【作者】 张贞；

【导师】 魏朝富； 谢德体；

【作者基本信息】 西南大学 ， 农业资源利用， 2009， 博士

【摘要】 针对土地退化、生态环境恶化等一系列问题,如何保证土地资源的农业生产能力,实现可持续利用已成为土地科学的研究重点。土地质量评价则是了解农业生态系统状态,监测农业行为对环境的影响,保障可持续土地利用最有效的方法。目前土地质量评价研究主要集中在指标体系和评价方法。国内普遍采用的指标体系都是源于土壤质量指标体系,在其基础上有一定衍生;考虑土地是一个“生态-经济-社会”的复合生态系统提出压力-状态-响应(PSR)框架。指标体系的构建存在指标选取不够全面、区域性强的问题,侧重选择生态、土壤指标,经济方面考虑较少。评价方法则以传统方法居多,和GIS的结合还相对较弱,评价主观性强,土地利用格局和生态过程变化对土地质量的影响研究力度不足。不同尺度上,由于影响土地质量的主导因素差异显著,评价指标和方法选择也应不同。但是目前大部分研究仍停留在单一尺度上,较少考虑尺度之间的联系以及不同尺度不同行为主体对土地质量的影响。本研究以三种尺度:合川典型田块(地块尺度)、高龙土地整理项目区(村级尺度)、合川区(县域尺度)为例,考虑不同尺度上影响土地质量的主要因素,采用不同的方法对丘陵区多尺度土地质量进行研究,剖析多尺度土地质量研究的异同,丰富和完善了土地质量评价体系,为实现丘陵区土地资源的可持续利用提供理论依据。1地块尺度的土地质量研究地块尺度选取涵盖整个合川区常有种植模式的40个典型样点进行研究,重点考虑地块自然质量、农户干扰结合地形地貌加以讨论。基础资料主要源于对典型地块和相邻地块农户调查的137户有效问卷以及各地块的养分分析数据。运用随机前沿生产函数从管理者角度将农户的生产投入水平转化为农业生产效率,分析从2000年-2007年的变化以及原因;运用生态服务价值计算方法详细量化各个地块的生态子功能;将农户效率、生态价值、各地块地形地貌特征等综合考虑作出最后的土地质量评价,最后针对影响地块尺度的驱动力进行详细讨论。(1)随机前沿生产函数较好的考虑无效性问题,能合理描述农户行为。参数r=0.9217＞0.5,在1%水平下通过显著性检验,农户在生产过程中受到各种随机因素的影响而造成对生产效率的损失。常数项β0,β1、β2在1%水平上显著,同时所有β系数都为正,其中劳动力投入对产出的弹性系数为0.3494,生产资料投入对产出弹性系数为0.5215。因此资本和劳动力的投入在本研究中对农户产出的作用都是正向的。对于无效率函数而言,解释变量(地块户主年龄、家庭劳动力人数、耕地面积、受教育程度、距离合川区的远近以及政策因素)的系数均未通过显著性检验,这6个解释变量是引起技术非效率的主要原因。其中农户年龄系数δ1系数为负值,农户年龄无效性对农户收入有显著的负向影响,能够提高技术效率。而家庭劳动力人数、耕地面积、受教育程度、距离合川区的远近以及政策因素的系数为正值,对技术无效性有正向的影响,这5个解释变量对农户技术效率的提高有一定的制约作用。(2)地块生产率的差异源于农户行为,但地块的生态功能对其没有太大的贡献作用。农业生产效率同生产产品功能、调节大气功能、涵养水源功能、保持土壤功能和养分循环功能的关系在0.05显著性水平下,5个子功能中都未能进入方程,各单位地块总价值与农业生产率二者之间没有显著线性相关性。不同种植模式下单位面积上生态服务价值菜园＞麦玉苕＞果园＞水旱轮作＞冬水田,其中又以生产功能和养分循环带来的价值量最大。(3)40个地块以中、优等质量耕地为主,总体质量状况良好。选择海拔、坡度、坡向等10个指标进行地块土地质量的综合评价。土地质量是地形、土壤、生态以及农户因素共同作用的结果,任何一个因子也无法替代其他因子来表达对土地质量的贡献作用。2村尺度的土地质量研究土地整理对村尺度土地质量有显著的影响。以重庆市合川区高龙土地整理项目区为例,分析土地整理驱动下的土地质量变化,主要考虑景观格局、农户行为、生态(生态功能、土壤质量等自然属性)三个方面。基础数据主要源于整理区1:2000的土地利用现状图与土地利用规划图,项目所涉及三个行政村的163份有效问卷及土壤相关基础数据。(1)整理后土地利用图斑趋于规整,形状复杂度降低,优势度增强,景观的异质程度降低,景观格局趋于简化。采用Fragstats3.3进行整理前后的景观格局对比分析发现,土地整理后图斑合并引起了土地分布特征的变化,整体形状复杂度降低,图斑形状趋近于圆形和方形。土地平整工程的实施改变了农田景观原有格局,有利于农业生产,单位斑块的面积增大,斑块边界的复杂弯曲程度降低,便于规模化种植。渠系35m的缓冲区距离为一个人类干扰的临界点,而田间道缓冲区距离在25m-50m之间斑块最容易受到人为干扰。(2)整理前后技术进步是推动全要素生产率变化的主要因素,土地整理项目的实施对三个村的影响总体趋势还是正向发展。运用数据包络分析(DEA)将调查资料量化以表达农户对土地整理的响应,三个行政村技术效率、技术进步、纯技术效率、规模效率和全要素生产率在整理前后的变化规律不一致,但是技术进步是全要素生产率变化的主要驱动力。技术效率以及由技术效率分解的纯技术效率和规模效率以吉福村最好,可以通过扩大生产规模使资源配置更加合理,而其余两村则需要缩小生产规模才能减少同最优规模之间的差距。进一步分析2004～2007年农户农业收入同农户投资结构之间的变化规律表明,三个村的农业收入还是以养殖业为主,牲畜的投入直接影响农户收益。有些年份农业收入除了受牲畜投入的影响,还受农药、化肥等种植业投入的影响,但都是单个种植业投入进入方程,因此技术进步是制约全要素生产率变化的主要因素。(3)整理后项目区具有较弱的生态压力和较少的生态服务价值的损失,这些变化都是在生态承载力范围内所允许的。运用生态服务价值方法详细分析研究区在整理前后的生态功能变化。整理后所有单个服务系统的经济价值较整理前都有不同程度的增加,其中增加最大的是养分循环功能,变化最小的则是保持土壤功能。通过道路、土地平整工程直接给生态系统增加了123.56万元,因此土地整理工程保证整个生态系统朝良性方向发展。选用成本效益率(BCR)评估土地整理项目的可行性,项目实施前六年BCR保持不变,从第7年到35年处于高速增长阶段,后面年份增加缓慢并最终趋于稳定。BCR＞1说明土地整理后研究区具有较弱的生态压力和较少的生态服务价值的损失。(4)系统动力学将土地质量系统(包括土地利用子系统、生态子系统、经济子系统、社会子系统)有机统一在一起进行评价。在详细分折构成土地质量系统内部循环关系的基础上,运用系统动力学构建整个土地质量系统流图,并引入“土地质量的货币化表示”预测到2020年的土地质量变化。在现行模式下,到2020年土地质量的货币化表示为79548300元。假定剩余田土坎全部转为耕地,未利用地也全部开发转为耕地,整理后新增耕地率都保持在3%不变预测未来土地质量变化趋势,结果2012年成为极限状态下农户收入、土地质量货币化表示、耕地生态系统服务功能和生态系统总服务价值的拐点,这些变量都是在2012年达到最大,随后逐渐降低。2012年出现的土地质量货币化最大值也是各种情景模拟有可能出现的最大土地经济价值极限值10957万元。3县域尺度的土地质量研究县域尺度主要以合川区为例,从耕地质量、空间格局两大方面讨论土地质量问题。主要运用农用地分等定级的基础数据,并对相关农户进行调查,获得234份有效问卷作为最原始的基础资料。(1)引入“影响力系数”修正耕地自然等级指数,能够真实反映土地生产能力的质量等级。农户对耕地投入产出水平的增加使得耕地自然质量呈现下降的趋势,仅考虑耕地自然质量不能全面反映耕地现实生产能力。用人为驱动力的干扰作用修正自然质量系数,新获取的修正质量系数具有了经济学意义,能真正反映土地的现实生产力。采用曲线分析算法,建立各个研究乡镇的水田和旱地收益状况与自然质量等级指数之间的相关性表明,旱地质量指数与农户收益的关系用三次项模型拟合最优的是盐井镇、钱塘镇和双槐镇,用二项式模型拟合最优的是隆兴镇和南办处,清平镇用对数方程模型最佳,肖家镇则采用倒数方程模拟最好,云门镇用指数模型进行拟合效果最好。水田质量指数同农户收益的关系用三项式拟合最优的是隆兴镇和盐井镇,而南办处和清平镇采用幂模型模拟最优,肖家镇和云门镇采用指数模型模拟,钱塘镇和双槐镇采用S模型模拟最优。(2)运用GeoDA软件进行空间自相关分析表达耕地质量在空间上的聚集状态。水田和旱地的修正质量等级分布存在非常显著的空间正相关关系,空间分布集聚现象显著。水田斑块中有538个斑块属于“HH”型,有409个属于“LL”型,它们同相邻斑块之间都具有较高的空间正相关性;有93个斑块属于“LH”型,296个斑块属于“HL”型,它们同相邻斑块之间存在较高的空间负相关关系。