Evaluating a global soil moisture dataset from a multitask model (GSM3 v1.0) with potential applications for crop threats. Liu, J., Hughes, D., Rahmani, F., Lawson, K., & Shen, C. Geoscientific Model Development, 16(5):1553–1567, March, 2023. ![Evaluating a global soil moisture dataset from a multitask model (GSM3 v1.0) with potential applications for crop threats [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex Abstract. Climate change threatens our ability to grow food for an ever-increasing population. There is a need for high-quality soil moisture predictions in under-monitored regions like Africa. However, it is unclear if soil moisture processes are globally similar enough to allow our models trained on available in situ data to maintain accuracy in unmonitored regions. We present a multitask long short-term memory (LSTM) model that learns simultaneously from global satellite-based data and in situ soil moisture data. This model is evaluated in both random spatial holdout mode and continental holdout mode (trained on some continents, tested on a different one). The model compared favorably to current land surface models, satellite products, and a candidate machine learning model, reaching a global median correlation of 0.792 for the random spatial holdout test. It behaved surprisingly well in Africa and Australia, showing high correlation even when we excluded their sites from the training set, but it performed relatively poorly in Alaska where rapid changes are occurring. In all but one continent (Asia), the multitask model in the worst-case scenario test performed better than the soil moisture active passive (SMAP) 9 km product. Factorial analysis has shown that the LSTM model's accuracy varies with terrain aspect, resulting in lower performance for dry and south-facing slopes or wet and north-facing slopes. This knowledge helps us apply the model while understanding its limitations. This model is being integrated into an operational agricultural assistance application which currently provides information to 13 million African farmers.
@article{liu_evaluating_2023,
title = {Evaluating a global soil moisture dataset from a multitask model ({GSM3} v1.0) with potential applications for crop threats},
volume = {16},
copyright = {https://creativecommons.org/licenses/by/4.0/},
issn = {1991-9603},
url = {https://gmd.copernicus.org/articles/16/1553/2023/},
doi = {10.5194/gmd-16-1553-2023},
abstract = {Abstract. Climate change threatens our ability to grow food for an ever-increasing population. There is a
need for high-quality soil moisture predictions in under-monitored regions
like Africa. However, it is unclear if soil moisture processes are globally
similar enough to allow our models trained on available in situ data to
maintain accuracy in unmonitored regions. We present a multitask long
short-term memory (LSTM) model that learns simultaneously from global
satellite-based data and in situ soil moisture data. This model is evaluated in
both random spatial holdout mode and continental holdout mode (trained on
some continents, tested on a different one). The model compared favorably to
current land surface models, satellite products, and a candidate machine
learning model, reaching a global median correlation of 0.792 for the random
spatial holdout test. It behaved surprisingly well in Africa and Australia,
showing high correlation even when we excluded their sites from the training
set, but it performed relatively poorly in Alaska where rapid changes are
occurring. In all but one continent (Asia), the multitask model in the
worst-case scenario test performed better than the soil moisture active
passive (SMAP) 9 km product. Factorial analysis has shown that the LSTM model's
accuracy varies with terrain aspect, resulting in lower performance for dry
and south-facing slopes or wet and north-facing slopes. This knowledge
helps us apply the model while understanding its limitations. This model is
being integrated into an operational agricultural assistance application
which currently provides information to 13 million African farmers.},
language = {en},
number = {5},
urldate = {2024-11-15},
journal = {Geoscientific Model Development},
author = {Liu, Jiangtao and Hughes, David and Rahmani, Farshid and Lawson, Kathryn and Shen, Chaopeng},
month = mar,
year = {2023},
pages = {1553--1567},
}
Downloads: 0
{"_id":"cxxX6XwGMw2wFmCGq","bibbaseid":"liu-hughes-rahmani-lawson-shen-evaluatingaglobalsoilmoisturedatasetfromamultitaskmodelgsm3v10withpotentialapplicationsforcropthreats-2023","author_short":["Liu, J.","Hughes, D.","Rahmani, F.","Lawson, K.","Shen, C."],"bibdata":{"bibtype":"article","type":"article","title":"Evaluating a global soil moisture dataset from a multitask model (GSM3 v1.0) with potential applications for crop threats","volume":"16","copyright":"https://creativecommons.org/licenses/by/4.0/","issn":"1991-9603","url":"https://gmd.copernicus.org/articles/16/1553/2023/","doi":"10.5194/gmd-16-1553-2023","abstract":"Abstract. Climate change threatens our ability to grow food for an ever-increasing population. There is a need for high-quality soil moisture predictions in under-monitored regions like Africa. However, it