Temporal And Spatial Patterns Of Citrus Orchard And Its Features Of Soil Temperature And Humidity In Hilly Red Soil Region Of South Jiangxi Province

The red soil hilly area of south Jiangxi is the main producing area of citrus in China.Accurately obtaining citrus planting information is of great significance to the formulation and management of regional agricultural policies.However,the research on remote sensing extraction of orange orchard in red soil hilly area of south Jiangxi is relatively weak and needs to be strengthened.Therefore,this paper takes Nanfeng county,Jiangxi province as the research area,builds high spatial and temporal resolution data with Landsat and MODIS data based on enhanced Spatial and temporal adaptive fusion model?ESTARFM?,combines DEM data,and adopts the random forest classification method to extract the information of orange orchard in the research area from 2002 to2014 by remote sensing.Meanwhile,based on the data of soil water content,soil temperature and the meteorological data of the same period,this study discussed the change characteristics of soil temperature and humidity of orange orchard and natural forest and the correlation with the main meteorological factors,and compared the soil moisture content of orange orchard and natural forest after different intensity precipitation events to measure the water holding capacity.The main conclusions are as follows:?1?High space-time resolution data in land use/cover classification has very important significance,this study based on ESTARFM generated image compared with the original image significantly correlated?p 0.01?or less,can effectively make up because of the cloud of pollution caused by the lack of time series data,remote sensing and a more detailed characterization vegetation growth process of years,with more precipitation and cloud heavy pollution region of land use/cover classification has important practical significance;?2?Based on the high spatial and temporal resolution data constructed by the spatial temporal fusion model,the spatial distribution characteristics of the study area in 2002and 2014 were extracted by the random forest classification method,with the accuracy of 87%and 92%respectively,and the Kappa coefficient of 0.85 and 0.94.The results show that the random forest classification method based on high spatial and temporal resolution data has a good effect in the information extraction of orchard.Among them,surface reflectance data and DEM data play an important role in improving the classification accuracy of orange orchard.?3?According to the results of remote sensing extraction,the main land use types in the study area include forest,orange orchard,residential area,water body and cultivated land,and the forest and orange orchard are mainly distributed.The forest is mainly distributed in the higher mountain area.orange orchard is mainly distributed in central hilly area and along the river area.From the perspective of time change,from2002 to 2014,the area of orange garden in the research area increased significantly,and the increase area was about 13.64km^2.Woodland was significantly reduced,with a reduction area of about 36.76 km².According to the land use transfer matrix analysis,from 2002 to 2014,the natural forest was mainly converted into orange orchard and cultivated land,with the converted area of 157.52km² and 17.64km² respectively.The water body is transformed into the area of other land use type least.?4?Based on the analysis of soil temperature and humidity data and contemporaneous meteorological data,it was found that the annual change of soil temperature was consistent with the trend of monthly temperature change in orange orchard and natural forest,and the annual difference of soil temperature in natural forest was larger than that in orange orchard.Secondly,except in May and June,the soil temperature of orange orchard was higher than that of forest.The soil moisture of orange orchards and natural forests changed regularly during the year,with more in summer and less in winter and spring.However,there are some differences,as follows:the soil water content of orange garden is lower than that of natural forest,and the change range is smaller than that of natural forest.The correlation analysis showed that there was a significant correlation between soil temperature and air temperature in orange orchard and natural forest.However,there is a certain difference in humidity.The closer the orange orchard is to the surface,the stronger the correlation between soil humidity and precipitation is.The correlation between soil moisture and precipitation in natural forest was 20cm>30cm>10cm from large to small.At the depth of 20cm,there was a significant correlation between soil moisture and precipitation?p<0.01?.At the depth of 10cm,the correlation was the smallest.?5?Within 72 hours after a heavy rainfall event,the soil moisture content of orange orchard and natural forest reached the highest value at 20cm.Moreover,the water content of the natural forest was significantly higher than that of the orange orchard,and the water content of the natural forest continued to increase with the extension of the distance from the precipitation event,while the orange orchard gradually decreased.After a period of weak precipitation,the moisture of orange orchard and natural forest changed with depth,and the soil moisture of all layers of natural forest was higher than that of orange orchard.With the extension of the time from the precipitation event to the depth of more than 20cm,the soil moisture of orange orchard and natural forest decreased,and the decrease of natural forest was more significant.