研究区有273个“HH”型和661个“LL”型旱地斑块,说明该类相邻的旱地斑块具有较高的空间正相关关系;有236个“LH”型和149个“HL”型旱地斑块,该类斑块具有较高的空间负相关关系。(3)1993-2006年度合川土地格局变化以林地增加最大,其次是住宅用地。草地减少最大;耕地是研究区的优势土地利用类型,也以1.11%的年减少率变化。景观格局上2006年的斑块密度比1993年有所增加,最大斑块指数有所降低,分维度、面积加权平均斑块分维数都较1993年有所减少,但变化幅度不是很大;Shannon多样性指数和Shannon均匀度指数都有所增加。应用Costanza提出的相应生态系统类型生态价值系数进行量化。1993～2006年间林地和园地的生态服务价值增加显著,但是由于耕地的占用、草地退化以及水域面积的缩小使得它们的生态功能减少。在整个研究区以水域和水利设施用地的灵敏度变化较大,从1993年的0.7180变化为2006年的0.6407。4多尺度土地质量评价体系的构建本部分从评价单元、指标体系和评价方法三个方面对比分析点、村和县域尺度土地质量研究的差异,在此基础上提出基于Multi-agent土地质量研究体系框架。(1)景观格局都是以栅格作为研究对象最后归纳为景观类型,栅格大小根据研究区面积判断;农户研究都是以单个地块(斑块)为单元,只是调查农户分布依据不同;研究自然质量是以土地利用现状图的耕地斑块为准,地块尺度的单个地块也可以看作闭合的斑块;生态功能研究单元是单一土地利用类型,如果能够获取单个斑块的属性数据以斑块为研究单元。指标都可以归为土地的生态质量指标和生产质量指标,生态质量强调土地属性和生态属性,生产质量强调土地生产能力、承载力和经济能力。研究方法上地块和县域尺度农户行为采用效率模型量化,而县域尺度基于耕地自然质量系数,引入农户参与的“影响力系数”加以修正;生态功能在三尺度上均采用生态服务价值功能量化。景观格局的研究均采用景观格局指数表示,而反映土地利用格局则采用空间自相关分析。(2)土地质量评价体系框架主要分为指标agent、方法agent,其中将指标agent归纳为土地属性agent、景观生态agent和农户agent,所有层次的研究都以这两个子agent为基础进行指标选择。方法agent实质是指标agent之间相互协作关系的实现,勾绘了土地质量交互性的评价体系框图。基于Multi-agent的评价框架可以与GIS技术、数学统计方法结合起来,通过构建agent数据库,完成尺度的推衍,逐渐实现土地质量的智能化评价,指导政府决策。综上所述,本文系统研究南方丘陵区不同尺度的土地质量状况,不再局限于某一单一尺度或者采用某一数学方法得出质量等级的传统研究思维,构建了适合多尺度研究的土地质量研究体系;同时对目前大力开展的农村基础设施建设对土地质量带来的影响进行系统研究,弥补国内在土地整理中土地质量系统性研究的不足。研究方法上采用随机生产函数、数据包络分析方法(DEA)将反映农户行为的繁琐数据归纳为农户生产效率,从管理者的角度研究农户与土地之间的关系。将系统动力学的理论运用在土地质量的研究中,详细分析了土地系统的内部结构,动态模拟土地系统的变化。但是由于某些尺度的研究只能获取到一个时间点的土壤数据,无法反映土地质量的变化过程。同时需要采用量化的方法将三个尺度的研究耦合在一起,因此未来应继续加大时间段上的研究和尺度耦合的研究,以便应用到实践中。更多还原

【Abstract】 It had become a key issue for land science how to maintain the agricultural production capacity of land resource because of land degradation,eco-environment deterioration and so on.By monitoring response of land use change to agro-ecosystem and environment,land quality assessment has recently become an effective measure for maintaining sustainable land use.As far as land quality assessment is concerned, indicator systems and evaluation methods have become its focus.At the present,Indicator systems which had been widely used in China were originated from soil quality indicator system.Pressure-state-response (PSR) assessment indicators system was established based on the theory of ecology-economy-society complex ecosystem.There have still been many problems in construction of indicator system,for example indexes were not chosen completely and impressed in some regions.Moreover,the assessment indicators in Pressure-state-response(PSR) assessment indicators system have given more emphasis to characteristics of ecology and soil and little economy.Traditional land quality assessment methods were little combined with GIS,so that their evaluations were sometimes arbitrary.It has been paid little attention to study the effect of land use pattern and ecological procedure on land quality.Most studies emphasized single scale and ignore of the relationship among scales.So the role of land monitoring and policy establishing were not reflected in land quality assessment.Therefore,this paper dealt with multi-scale land quality assessment in hilly-land regions under consideration of major influencing factors with various scale land use pattern,taking representative blocks of Hechuan,Gaolong land consolidation project area and Hechuan city for example.This study would enrich and improve land quality assessment system by anatomizing similarities and differences of multi-scale land quality in hilly-land regions.1.Land quality at land patch scaleForty plots chosen in Hechuan were covered with cropping patterns in whole city,and field parcels’ natural quality,households’ disturbance and topography landforms should be mainly considered.The basic data were obtained from peasant household investigation to typical and adjacent blocks,getting 137 valid questionnaires,and analytic dates of soil.The change of agricultural production efficiency from 2000 to 2007 was analyzed by using stochastic frontier production function.The ecosystem services value was used to quantify sub eco-function in each patch.Farmers’ efficiency,ecological value and geomorphic characteristics were integrated to evaluate,and further discuss the driving force for this scale.