|摘要: ||位於中亞的鹹海是廣為人知的高鹽分內陸流域，1960年時其面積約68,000平方公里，水體體積約1,093立方公里，自1960年開始，由於作為灌溉用水，使鹹海面積開始縮減，由於湖水補注不足，造成在1987年時鹹海縮減且分裂成南、北兩個湖（南方的湖面較大，故本文後以大、小湖區分）根據調查於2011年時整個鹹海已較早期縮小了83% 的湖面積以及93% 的水體體積，同時對其鄰近環境造成嚴重的衝擊。|
研究中，利用整個鹹海流域於1960年到2000年之降水、逕流以及蒸發量資料，其計算水體積與鹽度，由Nash-Sutcliffe model 預期水體積有效係數為0.975及鹽度為0.974，利用回歸方程式計算2010年到2021年之降雨、逕流、蒸發量，再進一步推估出2021年時，鹽度為200 g/l以及水體體積將減少83 km3。
研究中使用皮爾森相關係數以及複線性回歸，分別計算兩湖之水體體積、降水、逕流、蒸散量、水位高，並使用Holt''s Linear Trend Forecasts Method預估2010年到2025年兩湖之水體體積與鹽度。
在小湖的計算結果中，皮爾森相關係數於水體面積與水體體積為 (r = 0.942)、蒸發量與Syr Darya 逕流為(r = 0.722)，兩者皆為高顯著相關；相關性於水面高與水體體積為(r = 0.547, p = 0.005)、Syr Darya 逕流與降水為(r = 0.523, p = 0.031)兩者皆為顯著相關；相關性於水體體積與蒸發量為(r = -0.444)、蒸發量與水體面積(r = -0.449, p = 0.070)，兩者皆為負相關。
本研究中複線性回歸回歸結果顯示，R=0.972、R2 =0.94、Adjusted R2 = 0.918，小湖的變異數分析為高顯著性(F = 36.989, p–value was 0.000)根據統計推論，小湖於2025年時，水體體積將增加22.8 km3、鹽度提升至接近190 g/l。
在大湖的計算結果中，皮爾森相關係數於水體體積與降水為 (r =0.709)、蒸發量與水體體積為(r = 0.743)、水體面積與水體體積為(r = 0.962)、水面高與水體體積為(r = 0.784)、Syr Darya 逕流與降水為(r = 0.845)皆為高顯著性；
本研究中複線性回歸回歸結果顯示，R =0.984、R2 = 0.969 、Adjusted R2 = 0.951，大湖的變異數分析為高顯著性(F = 55.798, p= 0.000)根據統計推論，大湖於2025年時，水體體積將減少2.4 km3、鹽度提升至接近200.7g/l。
The Aral Sea is a landlocked and well-known saline endorheic basin in Central Asia. It was covered an area of 68,000 km2 and had a volume of 1,093 km3 in 1960. From 1960 onwards the lake started to shrink because of rivers water utilized for irrigation purposes. Due to irrigation, water was used enormously activities, without sufficient water supply for its watershed, the lake has shrunk and separated into two parts: the Small Aral Sea in the north and the Large Aral Sea on the south at 1987. According to the recent survey, the entire Aral Sea has lost 87% of water surface area, 93% of water volume in 2011, and has an environmental disaster areas its surround.
The Aral Sea (entire) observed data of precipitation, runoff, and evaporation in the watershed were used to estimate water volume and salinity in the lake from 1960 to 2009; the Nash-Sutcliffe model efficiency coefficient for predictive water volume and salinity were 0.975 and 0.974 respectively. Regression equations were used to estimate corresponding magnitudes of precipitation, runoff, evaporation, and salinity from 2010 to 2021, and applied in the estimations of the water volume and salinity. Our estimates suggest that salinity will increase to around 200 g/l and that water volume will decrease to around 83 km3 in 2021.
We have computed the water volume, runoff, precipitation, evaporation, water surface area, and water level of the Small Aral Sea and Large Aral Sea by using the Pearson correlation coefficient and a multiple linear regression. Holt''s Linear Trend Forecasts Method was used to predict the water volume and salinity of the Small Aral Sea and the Large Aral Sea from 2010 to 2025.
In the Small Aral Sea, the Pearson correlation between the water surface area and water volume (r = 0.942), the Evaporation and Syr Darya runoff (r = 0.722) were highly significant. The relationship between the water level and water volume (r = 0.547, p = 0.005), the Syr Darya runoff and precipitation (r = 0.523, p = 0.031) were significant. The correlation among the water volume and evaporation (r = -0.444), the evaporation and water surface area (r = -0.449, p = 0.070) were silently negative as our resulted. In this study, multiple linear regression analysis demonstrated that the R stands for the relation between the dependent and independent variables were 0.972, R2 values indicates that 94%, and the Adjusted R2 = 0.918. Our ANOVA results of the Small Aral Sea water volume was highly significant, F = 36.989, p–value was 0.000. The prediction results were indicated that the water volume of the Small Aral Sea will be increased to around 22.8 km3, and salinity of the Small Aral Sea will rise to approximately 190 g/l by 2025.
The Large Aral Sea Pearson correlation between the water volume and precipitation (r = 0.709), the water volume and evaporation (r = 0.743), the water volume and water surface (r = 0.962), the water volume and water level (r = 0.784), the Amu Darya runoff and precipitation (r = 0.845) were high significantly. The multiple linear regression analysis showed that the R stands for the correlation between the water volume and independent variables were 0.984, R2 = 0.969, and the Adjusted R-square = 0.951. ANOVA results shows that the contrast was highly significant, F = 55.798, p= 0.000. Our prediction suggests that the Large Aral Sea water volume will decreases to around 2.4 km3, and salinity will be raised to around 200.7 g/l in 2025.