附有等式约束病态模型正则化解的单位权中误差无偏估计

利用平差参数间合理的等式约束虽能提升病态模型解的精度,但其本质仍是通过引入正则化参数来改善模型的病态性,由于改变了观测方程的结构,所得的估值残差及单位权中误差均有偏。针对这一不足,在病态模型正则化解的无偏单位权方差估计式基础上引入等式约束条件,根据约束正则化解的残差二次型期望公式,导出约束正则化解的无偏单位权中误差估计式,并用数值算例和病态测边网算例验证其正确性。结果表明,本文公式所估的单位权中误差精度优于传统公式所估结果。     

SWAT模型对高精度土壤信息的敏感性研究

土壤信息是SWAT模型的重要输入数据,通常认为,土壤信息的精度直接影响着模拟结果的准确性。本文以美国Brewery Creek流域(19.5km²)为例,在其他输入不变的情况下,通过比较不同精度土壤数据(美国农业部SSURGO土壤图与SoLIM方法获得的土壤图)的模拟径流,分析SWAT模型对高精度土壤信息的敏感性。应用结果显示,在模型的校正前后,两种土壤数据的径流模拟结果均近似,差别并不显著。这表明在小流域水文模拟中,SWAT模型的径流模拟对高精度土壤信息的敏感性较弱,模拟径流不能很好的体现一定精度基础上土壤信息的差别。本文将此现象主要归因于:SWAT模型所采用的SCS-CN径流计算方法,在计算CN值(Curve Number)时将不同土壤类型综合到四个土壤水文组的做法,概括了土壤信息,模糊了土壤之间的属性差别,损失了土壤精度信息。本研究发现了SCS-CN径流计算方法在利用高精度土壤数据时存在的问题,并进行了分析,为水文模拟中参数的确定和数据的准备提供了参考。     

青藏高原土壤热交换的计算方案及初步分析

从地温满足的热传导方程出发,导出了计算土壤平均和瞬时热通量的计算方案。该计算方案可同时计算出土壤热通量随时间和随深度的变化。它利用整层的地温信息来计算任一层的热通量,这种方案克服了用差分方案进行直接计算的局限性。然后使用中日亚洲季风观测实验期间的地温自动观测站资料和相应的常规观测资料,计算了青藏高原上土壤热通量及其变化,结果表明,不论是常规观测站还是自动观测站,其结果与青藏高原第一次观测实验所用热流板的直接观测结果是相近的,因而这种计算方案是实用而有效的。     

基于GIS和USLE的东圳库区土壤侵蚀量预测研究

运用GIS、遥感技术与USLE模型相结合预测东圳库区流域土壤侵蚀量,利用GIS平台建立东圳库区基础地理数据库,以10×10m为栅格单元对库区进行离散化,在GIS空间分析功能支持下,结合USLE土壤侵蚀预测模型,根据合适的USLE因子算法生成栅格图层,通过图层运算,预测了东圳库区土壤侵蚀量,进而标识库区土壤侵蚀严重区域,为库区非点源污染控制及水土流失治理提供定量依据,并通过建立实用的泥沙输移比模型结合水文监测数据,对预测结果进行验证。根据USLE方程,采用改进后的SDR方程,预测的渡里流域年输沙量与流域实际监测得到的年输沙量仅相差14.08%,说明本文采用USLE方程模拟流域土壤侵蚀结果可信,可定量评价、拟定水土保持措施方案。     

加权总体最小二乘和RBF神经网络的三维坐标转换

分析部分变量误差加权总体最小二乘法（PWTLS）、加权总体最小二乘法（WTLS）和最小二乘法（LS）在三维坐标转换模型参数求解中的应用与影响,提出PWTLS与RBF神经网络组合的坐标转换方法。结果表明,当三维坐标转换模型系数矩阵中同时存在常数元素和重复元素时,PWTLS方法计算的单位权中误差和内符合精度均优于LS方法,且源坐标改正数较WTLS方法更加合理。PWTLS+RBF组合方法能够使PWTLS的求解参数得到有效使用,提高坐标转换精度。     

