异质景观年平均蒸发量空间格局模拟

Simulating the spatial pattern of annual mean evapotranspiration of a heterogeneous landscape

理解景观格局如何影响生态过程是景观生态学的核心问题。建立了一个多尺度的空间显式景观过程模型 EPPML(Ecosystem Productivity Process- based Model at L andscape Scale) ,对中国东北长白山自然保护区生态系统碳 -水循环变量和生产力的时空格局进行了模拟 ,其中年平均蒸发量和蒸散量的空间格局是模型的主要输出结果之一。模拟值与实测值的数量级一致 ,这表明 EPPML 可以比较合理而准确地模拟该保护区主要生态系统的年平均蒸发量和蒸散量 ,但仍需进一步的模型验证和不确定性分析。年平均蒸发量的模拟值平均为 0 .198± 0 .0 93m / a,空间格局随海拔变化的趋势不明显 ,其中最高为云冷杉林(0 .2 76± 0 .0 81m/ a) ,最低为阔叶林 (0 .0 94± 0 .0 30 m / a)。环境因子 (气象因子和土壤含水量 )和植被因子 (植被类型和叶面积指数 L AI)等景观要素在空间格局上的变化会直接或间接地影响蒸发过程 ,进而调控整个景观的水平衡。环境因子对蒸发的影响比对蒸腾的影响要复杂得多。相对湿度对蒸发的影响最大 (R=0 .4 0 ) ,其它依次为气温、总辐射、降水量、风速和土壤含水量。年平均蒸发量与 L AI负相关 (R=- 0 .39) ,但两者并不呈简单的反比关系 :当 L AI较小时 ,蒸发量随 L AI的增加迅速地降低 ;当L

Understanding how landscape pattern affects ecosystem processes is a central issue in landscape ecology. In this study, we examined the effects of landscape heterogeneity on ecosystem evapotranspiration in the Changbai Mountain Nature Reserve in northeastern China, using a process-based, spatially explicit model, EPPML (Ecosystem Productivity Process-based Model at Landscape Scale). We developed EPPML to investigate the spatiotemporal patterns of carbon and water cycling and ecosystem productivity, and the spatial pattern of annual mean evaporation is one of the model outputs. Data on vegetation, soils, elevation, slope, and aspect were derived from digitalized maps. The spatial pattern of daily LAI at the landscape scale was estimated from NDVI (Normalized Difference Vegetation Index), which was derived from TM remote sensing imagery and field measurements and surveys. The daily meteorological variables were simulated based on related topographic and climatologic principles and daily meteorological point data in the Changbai Mountain Forest Station for the year of 1995. The daily soil water content was an important output of the daily water cycle, and the daily spatial patterns of all the variables of interest were stored in grid format in ArcInfo with a spatial resolution of 30m and Alberts projection. EPPML integrated these local-scale outputs to simulate the daily spatial patterns of carbon and water cycling variables, including those of evaporation and transpiration. Evaporation from canopy surface was mainly dependent on canopy interception, which was related to vegetation type, cover, canopy structure, LAI, and precipitation. Soil evaporation was simulated by Penman- Monteith equation. Thus, using the GIS-based process model, we scaled up the biophysical variables from the local site to the entire landscape using the scaling method, direct extrapolation. Finally we computed annual mean evaporation, total precipitation, mean air temperature, total solar radiation, mean relative moisture, mean wind speed, and mean LAI in each pixel, and analyzed how evaporation was related to environmental variables and LAI. The estimated mean evaporation for all the vegetation types in the reserve was 0.198m/a, ranging from 0.0 to 0.740m/a. The spatial pattern of evaporation did not seem correlated with elevation. Evaporation rates were ranked from high to low as follows: spruce-fir forests (0.276±0.081m/a), Changbai larch forests, alpine grasses, mixed broad-leaved and Korean pine forests, shrubs, alpine tundra, meadows, Betula ermanii forests, and broad-leaved forests (0.094±0.030m/a). The estimated mean evapotranspiration for all the vegetation types in the reserve was 0.836m/a, ranging from 0.0 to 1.188m/a. The spatial pattern of evapotranspiration obviously correlated with elevation. Among different vegetation types, the mixed broad-leaved and Korean pine forests had the highest evapotranspiration (1.057±0.173m/a). The spatial patterns of environmental factors and LAI directly or indirectly influenced evaporation processes, and further controlled the water balance at the landscape scale. The effects of spatial heterogeneity of environmental factors on evaporation were much more complicated than on transpiration. Evaporation was limited mostly by relative moisture (R = 0.40) and secondarily by air temperature, total radiation, precipitation, wind speed, and soil water content. Mean evaporation was negatively correlated with mean LAI (R=-0.39), but they were not related as a simple inverse ratio. When LAI was low, evaporation rapidly decreased with the increase of LAI; but when LAI further increased, evaporation varied little with LAI. Different vegetation types responded to LAI and environmental factors quite differently. Most of the simulated values fell within the same order of magnitude as field observations. This suggests that EPPML was able to simulate, with reasonable accuracy, the annual mean evaporation and evapotranspiration of various vegetation types in Changbai Mountain Nature Reserve. But further model va