(1)Stochastic frontier production function was suit for describing household behavior,taking into account agricultural unavailability problems.Under a significance level of 0.01 andγ=0.9217＞0.5, results showed that farmers were influenced by many random factors,causing production efficiency loss in actual production process.The constant term ofβ1,β1 andβ2 in this study on 0.01 level were obvious,and all theβcoefficient were positive.The elastic coefficient of labor input was 0.3494,and productive investment was 0.5215.So it suggested that these inputs can improve agricultural output.The coefficient of explanatory variables such as farmers’ age,number of working force,cultivated area,education level, policy factor and the distance between plot and Hechuan city center had no significant difference.And it meant that there explanatory variables caused technical inefficiency.The coefficient of farmers’ age was negative,and it showed that the age unavailability had negative influence on household income.That is to say,it can enhance technical efficiency.But the coefficients of other explanatory variables were positive values which had a positive correlation with technology ineffectiveness,and they restricted technical efficiency.(2)The differences among production in each patch were derived from farmers’ behavior,and ecological function had not contributed to it.By analyzing the relationship among agricultural production efficiency,food production,gas regulation,water supply,soil formation and nutrient cycling ecosystem services,it was found these sub ecosystem functions can’t enter the regression equation (y=-E-05x+0.874,R~2=0.0822) under a significance level of 0.05.In other words,there was not relationship between ecological service value and agricultural production efficiency.The ecological service values in the unit area under various cropping patterns were garden＞wheat maize sweet potato rotation cropping system＞orchard＞paddy-upland rotation＞winter paddy field.Food production and nutrient cycling ecological functions had the most important service values.(3)Selecting such as elevation,slope and aspect ten indicators to evaluate land quality,these patches were dominated by medium and finest grade farmland,and the overall land quality was better.By analyzing the relation among indices,it suggested that land quality was the results of common action among terrain,soil,ecology and farmers,and those can’t be replaced each other.2.Land quality at village scaleLand consolidation projects had significant effect on land quality at village scale.This paper detailedly analyzed the changes of land quality from three respects:landscape pattern,household behavior and ecological functions,which were driven by land consolidation taking Gaolong land consolidation project for example.The fundamental datum were land use map,land use planning map(map scale is 1:2000) and 163 valid questionnaires.