PSO-GM(1,1,N,p,ξ)模型在变形预测中的应用

通过对传统GM(1,1)缺陷分析和改进的基于权的PGM(1,1)建模机理描述,顾及PGM(1,1)中背景值构造时取相同的参数不能充分降低模型的预测误差,对不同的时刻引入不同的参数来改进GM(1,1)背景值序列的计算公式,将这种背景值构造方法和灰元N引入GM(1,1)建立了新的白化方程。在建立的新的白化方程基础上,用龙格-库塔法以修正的初始值计算累加值的模拟序列。针对引入的参数较多问题,采用粒子群算法寻找满足相对误差均值最优的参数,从而建立了基于粒子群优化算法和加权灰色组合的PSO-GM模型。工程实例应用表明,新模型的拟合精度高,预测效果好,相对其他两种原有模型预测精度有明显提高。     

GM(1,1)模型的改进及其在变形预测中的应用

为了提高GM(1,1)模型预测精度,采用积分优化、二次拟合优化以及残差改化方法,分步对GM(1,1)模型进行改进,建立灰色多重修正模型.具体改进步骤为:首先,利用积分优化方法对背景值进行纠正,减小模型误差并提高预测精度;接着,对模型参数(发展系数和灰作用量)进行二次拟合优化,使参数更加接近理论真值;然后,根据预测结果进行适当的残差改化,提高模型整体的预测精度;最后,建立根据GM(1,1)模型改进的灰色多重修正模型.以重庆南川地区甄子岩崩塌为例,建立灰色多重修正模型对危岩裂缝累计位移值进行模拟和预测,并与GM(1,1)模型进行对比.精度检验结果表明:灰色多重修正模型后验差比值(0.082 39)明显好于GM(1,1)模型(0.192 67),平均相对残差比(0.073 9)更远好于GM(1,1)模型(0.259 6),表明灰色多重修正模型在预测精度上有较大提高,可靠性更好.     

基于熵权-IOWGA算子的变权组合预测新方法

针对熵权法组合权重分配固定不变的问题,提出一种基于IOWGA算子的权重分配方法,建立基于熵权-IOWGA算子的变权组合模型。该方法以各单一模型的预测精度作为诱导值,按每一期诱导值的大小对单一模型进行有序赋权,并将其应用到变形数据分析中。实验表明,该方法赋权过程合理,模型预测精度更高,更加适用于变形分析预测过程。     

土壤有机质含量估测及其影响因素的光谱分析

 土壤颗粒大小差异使土壤反射光谱产生相应变化,影响土壤有机质含量等属性的光谱预测精度。本研究准备了颗粒粒径分别为2、0.25和0.15mm的土样,测定土壤有机质(Soil Organic Matter,SOM)含量,并于室内模拟条件下测定其反射光谱。通过分析不同粒径土样的原始(Raw)、多次散射校正(Multiple scattering correction, Msc)、一阶微分(First derivative, Fd)、连续统去除(Continuum removal, Cr)光谱与SOM含量之间的关系,筛选出与SOM含量相关性最强的Fd光谱单波段(2250nm, r=0.82, P<0.01),并建立线性回归模型;利用全波段光谱反射率,以偏最小二乘回归(Partial least square regression, PLSR)方法,确立2mm土样Msc处理光谱的PLSR模型为最优模型(RPD=3.56、R²=0.90、RMSEP=1.96g/kg)。土壤颗粒粒径对土壤光谱反射率变化有明显影响,但二者之间并非简单的线性关系,可能存在一个转折点;单变量(单波段光谱反射率)线性回归模型的预测能力,明显低于全波段反射光谱(Msc处理)-PLSR模型;土样样本容量对SOM含量预测精度有显著影响。因此,根据样本容量大小,选择合适的土壤颗粒粒径与光谱预处理方法组合可以提高预测精度。     

基于支持向量机的水资源预测模型

水资源信息的预测可以为水资源的合理调配、管理和规划提供依据,为提高拟合和预测精度,建立基于支持向量机的水资源信息预测模型,采用捕鱼算法对其参数优化,并应用于地下水水位预测的实例分析,且拟合及预测精度均较高,与神经网络模型方法所取得的结果进行比较,能取得更好效果,表明将回归支持向量机用于水资源信息的拟合、预测是可行的.     

