CAJ | 학술논문

为科学地引导西南山地丘陵区永久性基本农田的划定，该文以重庆市垫江县为例，通过层次分析法、熵权系数法、TOPSIS法（technique for order preference by similarity to an ideal solution）和约束因子建立决策模型体系，结果表明：耕地指标中灌溉保证率、平均降雨量、交通和农村居民点辐射的自身优势突出。土壤质量各指标平均值（介于0.2～0.75之间）、标准差（介于0.09～0.3之间）、变异系数（介于13～75之间）为中等以下水平。自然条件中，灌溉保证率和耕地连片性的综合权重最大，大于10℃积温和平均降雨量的综合权重最小、且指标值的区间变化最大；土壤质量中，有机碳在层次分析法下权重最大；距离条件中，入选耕地受城镇的辐射作用明显、区位优势突出。耕地在入选重要程度上，依次排序前4位的是灌溉保证率、耕地连片性、城镇的辐射和土壤有机碳，测度和贴近度显示，地块间优劣接近，距离条件整体较优。耕地的入选从贴近度等级划分为：重点型、适宜型、调控型、后备型，约束条件下调出的地块为优势耕地（即重点型和适宜型）。最终划定成果显示适宜型和调控型比例增大，重点型和后备型比例减小。并利用数据验证了该决策模型体系的合理性、规范化和可行性。
위과학지인도서남산지구릉구영구성기본농전적화정，해문이중경시점강현위례，통과층차분석법、적권계수법、TOPSIS법（technique for order preference by similarity to an ideal solution）화약속인자건립결책모형체계，결과표명：경지지표중관개보증솔、평균강우량、교통화농촌거민점복사적자신우세돌출。토양질량각지표평균치（개우0.2～0.75지간）、표준차（개우0.09～0.3지간）、변이계수（개우13～75지간）위중등이하수평。자연조건중，관개보증솔화경지련편성적종합권중최대，대우10℃적온화평균강우량적종합권중최소、차지표치적구간변화최대；토양질량중，유궤탄재층차분석법하권중최대；거리조건중，입선경지수성진적복사작용명현、구위우세돌출。경지재입선중요정도상，의차배서전4위적시관개보증솔、경지련편성、성진적복사화토양유궤탄，측도화첩근도현시，지괴간우렬접근，거리조건정체교우。경지적입선종첩근도등급화분위：중점형、괄의형、조공형、후비형，약속조건하조출적지괴위우세경지（즉중점형화괄의형）。최종화정성과현시괄의형화조공형비례증대，중점형화후비형비례감소。병이용수거험증료해결책모형체계적합이성、규범화화가행성。
In order to guide the delimiting permanent basic farmland in the mountain and hill areas, this paper sets up a multiple-objective decision model by using Chongqing Dianjiang as an example. The model is supported by data analysis using ArcGIS 10.1 and other software based on the 2012 Second Land Survey data and Land Use Overall Planning data of 25 towns, basic farmland planning data, soil data, and investigation of county and country Land and Resource departments using spatial distribution products of more than 10℃ accumulated temperature in the cloud platforms of Geography National Condition monitoring. The model is also supported by the AHP (Analytic Hierarchy Process), Entropy Coefficient Method, TOPSIS (Technique for Order Preference by Similarity to an Ideal Solution), and constraint fact. First, the subjective weight of each index is given based on the AHP method by analyzing and erecting the judgment matrix. The objective weight is obtained by using the Entropy Coefficient Method by building the judgment matrix to a normalized process. The value of measurement and closeness of positive and negative ideal points is reflected through erecting Euclidean distance in the normalized matrix. The higher was the value of measurement and closeness, the higher was the priority of land. The call-out qualifications of farmland are confirmed by the constraint fact method in 4 investigations, which supplies the conditions of the Construction land or general farmland. The paper concludes that in the standardized normal outcome of the farmland decision index, the optimum conditions are farmland irrigation, rainfall, transportation, and rural residential radiation. For most soil quality indexes, the range of equal value, STD, and CV are [0.2, 0.6], [0.1, 0.3], and [13, 75]. Overall, the value of CV is below medium while existing comparable large areas have a high value or a low value. Variation in subjective and objective weights of these indexes, after normalization, causes the maximum weight to be farmland irrigation and farmland connectivity. The minimum weight is more than 10℃accumulated temperature, average rainfall in natural conditions, organic carbon in soil quality by analytic hierarchy process, transportation, rural residential radiation, and towns radiation in distance condition. The most important conditions for selecting permanent farmland are farmland irrigation > farmland connectivity > towns radiation > soil organic carbon. According to the value of measurement and closeness of positive and negative ideal points, the quality of farmland is generally close to each other, and the distance condition is at it’s best and comparatively stable. Therefore, whole performance is better. Four grades are divided as key, proper, adjusting, and reserved, in which farmland is chosen as permanent basic farmlands using the natural breaks method to reclassify the closeness degree. There are special differences among different grades in statistics and spatial patterns. Local departments prefer to select superior farmlands, which are in the‘key’ and‘proper’ grades, in the selection under constraint conditions. In the ultimately delimited results, the proportion of proper and adjusting grades increases, while the proportion of key and reserved grades decreases. The rationality, standardization, and practicability of the model are verified by comparison among ultimately delimited results, high weight index in the model, and a basic farmland protecting area plan. The multiple-objective decision method is not only suitable for delimitation of permanent basic farmland in the mountain and hill areas in southwest China, but also ensures red line and economic growth, which supplies the theoretical basis and the decision reference for the study of complementary delimitation in the mountain and hill areas.