(1)Land consolidation,as a type of human disturbance,can improves land production efficiency and changes landscape distribution through their land parcels reallocation.FRAGSTAT3.3 and buffer zone analysis were used to reflect those changes.The results indicated that the shape of land use patch become more regular and agglomeration of little-patches contributes to changes in the distribution of land cover after LCP in total area at the landscape level,but the changes of dry land and paddy land were much more intense than other land use patterns at class level.Land leveling project decreased the slope of dry land and made them flatten to convenient for agricultural production and reduce soil erosion.Ditch system and road network were representative ecological corridors,so it was important to study those corridor effects using buffer zone analysis.Those systems strengthen infrastructure construction,but increased fragmentation. LCP was not just the reallocation of fragmented parcels,but also improved agricultural conditions. Meanwhile,35m buffer zone of ditch system was critical point of human interference,and the patched from 25m to 50m buffer zone of road network were changed most markedly with obvious.(2)This paper analysis technology efficiency,scale efficiency and total factor productivity trend of farmers’ behavior and willingness response to land consolidation using Malmquist-DEA and C~2R-DEA model,respectively.The results indicated that technical progress was the main factor to constrain the development.The implementation of LCP can promote the development of studying villages,but its performance demonstration was later at Jifu village.Technical progress was the main factor to constrain the development,and technologies innovation was still lower.Technical efficiency,pure technical efficiency and scale efficiency,of Jifu village,were better than others,which suggested it can expand production scale to make reasonable resources allocation.But other villages needed to reduce scale.The relationship between land consolidation and farmers’ material capital investment can be further analyzed through the change rule between farmers’ income and investment structure from 2004 to 2007.The agricultural incomes were mainly from breeding industry,and affected mainly by livestock input.In some year,the agricultural income was also affected by seed input,chemical fertilizer input and other farming input.This result further indicated that technical progress was the major factor to affect TFP.(3)Shadow price method,reflection engineering method and reforestation cost method were used to quantitatively reflect the significance of special ecological functions under land consolidation impacts.In addition,engineering design and shadow price method were integrated to quantify the value of disturbance regulation,and it is a new method for eco-benefit evaluation in other land consolidation projects.And meanwhile,Benefit and cost analysis was used to reveal its’ economic significance to forecast the ecological variation of research area in the future.The disturbance function was the centralized reflection of irrigation and drainage engineering,roading engineering and land levelling project.The ecosystem service values before and after land consolidation were 5995.98 million yuan and 8330.11 million yuan respectively.All sub-functions had increased in a certain extent after this project.Nutrient cycling function increased most and soil conservation function increased least.Land leveling project increased 123.56 million yuan.The benefit-cost rate maintained the fast growth from the seventh year to thirty-fifth year,and it tended to be stable after the thirty-sixth year.It illustrated there were weak ecological pressure and less losses of ecosystem service value,and land consolidation project made eco-environment positively developed and improved.