Soil erosion calculation in the hydro-fluctuation belt by adding water erosivity factor in the USLE model

Soils in the hydro-fluctuation belts of the reservoirs are most highly influenced by the special hydro-conditions and reservoir operation, leading to unique soil erosion process and largely accelerate soil erosion intensity. The present study aimed to estimate soil erosion rate in the hydro-fluctuation belt of the Pubugou Reservoir, Southwest China, based on the framework of Universal Soil Loss Equation(USLE).An attempt has been made to modify the original USLE by including the reservoir water erosivity(W), a new factor into the model. Soil erosion rate from different land use types were quantitatively estimated,using the USLE and the modified USLE respectively.Field observation showed that soil erosion rate in dry farmland, bare land and grassland was 4700, 44600 and 5050 t/km2, respectively. The erosion rate assessed by the modified USLE was closely related to that recorded from the field monitoring data. The findings of this study clearly highlight the importance of inclusion of the W factor to the original USLE model while assessing soil erosion in the reservoir hydro-fluctuation belt.     

Study on the model for predicting soil erosion and its application in arid area

１　ＴＨＥＡＮＡＬＹＳＩＳＯＦＳＯＩＬＥＲＯＳＩＯＮＦＡＣＴＯＲＳＴｈｅｉｎｔｅｎｓｉｔｙｏｆｓｏｉｌｅｒｏｓｉｏｎｉｎａｒｉｄａｒｅａｖａｒｉｅｓｗｉｔｈａｌｔｉｔｕｄｅ,ｍｏｉｓｔｕｒｅａｎｄｈｅａｔｃｏｎｄｉｔｉｏｎａｎｄｓｕｒｆａｃｅｌａｙｅｒ（Ｃｈｅｎ,１９９５）．Ｉｎｍｉｄｄｌｅａｎｄｈｉｇｈｍｏｕｎｔａｉｎｓｅｃｔｉｏｎｓ,ｔｈｅｒｅｉｓａｌｉｔｔｌｅｍｏｒｅｐｒｅｃｉｐｉｔａｔｉｏｎ（４５０－７９６．９０ｍｍ）,ｒｉｃｈｅｒｖｅｇｅｔａｔｉｏｎｃｏｖｅｒｍｏｒｅｔｈａｎ６０ｐｅ…     

Estimation of non-point source pollution loads under uncertain information

Many kinds of uncertainties are involved, such as random, fuzzy, grey, unascertained property and so on, in soil erosion process. To exactly predict the non-point source pollution loads, some uncertainties should be taken into consideration. Aiming at the deficiency of present blind number theory being helpless for fuzziness, a novel blind number, i.e. extended-blind number, was introduced by substituting a set of triangular fuzzy numbers （TFNs）, expressed as a-cuts, for interval values in present blind number, and the expected value of extended-blind number was also brought forward by referring to the current blind number theory. On the basis of denoting the parameters of Uni- versal Soil Loss Equation （USLE） as extended-blind parameters, a novel USLE model was established for quantitatively evaluating soil erosion loss and non-point source pollution loads. As a case, the uncertain USLE was employed for predicting the soil erosion loss and non-point source pollution loads of absorbed nitrogen and phosphorus in a dis- trict in the Hangbu-Fengle River basin, in the upstream of Chaohu Lake watershed. The results show that it is feasible in theory to extend blind number into fuzzy environment and reliable on conclusion to apply extended-blind number theory for predicting non-point source pollution loads.     

Inference of reference conditions for nutrient concentrations of Chaohu Lake based on model extrapolation

In the mid-eastern China,there are few or no lakes which are in the absence of anthropogenic disturbances,or their sediments remain undisturbed.As a result,the reference lakes distribution and paleolimnological reconstruction approaches usually are inappropriate to estimate lake reference conditions for nutrients.This yields the necessity of using the extrapolation methods to estimate the lake reference conditions for nutrients within those regions.The lake reference conditions for nutrients could be inferred inversely from the law of mass conservation,current lake nutrient concentration,and the loadings from watershed.Considering the scarcity of hydrological and water quality data associated with lakes and watersheds in China,as well as the low requirement of the watershed nutrient loadings models for these data,the soil conservation service(SCS) distributed hydrological model and the universal soil loss equation(USLE) were applied.The SCS model simulates the runoff process of the watershed,thereby calculating dissolved nutrients annually.The USLE estimates the soil erosion and particulate nutrients annually in a watershed.Then,with the loadings from atmospheric deposition and point source,the previous annual average nutrient concentrations could be acquired given the current nutrient concentrations in a lake.Therefore,the nutrient reference conditions minimally impacted by human activities could be estimated.Based on the proposed model,the reference conditions for total nitrogen and total phosphorus of Chaohu Lake,Anhui Province,China are 0.031 mg/L and 0.640 mg/L,respectively.The proposed reference conditions estimation model is of clear physical concept,and less data required.Thus,the proposed approach can be used in other lakes with similar circumstances.     

