Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation
Surface albedo, as a key parameter determining the partition of solar radiation at the Earth’s surface, has been developed into a satellite-based product from various Earth observation systems to serve numerous global or regional applications. Studies point out that apparent uncertainty can be intro...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-04-01
|
| Series: | Remote Sensing |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2072-4292/14/9/2001 |
| _version_ | 1797503107517644800 |
|---|---|
| author | Jian Yang Yanmin Shuai Junbo Duan Donghui Xie Qingling Zhang Ruishan Zhao |
| author_facet | Jian Yang Yanmin Shuai Junbo Duan Donghui Xie Qingling Zhang Ruishan Zhao |
| author_sort | Jian Yang |
| collection | DOAJ |
| description | Surface albedo, as a key parameter determining the partition of solar radiation at the Earth’s surface, has been developed into a satellite-based product from various Earth observation systems to serve numerous global or regional applications. Studies point out that apparent uncertainty can be introduced into albedo retrieval without consideration of surface anisotropy, which is a challenge to albedo estimation especially from observations with fewer angular samplings. Researchers have begun to introduce smoothed anisotropy prior knowledge into albedo estimation to improve the inversion efficiency, or for the scenario of observations with signal or poor angular sampling. Thus, it is necessary to further understand the potential influence of smoothed anisotropy features adopted in albedo estimation. We investigated the albedo variation induced by BRDF smoothing at both temporal and spatial scales over six typical landscapes in North America using MODIS standard anisotropy products with high quality BRDF inversed from multi-angle observations in 500 m and 5.6 km spatial resolutions. Components of selected typical landscapes were assessed with the confidence of the MCD12 land cover product and 30 m CDL (cropland data layer) classification maps followed by an evaluation of spatial heterogeneity in 30 m scale through the semi-variogram model. High quality BRDF of MODIS standard anisotropy products were smoothed in multi-temporal scales of 8 days, 16 days, and 32 days, and in multi-spatial scales from 500 m to 5.6 km. The induced relative and absolute albedo differences were estimated using the RossThick-LiSparseR model and BRDFs smoothed before and after spatiotemporal smoothing. Our results show that albedo estimated using BRDFs smoothed temporally from daily to monthly over each scenario exhibits relative differences of 11.3%, 12.5%, and 27.2% and detectable absolute differences of 0.025, 0.012, and 0.013, respectively, in MODIS near-infrared (0.7–5.0 µm), short-wave (0.3–5.0 µm), and visible (0.3–0.7 µm) broad bands. When BRDFs of investigated landscapes are smoothed from 500 m to 5.6 km, variations of estimated albedo can achieve up to 36.5%, 37.1%, and 94.7% on relative difference and absolute difference of 0.037, 0.024, and 0.018, respectively, in near-infrared (0.7–5.0 µm), short wave (0.3–5.0 µm), and visible (0.3–0.7 µm) broad bands. In addition, albedo differences caused by temporal smoothing show apparent seasonal characteristic that the differences are significantly higher in spring and summer than those in autumn and winter, while albedo differences induced by spatial smoothing exhibit a noticeable relationship with sill values of a fitted semi-variogram marked by a correlation coefficient of 0.8876. Both relative and absolute albedo differences induced by BRDF smoothing are demonstrated to be captured, thus, it is necessary to avoid the smoothing process in quantitative remote sensing communities, especially when immediate anisotropy retrievals are available at the required spatiotemporal scale. |
| first_indexed | 2024-03-10T03:45:47Z |
| format | Article |
| id | doaj.art-4ccf397049a842e8abb1a8d4a9e76ba1 |
| institution | Directory Open Access Journal |
| issn | 2072-4292 |
| language | English |
| last_indexed | 2024-03-10T03:45:47Z |
| publishDate | 2022-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Remote Sensing |
