resulted in higher canopy storage and evaporation rates. More than 55% of trees in the rural SubRAD were large (tree height > 15 m [50 ft]) and the LAI was 6.1, while more than 58% of the trees in the city SubRAD were medium size (height between 10 and 15 m [33 and 50 ft.]) with LAls of 3.7 (Table 3).
The winter storm event (December 3, 1994) was much longer (44 hours) and larger (45 mm [1.78 in.]) than the summer event (Figure 3). Canopy storage steadily increased for about 6 hours, then declined once water began to drip off leaves and stems of saturated canopies. This pattern was repeated throughout the storm event as the canopy intercepted and lost rainfall in response to precipitation, leaf drip, and evaporation. It should be noted that evaporation rates were relatively low during the winter event. Compared to the summer event, air temperatures were cooler, relative humidity was higher, and net radiation was lower. Lower evaporation rates and lower LAI due to trees in a leaf-off condition (hence less canopy storage capacity) were primarily responsible for 14% (rural) to 26% (city) less interception during the winter event than the summer event.
At the urban forest canopy level, interception was 10% and 4%, respectively, in the rural and city SubRADs. Broadleaf deciduous trees were leafless in December, which reduced LAls to 5.2 and 1.8, respectively, for the rural and city SubRADs. During winter, condensation sometimes occurs on plant surfaces from dew and fog. Higher LAls and more evergreen trees in the rural compared to city SubRAD account for increased fog trapping and interception.
Total canopy interception for the winter event in rural and city SubRADs was 4,212 m3 (3.41 acf) and 6,103 m3 (4.95 acf), respectively. This volume of water would increase detention storage of a 1 km2 (247 ac) basin by a depth of 19 mm (0.75 in). Because tree crowns provide a type of detention storage, these results could be used as the basis for determining the economic value of canopy surface water storage.
Flood events. Canopy interception for 5 flooding events was greater for smaller, shorter storms than for larger and longer storm events (Table 4). During small events, a relatively large percentage of gross precipitation was required to fill canopy storage to capacity. Once storage was filled, relatively little precipitation was needed to maintain canopy saturation. Therefore, canopy interception had a minor impact on major flood events. For example, during the 200-year storm event, leaf-on interception loss was only 9% for the rural SubRAD and 5% for the city SubRAD (Table 4). In contrast, leaf-on interception was 37% and 20% for the 2-year event in the Rural and City SubRADs, respectively.
Differences between canopy interception for the leaf-on and leaf-off events reflected the impact of broadleaf evergreens and conifers in each SubRAD. Greatest interception loss occurred during the leaf-on season in both SubRADs. However, differences between leaf-off and leaf-on interception were greatest in the city, where leaf-on loss was about 70% to 100% greater than leaf-off loss due to the relative abundance of broadleaf deciduous trees (Table 3). The large evergreen component in the rural SubRAD accounted for a smaller seasonal difference of about 20%. As previously noted, greater overall leaf area in the rural versus city SubRAD was responsible for higher interception loss for all storm events.
Limitations of the model. This canopy interception model allows
water to drip from leaves only after canopy storage exceeds saturated canopy
storage. Because some leaf drip begins before canopy saturation, the model
overestimates actual interception. During winter rainfall, water stored
on stem surfaces is a large proportion of rainfall interception and temporary
canopy water storage. By ignoring stem surface water storage, the model
underestimates interception, especially for urban forest stands dominated
by deciduous trees. In this study, only rainfall interception by trees
is modeled. Shrubs and grasses also contribute to total interception. A
full water budget includes contributions from all vegetation layers. This
model has not been calibrated or validated with measured data from individual
trees or an urban watershed. Thus, findings are approximations.
At the urban forest canopy level, the mix of tree species and their size structures influenced interception. In Sacramento, evergreen trees played the most important role in interception because most precipitation occurs in winter. Large trees with evergreen foliage contribute to greater interception than smaller, deciduous trees. In many climates with summer precipitation, deciduous trees make a substantial contribution to rainfall interception. Planting trees, as well as maintaining existing trees in a healthy condition, will reduce the volume of stormwater runoff over the long term.
These results indicate that urban forests become increasingly less effective
at reducing stormwater runoff as the amount of precipitation per storm
increases. Although trees reduce runoff, they may not be very effective
for flood control. Floods usually occur during major storm events, well
after canopy storage has been exceeded. However, by substantially reducing
the amount of runoff during less extreme events, urban forests may protect
water quality. Small storms, for which urban forest interception is greatest,
are responsible for most annual pollutant washoff. Infrequently occurring
large storms usually produce greatest flooding damage, and although they
may contain significant pollutant loads, their contribution to the annual
average pollutant load is quite small (Chang et al. 1990). Also, because
of the infrequent occurrence of large storms, receiving waters have relatively
long periods of recovery between events (Claytor and Schueler 1996). Therefore,
urban forests are likely to produce more benefits through water quality
protection than through flood control. Research is needed to better understand
the interception process for open-grown urban trees, as well as the impacts
of canopy interception on water quality.
Aston, A. R. 1979. Rainfall interception by eight small trees. J. Hydrol. 42:383-396.