Land use optimization for sustainable development of mountain regions of western Serbia

This study analyses the possibilities for optimal land use structure in mountainous areas of rural Serbia, which is faced with problems such as unfavorable terrain characteristics and inadequate land use patterns causing economic underdevelopment and increasing population migration. In the case of a small privately owned farm in Western Serbia, optimal land use pattern is determined using the linear programming model. The objective function-cost-effectiveness or soil erosion reduction, as well as constraints, are formulated using the benefit-cost ratio (BCR) for a 15-year production period and soil losses data calculated using the Universal Soil Loss Equation (USLE). The best results are achieved using the scenario with the aim to minimize soil erosion, considering the BCR in the constraints. This scenario is proposed since it was registered that the benefit for landowners would increase by five times, while the losses of land would decline about 80%. The sensitivity analysis of investments is performed to check whether the price variation affects the cost-effectiveness of proposed land use. Results show that market changes (nine cases are considered) do not lead to significant changes in land use allocation. Proposed land use pattern improves the environmental condition, strengthens the rural economy and thus offers the greatest potential for sustainable development of mountainous rural areas.     

Soil Erosion under Different Land Use Types and Zones of Jinsha River Basin in Yunnan Province, China

Severe soil erosion in the middle and upper reaches of Yangtze River has been regarded as a major environmental problem. The on-site impact of soil erosion on agricultural production and the off-site impact on floods and sedimentation in Yangtze Rive are well known. A quantitative assessment of soil erosion intensity is still scanty for developing appropriate soil erosion control measures for different land use types and zohes in this region. This article constructs a localized USLE and estimates the average soil loss in the Jinsha River Region in Yunnan Province, one of the priority areas for soil erosion control in the middle and upper reaches of Yangtze River. The estimation is done under different land uses and zones in this basin. The estimation shows that while soil erosion in the cultivated land is the most severe, 36～40% of the garden and forest land suffers from soil erosion of various degrees due to lack of ground cover and other factors. Soil erosion in the pasture is modest when the ground cover is well maintained. It also confirmed that terracing can reduce soil erosion intensity significantly on the cultivated land. Research findings suggest that sufficient attention must be paid to regeneration of the ground cover in reforestation programs. In addition to mass reforestation efforts, restoration of grassland and terracing of the cultivated land should also play an important role in erosion control.     

利用数字摄影测量方法估算半干旱区小流域沟谷侵蚀产沙

土壤侵蚀不仅表现黄土高原强烈的现代地貌过程,也是其生态环境恶化的象征,其地貌演化和土壤侵蚀互为因果,现代研究表明黄土高原半干旱小流域的沟谷侵蚀产沙占其水土流失总量的重要部分,近年有关沟谷侵蚀的定量研究成果,加深了人们对沟道侵蚀危害的认识,为土地资源的合理利用、沟谷危害的评价等提供了科学依据。传统的沟谷侵蚀定量研究是采用实地量测的方法,现代的数字摄影测量技术的发展为估算沟谷侵蚀研究提供了一个经济高效方法。本研究采用的髙分辨率数字高程模型是由3期历史航空照片根据数字摄影测量方法制作而成,利用它对黄土高原半干旱区一个小流域的1959～1981年及1981～1999年两个时段内沟谷侵蚀产沙量进行估算,将流域侵蚀产沙的实测值与估算结果进行评价。并指出此方法在沟谷侵蚀产沙量估算方面的统计学意义及在未来应用时要注意的问题;根据流域侵蚀产沙结果分析历史时期人类活动的影响:人类活动在一定程度上改善了环境,绝大部分泥沙被拦截在流域里,同时沟道侵蚀加剧,如何采取措施减少流域坡面和沟道的侵蚀,是目前人们所面对的问题;此外,结合流域水沙统计数据的DEM的侵蚀空间分析,结果表明坡面流、重力侵蚀及沟谷的下切是导致流域土壤侵蚀的主要原因。     