(4)System dynamics based on feedback control theory and computer simulating techniques was an effective methodology for modeling and analyzing complex social,economic,and ecological systems. Land use sub-system,ecology sub-system,economy sub-system and society sub-system were integrated into land quality system,and system dynamics were used to analyze these relationships among sub systems. The flow map of land quality system was established based on internal circular relation,and the monetization representation of land quality was introduced to predict the change of land quality from 2004 to 2007.The monetization representation of land quality was 79548300 yuan under current model. Supposing the residual field ridge changed into cropland,unutilized lands were exploited into cropland,and new-increased cultivated land rate was 3%all the time.Household income,the monetization representation of land quality,cultivated land’s ecological service value and total value reached maximum in 2012, subsequently decreased.This monetization representation of land quality was also the maximum value in various scenarios.3.Land quality at county scaleLand quality in county scale was discussed from farmland quality and its spatial distribution pattern.Data from 234 valid questionnaires in whole county were analyzed based on agricultural land classification and gradation.(1)Influence coefficient was used to modify natural quality index based on the investigation to households.It turned out that the farmers’ input-output made cultivated land natural quality reduced.Only considering natural quality,it can’t reflect farmlands’ actual production capacity.Therefore,this influence coefficient made adjusted quality indices economization.To simulate the relationships between natural quality index and household incomes,curve equations were established.It suggested that trinomial equation were fitted to reflect relationships under dry land among Yanjing town,Qiantang town and Shuanghuai town,and binomial equation were suitable for Longxing town and Nanjin subdistrict office. But the better equation of Qingping town,under dry land were logarithmic,inverse and exponential model, respectively.Under paddy land,trinomial equations were fitted to Longxing town and Yanjing town,and power model were suitable for Nanjin subdistrict office and Qingping town.Xiaojia town,and Yunmen town used exponential model,and Qiantang town and Shuanghuai town used S model.Household incomes are not only depended on natural attributes of cultivated land,but also affect by input and management.In addition,household incomes with investment behaviors are estimated Cobb-Douglas production function. Household incomes are restricted by one of investment factors,so incorrect investment structure can be adjusted from elasticity coefficient to improve cultivated land quality.(2)Spatial autocorrelation analysis was used to reflect the spatial agglomeration of farmland by GeoDA.The distribution of adjusted quality grade under paddy land and dry land had remarkable positive correlation in space,and concentration phenomenon was obvious.There were 538 HH type paddy land patches,and 409 were LL type.That meant there were positive correlation between HH and LL type patch. 93 LH type and 296 HL type of paddy land had negative correlation in space.273 HH type and 661 LL type of dry land had positive correlation in space,and 236 LH type and 149 HL type of dry land had negative correlation in space.