Baldwin, J.l. 1938. Interception of snowfall by forests. New Hampshire For. & Rec. Dept. Forest Notes 6 (mimeo).
Brutsaert, W. 1988. Evaporation into the Atmosphere— Theory, History, and Applications. D. Reidel Publishing Company, Holland. 299 pp.
Calder, I.R. 1996. Rainfall interception and drop size— development and calibration of the two-layer stochastic interception model. Tree Physiol. 16:727-732.
Calder, l.R. 1977. A model of transpiration and interception loss from a spruce forest in Plynlimon, Central Wales. J. Hydrol. 33:247-265.
Chang, G., J. Parrish, and C. Souer. 1990. The first flush of runoff and its effect on control structure design. Environmental Resource Management Division. Department of Environmental and Conservation Services. City of Austin, Austin, TX. 36 pp.
City/County of Sacramento. 1996. Sacramento City/County Drainage Manual, Vol. 2, Hydrology Standards, Sacramento County Public Works Agency, Department of District Engineering, Water Resources Division, Sacramento, CA.
Claytor, R.A., and T.R. Schueler.1996. Design of Stormwater Filtering Systems. The Center for Watershed Protection, Silver Spring, MD.
Dong, A., S.R. Grattan, J.J. Carroll, and C.R.K. Prashar.1992. Estimation of daytime net radiation over well-watered grass. J. Irrig. Drain. Eng. 118(3):466-479.
Edward, H.l., and R.M. Srivastava.1992. Applied Geostatistics. Oxford University Press, New York. 561 pp.
Environmental Protection Agency. 1994. The quality of our nation's water: 1992. United States Environmental Protect Agency, #EPA-841-S-94-002. USEPA Office of Water, Washington, DC.
Gash, J.H.C., C.R. Lloyd, and G. Lachaud.1995. Estimating sparse forest rainfall interception with an analytical model. J. Hydrol. 170:79-86.
Gash, J. H.C.1979. An analytical model of rainfall interception by forests. Quart. J. R. Met. Soc. 105:43-55.
Gash, J.H.C., I.R. Wright, and C.R. Lloyd.1980. Comparative estimates of interception loss from three coniferous forests in Great Britain. J. Hydrol. 48:89-105.
Gash, J.H.C., and A.J. Morton. 1978. An application of the Rutter model to the estimation of the interception loss from Thetford forest. J. Hydrol. 38:49-58.
Hamilton, E.L., and P.B. Rowe. 1949. Rainfall interception by chaparral in California. State of California, Dept. of Natural Resources, Division of Forestry. 43 pp.
Horton, R.E.1919. Rainfall interception. Mon. Weather Rev. 47:603-623.
Hungerford, R.D. , R. R. Nemani, S.W. Running, and J.C. Coughlan. 1989. MTCLIM: A mountain microclimate simulation model, Research paper INT-414. USDA For. Ser. Intermoun. Res. Sta., Ogden, UT.
Jetten, V.G. 1996. Interception of tropical rain forest performance of a canopy water balance model. Hydrol. Process. 10(5):671-685.
Lloyd, C.R., J.H.C. Gash, and W.J. Shuttleworth.1988. The measurement and modeling of rainfall interception by Amazonian rain forest. Agric. For. Meteorol. 43:277-294.
Lormand, J.R. 1988. The effects of urban vegetation on stormwater runoff in an arid environment. Master's thesis, School of Renewable National Resources, Univ. Ariz., Tucson, AZ.100 pp.
Massman, W.J.1983. The derivation and validation of a new model for the interception of rainfall by forests. Agric. Meteorol. 28:261-286.
McPherson, E.G. 1998. Structure and sustainability of Sacramento's urban forest. J. Arboric. 24(4):174-190.
McPherson, E.G. 1984. Energy-Conserving Site Design. American Society of Landscape Architects, Washington, DC. 326 pp.
Monteith, J.L. 1973. Principles of Environmental Physics. American Elsevier Publ. Co., New York, NY. 241 pp.
Penman, K.L. 1948. Natural evaporation from open water, bare soil, and grass. Proc. Roy. Soc. London A198: 116-140.
Pitt, R.1994. Small storm hydrology. University of Alabama, Birmingham. Unpublished manuscript. Presented at Design of Stormwater Quality Management Practices, Madison, Wl. May 17-19.
Pruitt, W.O., and J. Doorenbos. 1977a. Background and development of methods to predict reference crop evapotranspiration (ETO). Appendix 11, 2nd edition, Crop Water Requirements, Irrigation and Drainage Paper No. 24. FAO, Rome, Italy, pp. 108-119.
Pruitt, W.O., and J. Doorenbos.1977b. Empirical calibration, a requisite for evapotranspiration formulae based on daily or longer mean climate data. International Round Table Conference on "Evapotranspiration." Budapest, Hungary. Intl. Commission on Irrigation and Drainage. 20 pp.
Roland, B. Stull. 1988. An introduction to boundary layer meteorology. Kluwer Academic Publishers, Dordrecht, The Netherlands. 666 pp.