is unclear if soil moisture processes are globally similar enough to allow our models trained on available in situ data to maintain accuracy in unmonitored regions. We present a multitask long short-term memory (LSTM) model that learns simultaneously from global satellite-based data and in situ soil moisture data. This model is evaluated in both random spatial holdout mode and continental holdout mode (trained on some continents, tested on a different one). The model compared favorably to current land surface models, satellite products, and a candidate machine learning model, reaching a global median correlation of 0.792 for the random spatial holdout test. It behaved surprisingly well in Africa and Australia, showing high correlation even when we excluded their sites from the training set, but it performed relatively poorly in Alaska where rapid changes are occurring. In all but one continent (Asia), the multitask model in the worst-case scenario test performed better than the soil moisture active passive (SMAP) 9 km product. Factorial analysis has shown that the LSTM model's accuracy varies with terrain aspect, resulting in lower performance for dry and south-facing slopes or wet and north-facing slopes. This knowledge helps us apply the model while understanding its limitations. This model is being integrated into an operational agricultural assistance application which currently provides information to 13 million African farmers.","language":"en","number":"5","urldate":"2024-11-15","journal":"Geoscientific Model Development","author":[{"propositions":[],"lastnames":["Liu"],"firstnames":["Jiangtao"],"suffixes":[]},{"propositions":[],"lastnames":["Hughes"],"firstnames":["David"],"suffixes":[]},{"propositions":[],"lastnames":["Rahmani"],"firstnames":["Farshid"],"suffixes":[]},{"propositions":[],"lastnames":["Lawson"],"firstnames":["Kathryn"],"suffixes":[]},{"propositions":[],"lastnames":["Shen"],"firstnames":["Chaopeng"],"suffixes":[]}],"month":"March","year":"2023","pages":"1553–1567","bibtex":"@article{liu_evaluating_2023,\n\ttitle = {Evaluating a global soil moisture dataset from a multitask model ({GSM3} v1.0) with potential applications for crop threats},\n\tvolume = {16},\n\tcopyright = {https://creativecommons.org/licenses/by/4.0/},\n\tissn = {1991-9603},\n\turl = {https://gmd.copernicus.org/articles/16/1553/2023/},\n\tdoi = {10.5194/gmd-16-1553-2023},\n\tabstract = {Abstract. Climate change threatens our ability to grow food for an ever-increasing population. There is a\nneed for high-quality soil moisture predictions in under-monitored regions\nlike Africa. However, it is unclear if soil moisture processes are globally\nsimilar enough to allow our models trained on available in situ data to\nmaintain accuracy in unmonitored regions. We present a multitask long\nshort-term memory (LSTM) model that learns simultaneously from global\nsatellite-based data and in situ soil moisture data. This model is evaluated in\nboth random spatial holdout mode and continental holdout mode (trained on\nsome continents, tested on a different one). The model compared favorably to\ncurrent land surface models, satellite products, and a candidate machine\nlearning model, reaching a global median correlation of 0.792 for the random\nspatial holdout test. It behaved surprisingly well in Africa and Australia,\nshowing high correlation even when we excluded their sites from the training\nset, but it performed relatively poorly in Alaska where rapid changes are\noccurring. In all but one continent (Asia), the multitask model in the\nworst-case scenario test performed better than the soil moisture active\npassive (SMAP) 9 km product. Factorial analysis has shown that the LSTM model's\naccuracy varies with terrain aspect, resulting in lower performance for dry\nand south-facing slopes or wet and north-facing slopes. This knowledge\nhelps us apply the model while understanding its limitations. This model is\nbeing integrated into an operational agricultural assistance application\nwhich currently provides information to 13 million African farmers.},\n\tlanguage = {en},\n\tnumber = {5},\n\turldate = {2024-11-15},\n\tjournal = {Geoscientific Model Development},\n\tauthor = {Liu, Jiangtao and Hughes, David and Rahmani, Farshid and Lawson, Kathryn and Shen, Chaopeng},\n\tmonth = mar,\n\tyear = {2023},\n\tpages = {1553--1567},\n}\n\n\n\n\n\n\n\n","author_short":["Liu, J.","Hughes, D.","Rahmani, F.","Lawson, K.","Shen, C."],"key":"liu_evaluating_2023","id":"liu_evaluating_2023","bibbaseid":"liu-hughes-rahmani-lawson-shen-evaluatingaglobalsoilmoisturedatasetfromamultitaskmodelgsm3v10withpotentialapplicationsforcropthreats-2023","role":"author","urls":{"Paper":"https://gmd.copernicus.org/articles/16/1553/2023/"},"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero/tereno","dataSources":["cq3J5xX6zmBvc2TQC"],"keywords":[],"search_terms":["evaluating","global","soil","moisture","dataset","multitask","model","gsm3","potential","applications","crop","threats","liu","hughes","rahmani","lawson","shen"],"title":"Evaluating a global soil moisture dataset from a multitask model (GSM3 v1.0) with potential applications for crop threats","year":2023}