
基于多源数据融合技术的绿洲灌区土壤水分反演
李华伟, 朱晓春, 张旭东, 隋喆, 周黎勇, 吴迪, 王叶, 白亮亮
基于多源数据融合技术的绿洲灌区土壤水分反演
Retrieval of Surface Soil Moisture at Field Scale in Oasis Irrigation Area Based on Multi-source Data
土壤水分是联系农业、生态和水文领域的重要环境变量,而卫星遥感是监测地表土壤水分的重要手段之一。针对微波遥感空间分辨率不足和光学遥感受云雨天气影响的问题,基于Landsat 8和MODIS光学影像、SMAP微波以及CLDAS再分析等多源数据,联合增强型自适应时空融合算法和随机森林模型对土壤水分进行定量反演,获得了绿洲灌区高时空分辨率田块尺度(30 m)土壤水分。结果表明:通过ESTARFM时空融合算法可有效获得日尺度30 m分辨率归一化植被指数(NDVI),融合后的NDVI与原始NDVI空间纹理特征一致,两者的相关系数(R)在0.85以上,均方根误差为0.05~0.08,融合效果较好。基于地表温度、NDVI、增强植被指数、叶面积指数、再分析土壤水分产品多特征参数组合下的随机森林模型反演效果最优,获得的高时空分辨率田块尺度土壤水分能够反映其时空变化,相关系数和均方根误差分别达到0.82和0.037 cm3/cm3。该方法可为灌区灌溉面积识别、旱情监测等提供技术支撑。
Soil moisture is an important environmental variable that connects the fields of agriculture, ecology, and hydrology. Remote sensing serves as a valuable tool for monitoring surface soil moisture. In response to the problems of insufficient spatial resolution in microwave remote sensing and the impact of cloud and rainy weather in optical remote sensing, field-scale (30 m) SSM was generated in this study based on random forest models combined with quality Landsat 8 and MODIS optical images, SMAP_L4 microwave data, and CLDAS reanalysis data, and in situ SSM measurements. The results showed that the ESTARFM algorithm could generate daily continuous normalized difference vegetation index (NDVI) image at a 30m resolution, exhibiting spatial consistency with the original NDVI image. The fused NDVI image achieved a correlation coefficient (R) of above 0.85 and a root mean square error ranging from 0.05 to 0.08, indicating a good fusion performance. The random forest model yielded the best results of soil moisture at high spatio-temporal resolution, based on the combination of multiple characteristic parameters of land surface temperature, NDVI, enhanced vegetation index, leaf area index and reanalysis of soil moisture products. The obtained soil moisture data at the field scale accurately captured its spatio-temporal variations, with a correlation coefficient of 0.82 and a root mean square error of 0.037 cm3/cm3. The research methods presented in this study can provide technical support for irrigation area identification and drought monitoring in irrigation districts.
土壤水分 / 遥感反演 / 时空融合算法 / 随机森林 / 数据同化 / 多源数据 {{custom_keyword}} /
surface soil moisture / remote sensing / spatial-temporal fusion / random forest / data assimilation / multi-source data {{custom_keyword}} /
表1 数据来源及时空分辨率Tab.1 Data source and spatiotemporal resolution |
数据类型 | 数据名称 | 含义 | 时空分辨率 |
---|---|---|---|
遥感数据 | Landsat-8 OLI C2L2 | 多光谱与热红外 | 30 m/8 d |
MOD11A1 | 地表温度 | 1 km/1 d | |
MOD13Q1 | 植被指数 | 250 m/16 d | |
MOD09GA | 地表反射率 | 500 m/1 d | |
MCD12Q1 | 土地覆盖类型 | 500 m/1 a | |
微波土壤水分产品 | SMAP L4_SM | 微波土壤水分 | 9 km/3 h |
中国气象驱动数据及产品 | GST | 地表温度 | 6.25 km/1 h |
PRE | 降雨 | 6.25 km/1 h | |
SSM | 0~10 cm土壤水分 | 6.25 km/1 h | |
土壤水分观测数据 | 体积含水率 | 0~10 cm土壤水分 | 站点/1 h |
图4 训练集与验证集土壤水分散点图Fig.4 Scatter plot of soil moisture for training and validation datasets |
表2 模型性能统计指标汇总Tab.2 Aggregated Statistical Indicators of Model Performance |
