| 540 | 17 | 157 |
| 下载次数 | 被引频次 | 阅读次数 |
土壤水分是"四水"转换的纽带,农业生产的基础,传统的野外定点测量土壤水分的方法难以实现空间上的展布,现代微波遥感数据可以得到大尺度的土壤水分,但分辨率低。本文利用CLDAS数据,将机器算法应用到遥感影像指数运算中,开展土壤水分的降尺度研究。论文分别采用OLS算法、Bagging算法、BRT算法和随机森林算法模型建立MODIS光学遥感数据(LST、Albedo、NDVI、ET)与土壤水分的关系模型。研究结果表明:四种算法中随机森林算法的拟合效果更优(R2=0.961 28,RMSE=0.006 99)。利用该算法算出降尺度后的土壤体积水分,可以得到大尺度且空间分辨率更高的土壤水分数据。大清河流域西北部土壤含水量高于东南部,土壤含水量差异可达0.2 mm3/mm3,在流域土壤含水量空间分布的季节变化显著,3月土壤水分低至0.16 mm3/mm3,9月土壤水分高达0.33 mm3/mm3。
Abstract：Soil moisture is the tie that water transformation and the function of agricultural production. The traditional method of field measuring soil moisture is difficult to achieve space distribution. At the same time, microwave remote sensing data can get large scale but low resolution. In this paper, soil moisture of CLDAS and machine learning algorithm is applied to the remote sensing image index operation. The study on soil moisture downscaling paper respectively using OLS algorithm, Bagging algorithm, BRT algorithm and random forest algorithm model is established between the MODIS data(LST, Albedo, NDVI and ET) and soil moisture. It is used to calculate the soil moisture by the random forest algorithm with better fitting effect(R2=0.961 28, RMSE=0.006 99), so as to obtain large-scale soil moisture data with higher spatial resolution. The soil moisture in the northwest of Daqing river basin is higher than that in the southeast, and the difference of soil water content can reach 0.2 mm3/mm3. The spatial distribution of soil moisture varied significantly with different seasons. The soil water content is as low as 0.16 mm3/mm3 in March and it is as high as 0.33 mm3/mm3 in September.
[1] 吴超.基于遥感影像对土壤水分反演的研究[D].西安:西安科技大学,2011.
[2] 裴浩,范一大,乌日娜.利用气象卫星遥感监测土壤含水量[J].干旱区资源与环境,1999(1):74- 77.
[3] 郭晓娟,周妍妍,郭建军,等.疏勒河流域土壤含水率反演[J].干旱区研究,2018(6):1317- 1326.
[4] CARLSON T.An Overview of the aTriangle Methoda for Estimating Surface Evapotranspiration and Soil Moisture from Satellite Imagery[J].Sensors,2007,7(8):1612- 1629.
[5] PETROPOULOS G,CARLSON T N,WOOSTER M J,et al.A review of Ts/VI remote sensing based methods for the retrieval of land surface energy fluxes and soil surface moisture[J].Progress in Physical Geography,2009,33(2):224- 250.
[6] MOOSAVI V,TALEBI A,MOKHTARI M H,et al.Estimation of spatially enhanced soil moisture combining remote sensing and artificial intelligence approaches[J].International Journal of Remote Sensing,2016,37(23):5605- 5631.
[7] PARK S,IM J,RHEE J,et al.AMSR2 soil moisture downscaling using multisensor products through machine learning approach[C]// Geoscience and Remote Sensing Symposium (IGARSS),2015 IEEE International.IEEE,2015.
[8] 刘文静.大清河流域可利用水资源量演变规律研究[D].邯郸:河北工程大学,2014.
[9] 赵立军.基于MODIS数据的北京地区土壤含水量遥感信息模型研究[D].北京:中国农业大学,2004.
[10] ZHAO W,LI A.A Downscaling Method for Improving the Spatial Resolution of AMSR-E Derived Soil Moisture Product Based on MSG-SEVIRI Data[J].Remote Sensing,2013,5(12):6790- 6811.
[11] 师春香,姜立鹏,朱智,等.基于CLDAS2.0驱动数据的中国区域土壤湿度模拟与评估[J].江苏农业科学,2018,46(4):231- 236.
[12] 庄媛.中国区域多源主被动微波遥感土壤湿度产品融合研究[D].南京:南京信息工程大学,2014.
[13] Breiman,Leo.Bagging predictors[J].Machine Learning,1996,24(2):123- 140.
[14] 沈学华,周志华,吴建鑫,等.Boosting和Bagging综述[J].计算机工程与应用,2000(12):31- 32.
[15] ELITH J,LEATHWICK J R,HASTIE T.A working guide to boosted regression trees[J].Journal of Animal Ecology,2008,77(4):802- 813.
[16] LI M,IM J,QUACKENBUSH L J,et al.Forest Biomass and Carbon Stock Quantification Using Airborne LiDAR Data:A Case Study Over Huntington Wildlife Forest in the Adirondack Park[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2014,7(7):3143- 3156.
[17] BREIMAN L.Random forests are a combination of tree predictors such that each tree depends on the values of a random vector sampled independently and with the same distribution for all trees in the forest[J].Machine Learning,2001,45(1):5- 32.
[18] CLEWLEY D,WHITCOMB J B,AKBAR R,et al.A Method for Upscaling In Situ Soil Moisture Measurements to Satellite Footprint Scale Using Random Forests[J].IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing,2017,10(6):2663- 2673.
基本信息:
DOI：10.13928/j.cnki.wrahe.2019.10.003
中图分类号:S152.71
引用信息:
[1]吴颖菊,朱奎,鲁帆,等.基于CLDAS数据和机器算法模型的大清河流域地表土壤湿度降尺度研究[J].水利水电技术,2019,50(10):18-24.DOI:10.13928/j.cnki.wrahe.2019.10.003.
基金信息:
国家重点研发计划课题(2018YFC0406501);; 国家自然科学基金项目(51679252)
2019-10-20
2019-10-20