黄土高原丘陵沟壑区切沟侵蚀与地形关系分析——以纸坊沟流域为例

沟蚀是土壤侵蚀研究的主要内容之一,地形地貌是沟蚀的一个重要影响因子。本文以安塞纸坊沟流域作为研究区域,选择土地利用方式、土壤类型、坡度坡长因子、平面曲率、坡向和地形湿度指数6个因子,通过因子内切沟所占比重/整个研究区切沟所占比重计算各个因子的权重值,通过空间叠加分析土壤侵蚀敏感性,并通过重分类的方法把土壤侵蚀敏感性分为基本无侵蚀、轻度侵蚀、中度侵蚀、强度侵蚀、剧烈侵蚀5个等级,来研究切沟侵蚀与地形的关系。结果表明：切沟多发生在坡度坡长较大、地表湿度较高的林草地区域和更容易发生侵蚀的黄绵土区域,并且多分布在阴坡的凹面;对比分析切沟侵蚀和土壤侵蚀敏感性,切沟大多分布在中等侵蚀敏感性以上的区域,约占总切沟的90%;实验权重值对验证区冲沟的响应精度为82.43%（中度侵蚀及其以后阶段）,与实际值90.53%相差不大,说明此种方法对黄土丘陵沟壑区具有一定适用性,对黄土丘陵沟壑区水土保持工作有重要意义。     

六盘水市土壤侵蚀时空特征及影响因素分析

六盘水市是我国生态地位极其重要,水土流失又较为严重的城市。近些年,六盘水市实施了一系列生态工程,为了定量分析六盘水市土壤侵蚀状况及其影响因素,本文基于RUSLE模型,利用降雨数据、遥感影像数据、土地利用数据等,对贵州省六盘水市1990-2015年土壤侵蚀模数和土壤侵蚀量进行定量模拟,分析其时空动态变化,利用地理探测器定量分析影响因素,并进行空间相关性分析,结果表明： ① 六盘水市土壤侵蚀以微度和中度侵蚀为主。土壤侵蚀严重地区主要分布在北盘江流域与南盘江流域交界处以及喀斯特山地地区,煤矿开采使植被覆盖等抑制土壤侵蚀因子减少作用,使局部地区土壤侵蚀程度加剧。② 1990-2010年平均土壤侵蚀模数整体为下降趋势,2010-2015年为上升趋势。其中2000年平均土壤侵蚀模数最大,2010年平均土壤侵蚀模数最小。该变化由降雨可蚀性因子和植被覆盖度因子综合影响所致。③ 植被覆盖度因子和多年平均降雨量因子是影响区域土壤侵蚀空间分布的重要因素。未利用土地、植被覆盖度小于0.3、坡度在25°以上和降雨量在1543~1593 mm之间的地区为高风险侵蚀区域。④ 植被覆盖度与土壤侵蚀在空间上全部呈负相关性,降雨因子与土壤侵蚀在空间上存在负相关性和正相关性。⑤ 土壤侵蚀改善区域大多分布在生态工程区域内,生态工程建设能够改善土壤侵蚀情况,不同生态工程保护侧重点不同导致土壤侵蚀改善程度不同。退耕还林还草工程对六盘水市土壤侵蚀的改善具有重要意义,六盘水市更宜退耕还林。     

A Participatory GIS Approach to Identify Critical Land Degradation Areas and Prioritize Soil Conservation for Mountainous Olive Groves （Case Study）

Climate change scenarios predict an increase in the frequency of heavy rainfall events in some areas. This will increase runoff and soil erosion, and reduce agricultural productivity, particularly on vulnerable mountainous agricultural lands that is already exhibiting high rates of soil erosion. Haphazard implementation of soil and water conservation （SWC） interventions on scattered fields is inefficient in reducing soil erosion. The objective of this study was to identify areas at high risk of erosion to aid the design and implementation of sustainable SWC using GIS analysis and farmers＇ participation approach. A 25 m digital elevation model （DEM） was used to derive layers of flow accumulation, slope steepness and land curvature, which were used to derive an erosion-risk （priority） map for the whole watershed. Boundaries of farmers＇ fields were mapped and verified by the community and each field was classified into high, moderate or low erosion risk. Fields with low flow accumulation （top of hill） and/or steep slope and/or convex slope were assigned high erosion risk and therefore high implementation priority. The study showed that more than 54% of the fields were classified into high erosion risk areas. Accordingly, a community-watershed plan was established, revised and approved by the community.Incentive loans to implement SWC measures were distributed to 100 farmers based on the priorities of their fields. Judged by local farmers and using 16 randomly selected fields, 90% of the targeted areas were correctly identified using the erosion risk map. After two years, the conservation measures had led to marked improvement of soil conservation. The approach is simple and easy to comprehend by the community and provides scientific basis to prioritize the implementation of SWC and to target the most degraded areas, which amplify the impact of these in reducing the vulnerability to land degradation.