源頭集水區表層土壤平均含水率推估之研究

Abstract

森林源頭集水區表層土壤水的分布及動態同時受到地形、植生分布及土壤性質等環境因子所影響而具有強烈的空間變異，此測量尺度下，除了以高時空解析度監測表層土壤含水率之時空變動外，若能建立以少量樣點推估森林源頭集水區表層土壤平均含水率，將有助於瞭解大氣及土壤界面之水文反應及交換作用。本研究以森林源頭集水區內表層土壤含水率的高時空解析度測量資料為基礎，嘗試減少樣本數並在最具代表性的位置取樣以估計源頭集水區的平均含水率。本研究利用手持式時域反射儀(TDR)對459個樣點0-12 cm的土壤含水率做了詳細的測量，一年內共進行了12次的調查。透過相對差異法分析各樣點的時間穩定性並篩選出代表性樣點。以不同數量樣本所得之平均含水率與樣區平均含水率之殘差平方和(RSS)及線性迴歸的r2值，並與野外調查中常用的系統取樣法比較，估計多少樣本數才足以代表源頭集水區平均含水率。確定所需的樣本數以後，利用環境因子及坡地區位為線索，分析代表性樣點的分布特性。本研究結果顯示以時間穩定指數(ITS)最小的代表性樣點進行估計時所需要的樣本數為8，系統取樣所需的樣本數則大於15。代表性樣點的分布位置與土壤、植生、地形等環境因子並未表現出顯著的相關性，而與坡地區位較為相關。全年的分析結果顯示，代表性樣點多分布於沖蝕溝及谷源中段西側，其中沖蝕溝的含水率具有強烈的空間變異，不具代表性與具有代表性的樣點分布區位重疊，因此該地不適合作為樣區平均含水率之採樣地點。邊坡的含水率雖具有高時間穩定性但總是相對乾燥，於此區域測得的數值會低於樣區平均含水率，而不具有代表性。代表性樣點的分布隨著樣區乾濕狀態的不同而變動，樣區較乾燥時，代表性樣點的分布較無明顯趨勢。當樣區較濕潤，地表、地中逕流等側向水分移動較旺盛時，其分布則向樣區中央集中。

Surface moisture (θ) in a forested headwater catchment would vary greatly in space and time due to the influences of topography, vegetation, and soil properties. In addition to the monitoring network with highly spatial and temporal resolutions, developing an effective method to measure the average surface moisture with few samples would be efficient to understand the hydrological responses and interactions at the atmosphere-soil interface in a forested headwater catchment. In this study, we tried to estimate average surface moisture with a small sample size and to find the locations with representativeness. We measured θ at the depths of 0-12 cm at 459 measurement points using portable TDR (time domain reflectometry) sensors. Twelve surveys were conducted within 1 year in a natural forested headwater catchment. Temporal stability of each measurement point was analyzed by the relative difference approach. The mean soil moistures determined by temporal stability and systematic sampling methods were compared to average surface moisture for each survey. The minimum sample size necessary to estimate average surface moisture was determined with the residual sum of squares (RSS) and r2 value of linear regression between mean soil moistures and average surface moisture. We also analyzed the relationships of the representative samples (Sr) and environmental factors (i.e. topography, vegetation, and soil properties) in space and time. The results showed that 8 samples with minimum index of temporal stability (ITS) were sufficient to estimate average surface moisture, while more than 15 systematic samples were necessary. The correlations between the spatial distributions of Sr and environmental factors were weak. Most of Sr located at the valley-head hillslope and gullies. Gullies were not necessarily good locations for Sr where the best and worst representative samples existed. The spatial distribution of Sr varied with the dry and wet conditions. When average surface moisture was lower, the spatial pattern of Sr could not be characterized by environmental factors. When average surface moisture was higher, Sr tended to locate around the center of the headwater catchment.