Rutter, A.J., A.J. Morton, and P. C.. Robins. 1975. A predictive model of rainfall interception in forests, 11. Generalization of the model and comparison with observations in some coniferous and hardwood stands. J. Appl. Ecol. 12:367-380.
Rutter, A.J., and A.J. Morton. 1977. A predictive model of rainfall interception in forests, 111. Sensitivity of the model to stand parameters and meteorological variables. J. Appl. Ecol. 14:567-588.
Rutter, A.J., K.A. Kershaw, P.C. Robins, and A.J. Morton. 1971. A predictive model of rainfall interception in forests, 1. Derivation of the model from observations in a plantation of corsican pine. Agric. Meteorol. 9:367-384.
Sanders, R A. 1986. Urban vegetation impacts on the hydrology of Dayton, Ohio. Urban Ecol. 9:361-376.
Soil Conservation Service. 1975. Technical Release No. 55. Urban hydrology for small watersheds. Soil Conservation Service, USDA, Washington, DC.
Ustin, S.L., W.W. Wallender, L. Costick, R. Lobato, S N. Martens, J. Pinzon, and Q.F. Xiao. 1996. Modeling terrestrial and aquatic ecosystem responses into Hydrologic Regime in a California watershed. In SNEP Science Team and Special Consultants (Eds.). Sierra Nevada Ecosystem Project final report to Congress: Status of the Sierra Nevada, Volume III: 275-307. Centers for Water and Wildland Resources, Univ. Calif., Davis, CA.
Vertessy, R.A.,R.G . Benyon , S. K. O'Sullivan, and P. R.Gribben. 1995. Relationships between stem diameter, et al. 1971): sapwood area, leaf area and transpiration in a young mountain ash forest. Tree Physiol. 15:559-567.
Xiao, QF., S.L. Ustin, W.W. Wallender. 1996. A spatial and temporal continuous surface subsurface hydrological model. J. Geophys. Res., D, Atmos. 101 (D23):29565-29584
Zinke, P.J. 1967. Forest interception studies in the United States, pp 137-161. In Sopper, W.E., and Lull, H.W. (Eds.). Forest Hydrology. Pergamon Press, Oxford, England.
Precipitation water balance on a canopy surface can be expressed as:
|
(A1)
|
Differentiating equation (A1) with time gives the general canopy interception
equation:
|
|
(A2)
|
Interception (I) is the sum of canopy surface water storage (C) and evaporation (E). Interception loss (Li) accounts for all of the water evaporated from canopy leaf and branch surface (E).
Canopy drip rate is described as an exponential function of canopy storage
and saturation storage capacity (Rutter et al. 1971; Lloyd et al. 1988;
Jetten 1996):
|
|
 C³
S |
(A3)
|
To calculate drainage from stem surfaces (stem flow), we assume that water available on stem surfaces for drainage is supplied mainly by the proportion of the gross precipitation (psp) and lost by both flow and evaporation. Evaporation from stem surface storage is small compared with evaporation from leaf surfaces. Rutter and Morton (1977) estimated it as 1% to 5% of the canopy evaporation value. Stem flow is calculated as directly proportional to precipitation (qstem = psp) Free throughfall is calculated as a fraction of gross precipitation (th = pfp), where pf is the canopy shading coefficient.
Canopy evaporation is described as (Rutter et al. 1971):
|
|
 C<S |
(A4)
|
 |
(A5)
|
 |
(A6)
|
 |
(A7)
|
 |
(A8)
|
In equation (A5), we use drying power of the air instead of aerodynamic resistance to calculate potential evaporation because the wind function (equation A8) is well studied in the study area (Pruitt et al. 1977a,1977b). Simulation accuracy should increase due to the way evaporation is estimated.
Net radiation is calculated from solar radiation (Monteith 1973; Roland 1988; Dong et al.1992). The wind profile at the meteorological station was retrieved from the wind speed measured at stand height (2 m [6.6 ft.] from ground surface) (Brutsaert 1988; Jetten 1996). We are not extrapolating air temperature and relative humidity from measurement height to actual canopy height because the vertical gradient is small.
Boundary and initial conditions must be determined before we can start solving these equations. Two flux boundaries are defined: upper boundary (at the canopy top) is determined by precipitation and evaporation rates, and lower boundary (at ground surface) is determined by canopy drainage (throughfall) and stem flow rates. To determine initial conditions, we assume that the canopy surface is dry before initiation of the precipitation event.
The model (equation A1) is explicitly solved using finite differences.
Numerical instability errors are reduced by limiting the maximum time step.
Assuming air temperature and relative humidity measured from meteorological
stations are representative of the canopy surface, these data can be used
directly without modification.
Acknowledgements. We thank Drs. Bruce Ferguson (University of
Georgia) and Eric Larsen (UC Davis) for their comments on an earlier version
of this manuscript. We appreciate Andrew Hertz for his assistance preparing
canopy leaf and crown projection area calculations, Klaus Scott for help
accessing and processing the meteorological data, James Goodridge (Division
of Local Assistance, California Department of Water Resources) for providing
Stonemead Station meteorological data, and Simon Eching (CIMIS program
Development and Outreach, California Department of Water Resources) for
his assistance using CIMIS data.