| spelling | doaj.art-4ccf397049a842e8abb1a8d4a9e76ba12023-11-23T09:09:05ZengMDPI AGRemote Sensing2072-42922022-04-01149200110.3390/rs14092001Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo EstimationJian Yang0Yanmin Shuai1Junbo Duan2Donghui Xie3Qingling Zhang4Ruishan Zhao5College of Surveying and Mapping and Geographic Science, Liaoning Technical University, Fuxin 123000, ChinaXinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi 830011, ChinaCollege of Surveying and Mapping and Geographic Science, Liaoning Technical University, Fuxin 123000, ChinaState Key Laboratory of Remote Sensing Science, Faculty of Geographical Science, Beijing Normal University, Beijing 100875, ChinaSchool of Aeronautics and Astronautics, Sun Yat-sen University, Guangzhou 510006, ChinaCollege of Surveying and Mapping and Geographic Science, Liaoning Technical University, Fuxin 123000, ChinaSurface albedo, as a key parameter determining the partition of solar radiation at the Earth’s surface, has been developed into a satellite-based product from various Earth observation systems to serve numerous global or regional applications. Studies point out that apparent uncertainty can be introduced into albedo retrieval without consideration of surface anisotropy, which is a challenge to albedo estimation especially from observations with fewer angular samplings. Researchers have begun to introduce smoothed anisotropy prior knowledge into albedo estimation to improve the inversion efficiency, or for the scenario of observations with signal or poor angular sampling. Thus, it is necessary to further understand the potential influence of smoothed anisotropy features adopted in albedo estimation. We investigated the albedo variation induced by BRDF smoothing at both temporal and spatial scales over six typical landscapes in North America using MODIS standard anisotropy products with high quality BRDF inversed from multi-angle observations in 500 m and 5.6 km spatial resolutions. Components of selected typical landscapes were assessed with the confidence of the MCD12 land cover product and 30 m CDL (cropland data layer) classification maps followed by an evaluation of spatial heterogeneity in 30 m scale through the semi-variogram model. High quality BRDF of MODIS standard anisotropy products were smoothed in multi-temporal scales of 8 days, 16 days, and 32 days, and in multi-spatial scales from 500 m to 5.6 km. The induced relative and absolute albedo differences were estimated using the RossThick-LiSparseR model and BRDFs smoothed before and after spatiotemporal smoothing. Our results show that albedo estimated using BRDFs smoothed temporally from daily to monthly over each scenario exhibits relative differences of 11.3%, 12.5%, and 27.2% and detectable absolute differences of 0.025, 0.012, and 0.013, respectively, in MODIS near-infrared (0.7–5.0 µm), short-wave (0.3–5.0 µm), and visible (0.3–0.7 µm) broad bands. When BRDFs of investigated landscapes are smoothed from 500 m to 5.6 km, variations of estimated albedo can achieve up to 36.5%, 37.1%, and 94.7% on relative difference and absolute difference of 0.037, 0.024, and 0.018, respectively, in near-infrared (0.7–5.0 µm), short wave (0.3–5.0 µm), and visible (0.3–0.7 µm) broad bands. In addition, albedo differences caused by temporal smoothing show apparent seasonal characteristic that the differences are significantly higher in spring and summer than those in autumn and winter, while albedo differences induced by spatial smoothing exhibit a noticeable relationship with sill values of a fitted semi-variogram marked by a correlation coefficient of 0.8876. Both relative and absolute albedo differences induced by BRDF smoothing are demonstrated to be captured, thus, it is necessary to avoid the smoothing process in quantitative remote sensing communities, especially when immediate anisotropy retrievals are available at the required spatiotemporal scale.https://www.mdpi.com/2072-4292/14/9/2001BRDFalbedo variationtemporal smoothingspatial smoothingmulti-angle retrieval |
| spellingShingle | Jian Yang Yanmin Shuai Junbo Duan Donghui Xie Qingling Zhang Ruishan Zhao Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation Remote Sensing BRDF albedo variation temporal smoothing spatial smoothing multi-angle retrieval |
| title | Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation |
| title_full | Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation |
| title_fullStr | Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation |
| title_full_unstemmed | Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation |
| title_short | Impact of BRDF Spatiotemporal Smoothing on Land Surface Albedo Estimation |
| title_sort | impact of brdf spatiotemporal smoothing on land surface albedo estimation |
| topic | BRDF albedo variation temporal smoothing spatial smoothing multi-angle retrieval |
| url | https://www.mdpi.com/2072-4292/14/9/2001 |
| work_keys_str_mv | AT jianyang impactofbrdfspatiotemporalsmoothingonlandsurfacealbedoestimation AT yanminshuai impactofbrdfspatiotemporalsmoothingonlandsurfacealbedoestimation AT junboduan impactofbrdfspatiotemporalsmoothingonlandsurfacealbedoestimation AT donghuixie impactofbrdfspatiotemporalsmoothingonlandsurfacealbedoestimation AT qinglingzhang impactofbrdfspatiotemporalsmoothingonlandsurfacealbedoestimation AT ruishanzhao impactofbrdfspatiotemporalsmoothingonlandsurfacealbedoestimation |