(3)Land use maps from 1993 to 2006 were used to analyze the spatial pattern of land use.The results showed that over a long period of time,the forestland and rural residential area increased day by day.But, grasslands and croplands decreased quickly.PD was slightly increased,while PAFRAC and FRACAM were decreased.SHDI and SHEI were increased.The factor of policy and the drives of economic interests were the main reasons that drive the landscape pattern change.Croplands’ leading position became weak, but the dominant position of forestlands was gradually presented due to the conversion of cropland to forest policy.Ecological service value coefficient assigning different categories of land use raised by Costanza was generally adopted.So we used this coefficient to quantify the change of ecological service value at county scale.The values of forestland and orchard were increased,while the values of other land use patterns were reduced owing to area decreasing from 1993 to 2006.Sensitivity analysis was used to reflect the dependence of our estimates of changes in ecosystem service values on the applied coefficients.The results showed that the changes of river’s sensitivity were highest from 0.7180 to 0.6407.4.Construction of land quality assessment system for multi-scaleA comparative analysis was conducted on three scales of patch,village and county scale,in view of evaluation units,indicator system and evaluation methods.Meanwhile,this paper presented the concept and theoretical specification of a multi-agent based on model for land quality assessment.(1)Landscape analysis was based on grid which was in accordance with the total area investigated. Independent block was the unit of farmer research,but the selected foundation of investigated household was different.Cropland plague of land use map was natural quality study unit,while independence land use pattern was the unit of ecological service function.All indices can be divided into ecological quality and production quality.Ecological quality emphasize on land attribute and ecology attribute,but production quality reflected production capacity,carrying capacity and economic ability.Efficiency model was used to analyze household behavior at patch and county scale,but influence coefficient was adopted to reflect household’s action.Ecology service value was an effective method to describe ecology structure. Landscape metrics was adopted at county and village scale,while spatial autocorrelation analysis was used to reflect the change of land use pattern at county scale.(2)A multi-agent based model for land quality assessment was represented by two components: indicator agent and method agent.Among them,indicator agent can be summarized as follow:land attribute agent,landscape ecology agent and household agent.The essential dimension of method agent was the implementation of cooperative relationship among indicators,and the evaluation interactive diagram of land quality was drawn.This model can be combined with GIS,mathematical method through agent database to achieve yardstick reasoning and intelligentized assessment.As what have mentioned above,this paper systematical study land quality at multi-scale in hilly reginons,south China,which wasn’t limited to certain scale or traditional research method.Therefore,this study can enrich and perfect land quality research system.It can make up the research on land quality systemic for land consolidation project by analyzing the affection of agricultural infrastructure to land quality.In view of managers,household study was expression in production efficiency using stochastic frontier production function and data envelopment analysis.System dynamic was used to present the internal structure of land system to predict the change of future land.But only soil datum in a given time was got at certain scale,so it can’t reflect the change process limited by datum.Based on that,the integration of multi-scale’s yardstick reasoning will be the focus of future study and strengthen practical application.更多还原