项目 | MAE/(cm3•cm-3) | RMSE/(cm3•cm-3) | R 2 |
---|---|---|---|
最大值 | 0.032 8 | 0.046 7 | 0.769 1 |
最小值 | 0.022 4 | 0.032 5 | 0.447 5 |
平均值 | 0.027 0 | 0.039 2 | 0.670 1 |
标准差 | 0.001 7 | 0.002 4 | 0.016 9 |
1 |
王丽丽, 王振龙, 索梅芹, 等. 基于神经网络的冬小麦生育期日土壤水分预测模型研究[J]. 节水灌溉, 2020(11) : 64-67.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
2 |
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, 35(6) : 1 317-1 326.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
3 |
郭文, 马梦梦, 孙培彦. 基于Sentinel-1A数据和BP神经网络的裸露地表土壤含水量反演研究[J]. 中国农村水利水电, 2023(1) : 89-94.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
4 |
高琪, 彭杰, 冯春晖, 等. 基于Landsat 8数据的荒漠土壤水分遥感反演[J]. 水土保持通报, 2021, 41(1) : 125-131, 151.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
5 |
厉芳婷,胡 菡,张 亮,等.基于土地利用分类的夜间灯光和GDP分量相关性研究[J].测绘地理信息,2023,48(1):82-85.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
6 |
艾璐, 孙淑怡, 李书光, 等. 光学与SAR遥感协同反演土壤水分研究进展[J]. 自然资源遥感, 2021, 33(4) : 10-18.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
7 |
邓小东, 王宏全. 土壤水分微波遥感反演算法及应用研究进展[J]. 浙江大学学报(农业与生命科学版), 2022, 48(3) : 289-302.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
8 |
姜红, 玉素甫江·如素力, 拜合提尼沙·阿不都克日木, 等. 基于支持向量机回归算法的土壤水分光学与微波遥感协同反演[J]. 地理与地理信息科学, 2017, 33(6): 30-36.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
9 |
李艳,张成才,恒卫冬,等.基于多源遥感数据反演土壤墒情方法研究[J].节水灌溉, 2020(8):6.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
10 |
马春芽, 王景雷, 黄修桥. 遥感监测土壤水分研究进展[J]. 节水灌溉, 2018(5) : 70-74, 78.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
11 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
12 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
13 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
14 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
15 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
16 |
郭二旺, 郭乙霏, 罗蔚然, 等. 基于Landsat8和Sentinel-1A数据的焦作广利灌区夏玉米土壤墒情监测方法研究[J]. 中国农村水利水电, 2019(7): 22-25, 34.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
17 |
王文婷, 郭乙霏. 多源遥感数据及其融合技术在土壤墒情监测中的应用[J]. 节水灌溉, 2018(6): 63-66, 70.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
18 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
19 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
20 |
尉毓姣, 朱琳, 曹鑫宇, 等. 基于ESTARFM模型的南京地区蒸散发降尺度研究[J]. 生态学报, 2022 (15): 1-11.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
21 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
22 |
罗亚东, 许齐, 郭宇宏, 等. 尺度效应对多光谱遥感监测土壤水分影响研究[J]. 节水灌溉, 2023(2): 20-27.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
23 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
24 |
匡中正,涂 超,李江沙,等.基于多源数据融合的城市雨污排水WebGIS系统及应用[J].测绘地理信息,2023,48(6):138-141.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
25 |
曾炜健,张棋斐,吴志峰,等.对比多源遥感数据在滨海湿地提取中的差异[J].测绘地理信息,2023,48(3):71-78.
{{custom_citation.content}}
{{custom_citation.annotation}}
|
26 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
27 |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
{{custom_ref.label}} |
{{custom_citation.content}}
{{custom_citation.annotation}}
|
/
〈 |
|
〉 |