基于层次分析法和GIS的秦岭地区魔芋潜在分布研究

Potential distribution of Amorphophallus rivieri in the Qinling Mountains based on analytic hierarchy process and geographic information system

魔芋(Amorphophallusrivieri)为我国传统的食用与药用植物,主产于秦岭以南地区。依据陕西秦岭地区28个魔芋分布点信息,利用秦岭地区45个气象台站1961—2010年气候资料、土壤数据(分辨率1 km)和DEM高程数据(分辨率30 m),结合前人研究,通过魔芋产量与环境指标的相关性分析,获取相关性显著的20个评价指标,包括气候指标13个、土壤指标4个和地形指标3个,运用GIS技术和多元回归模型对气候指标进行栅格化处理,基于层次分析法和加权平均法获得评价指标权重,建立陕西秦岭地区魔芋潜在种植分布模型,确定魔芋潜在种植的空间分布。结果显示:陕西秦岭地区魔芋最适宜种植区面积1 214.42 km2,占可种植区面积的10.18%;适宜种植区面积2 015.60 km2,占可种植区面积的16.90%;次适宜种植区面积3 115.03 km2,占可种植区面积的26.12%;不适宜种植区面积5 580.02 km2占可种植区面积的46.80%。适宜魔芋潜在种植区域主要分布在陕西汉中中南部、安康中南部以及商洛东南部。

Amorphophallus rivieri （the corpse flower） is a traditional edible and medicinal plant in China. This species is distributed in the south part of the Qinling Mountains, China. We assimilated data about A. rivieri cultivation, environmental information from 28 sampling sites in the Qinling Mountains, climate data from 45 weather stations in the Qinling Mountains from 1961 to 20i0, soil data with 1 km x 1 km spatial resolution and DEM data with 30 m × 30 m spatial resolution in the Qinling Mountains, A. rivieri data collected throughout China, and a specific report on A. rivieri in Shaanxi Province. We obtained 20 assessment factors that were significantly correlated when evaluating A. rivieri yield against environmental factors. The key environmental factors affecting the distribution of A. rivieri cultivation included 13 dominant climate factors, 4 dominant soil factors, and 3 dominant topographical factors. These dominant factors were 1 ） Frost-free duration （D）, 2） Annual average temperature （Tn）, 3） Annual total active temperature （ ≥ 10℃）（ T≥ 10djw ）, 4） Monthly mean maximum temperature from July to August （ T78zg ）, 5 ） Annual precipitation （ Pn）, 6） Monthly mean daily temperature range from July to September （ T79gc）, 7） Monthly mean temperature from May to October （ T510p ）, 8 ）Monthly mean temperature from July to August （T78p ）, 9） Monthly mean relative air humidity from July to August （Q78）, 10） June precipitation （P6）, 11 ） July precipitation （ P7 ）, 12） August precipitation （ P8 ）, 13 ） September precipitation （P9） , 14） Topsoil depth （H）, 15） Topsoil pH（H2O） （pH） , 16） topsoil texture classification （C） , 17） Topsoil organic matter （O）, 18） Aspect （A）, 19） Slope （S）, 20） Altitude elevation （h）. Using Geographic Information System （GIS） and a multivariate regression model, the climate factors were rasterized. Then, we used fuzzy mathematics analysis, analytic hierarchy process （AHP）, and the weighted means method to set up the subjection function and determine the weight of each factor. We set up a model of ecological suitability for A. rivieri in the Qinling Mountains of Shaanxi Province, and determined the spatial distribution of suitable planting areas for this species. The root-mean-square error （RMSE） was used to evaluate the accuracy of the model predictions. The RMSE value reached 7.8%, which indicated that the predictive accuracy of the model was ＂Excellent.＂ The results showed that the ecological planting suitability model identified a relationship between potential A. rivieri cultivation distribution and the environmental factors. Highly suitable, moderately suitable, marginally suitable, and unsuitable planting areas of 1 214.42 kmz, 2 015.60 km^2, and 3 115.03 km^2, and 5 580.02km^2 were identified for A. rivieri, respectively. The potential suitable planting areas for A. rivieri were mainly distributed in the south central part of Hanzhong district, the south central part of Ankang district, and the southeast part of Shanluo district. This information on the potential suitable planting area of A. rivieri is valuable for providing baseline data, scientific information, and a research platform for understanding the ecology, geography, and environmental science of this important medicinal species.