| In the context of the continuous reduction of cultivated land quantity,protecting and improving the cultivated land quality(CLQ)is an important measure to increase grain production and ensure national food security.Carrying out CLQ evaluation research can effectively grasp the current situation of CLQ,and analyzing the spatiotemporal variation characteristics of CLQ can predict the evolution trend of CLQ.The research results are helpful to optimize the allocation of cultivated land resources and promote the sustainable development of cultivated land.However,the traditional CLQ evaluation method has low efficiency and poor timeliness,which cannot meet the needs of practical application.With its advantages of wide coverage and high spatiotemporal resolution,remote sensing technology has promoted the development of CLQ evaluation in the direction of large area,long time series and high time efficiency.This study took the cultivated land in Lateritic soil region of Guangdong province as the research area.Firstly,based on multi-source remote sensing data,CLQ monitoring data,soil attribute data and meteorological data,the key indicators of CLQ: soil organic matter(SOM),soil p H,irrigation guarantee capacity and productivity capacity were estimated.Then,a comprehensive evaluation indicator system of CLQ was constructed by combining the above-mentioned key indicators with soil texture,soil thickness,slope,utilization intensity and concentrated contiguity.Finally,the CLQ comprehensive index method was used to evaluate the CLQ in the study area in 2010 and 2018,and its spatiotemporal characteristics were revealed.The main research contents and conclusions of this paper are as follows:(1)Research on the estimation of SOM in cultivated land based on multi-source data.First,based on MOD09 A1,climate,topography and soil attribute data,34 environmental variables were selected as predictors of SOM.Then,a decision tree,adaptive enhanced decision tree,random forest and support vector regression model were constructed to evaluate the importance of the selected predictors.Finally,the model with the highest accuracy was selected to estimate the spatiotemporal distribution of SOM of cultivated land in the study area in 2010 and 2018.Results showed that the first derivative of green band reflectance,average annual temperature,soil thickness and clay content are the key factors for estimating SOM.The average SOM of cultivated land in the lateritic soil region of Guangdong province in 2010 was25.16g/kg,and the average SOM of cultivated land in 2018 was 21.93g/kg.The overall spatial distribution of SOM was "high in the north and low in the south".Compared with 2010,the SOM content of cultivated land decreased significantly in 2018.(2)Research on the estimation of soil p H in cultivated land based on multi-source data.The predictors of soil p H value were selected based on remote sensing,climate,topography and soil attribute data.The prediction performance of decision tree,adaptive boosted decision tree,random forest and support vector regression model for estimating soil p H was compared.Random Forest model was used to estimate the spatiotemporal distribution of soil p H of cultivated land in the study area in 2010 and 2018.Results showed that the annual EVI＿Max,mean annual temperature and cation exchange capacity(CEC)were the key factors to estimate the soil p H of cultivated land.In 2010,the average soil p H of cultivated land was 5.45,and in 2018,it was 5.42.The soil of cultivated land was mainly strongly acidic(4.5<p H≤5.5)and acidity(5.5<p H≤6.5),which were distributed across the whole study area.From 2010 to 2018,soil p H value of cultivated land in the study area showed a downward trend.(3)Research on remote sensing evaluation of irrigation guarantee capacity of cultivated land.Based on remote sensing evapotranspiration data and meteorological data,the effective irrigation amount and irrigation water demand were calculated by using crop coefficient method and Penman formula,and the ratio of the two was used as the evaluation indicator of irrigation guarantee capacity of cultivated land.Based on this,this study evaluated the irrigation guarantee capacity of cultivated land in the lateritic soil region of Guangdong Province in 2010 and 2018.Results showed that in 2010,the proportion of irrigation guarantee capacity grade of “fully satisfied”,“satisfied”,“basically satisfied” and “not satisfied” was35.6%,23.4%,20.7% and 20.3%,respectively.In 2018,the proportion of four grades was 43.2%,22.9%,19.4% and 14.4%,respectively.The cultivated land in the western and eastern parts of the study area had strong irrigation guarantee capacity,while the cultivated land in the central part had relatively poor irrigation guarantee capacity.Compared with 2010,irrigation guaranteed capacity of cultivated land in the northern part of the study area improved significantly in 2018.(4)Research on remote sensing evaluation of cultivated land productivity.Based on temporal vegetation index and MOD17 A3 NPP data,the effectiveness of different vegetation indices and NPP in evaluating cultivated land productivity was compared.The mean of annual enhanced vegetation index(EVI＿Mean)was used to quantify the high-productivity capacity and stable-productivity capacity of study area in 2010 and 2018.Results showed that the high-productivity capacity of cultivated land in 2010 and 2018 was mainly "high" and "generally high".The stable-productivity capacity of cultivated land was mainly "stable" and "generally stable".The high-productivity capacity of cultivated land in the eastern and western parts of the study area was better than that in the central and northern parts.Compared with2010,The high-productivity capacity of cultivated land was improved in 2018,and the stable-productivity capacity of cultivated land did not change significantly.(5)Research on comprehensive evaluation of CLQ.Based on the research results of the first four chapters,a comprehensive indicator system for evaluating CLQ was constructed in combination with soil texture,soil thickness,slope,utilization intensity and concentrated contiguity,and the comprehensive evaluation indicator system was used to evaluate the CLQ in the study area in 2010 and 2018.Results showed that in 2010,excellent-,medium-and poor-quality cultivated land accounted for 21.4%,40.9%and 37.7%,respectively.In 2018,they accounted for 26.1%,40.0% and 33.8%,respectively.The cultivated land with excellent-quality was mainly distributed in the north and west of the study area,and the cultivated land with poor quality was distributed in the middile and south of the study area.Compared with 2010,the CLQ in the study area was improved in 2018. |
