{"title":"Forecasting Lakes' Chlorophyll Concentrations Using Satellite Images and Generative Adversarial Networks","authors":"Nikolaos Nagkoulis, Giorgos Vasiloudis, Anastasia Moumtzidou, Ilias Gialampoukidis, Stefanos Vrochidis, Ioannis Kompatsiaris","doi":"10.1029/2024wr037138","DOIUrl":null,"url":null,"abstract":"Satellite data are extensively used for water quality monitoring purposes, offering a significantly reduced cost compared to in situ data sampling. Using past measurements to predict future conditions remains a challenging task, because of the complexity of the natural phenomena that are involved, with great potential in terms of water resources management. This paper proposes a model that can be used to forecast Chlorophyll-<span data-altimg=\"/cms/asset/56d9a941-e73e-41ed-9d83-e30ef0d09f2c/wrcr27456-math-0001.png\"></span><mjx-container ctxtmenu_counter=\"118\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0001.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0001\" display=\"inline\" location=\"graphic/wrcr27456-math-0001.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container> (Chl-<span data-altimg=\"/cms/asset/400bae36-5108-47c6-9861-aebbf2fd9567/wrcr27456-math-0002.png\"></span><mjx-container ctxtmenu_counter=\"119\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0002.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0002\" display=\"inline\" location=\"graphic/wrcr27456-math-0002.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container>) values in water bodies, which are a common water quality indicator. The operation of the model lays on the fact that typically Chl-<span data-altimg=\"/cms/asset/82a2909b-be9e-4112-9864-6d6f58e3c390/wrcr27456-math-0003.png\"></span><mjx-container ctxtmenu_counter=\"120\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0003.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0003\" display=\"inline\" location=\"graphic/wrcr27456-math-0003.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container> increases and decreases periodically. First, we apply C2RCC, which is a common atmospheric correction algorithm, to Sentinel-2 images to get Chl-<span data-altimg=\"/cms/asset/6822565d-3634-4c2c-980f-23edf866c262/wrcr27456-math-0004.png\"></span><mjx-container ctxtmenu_counter=\"121\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0004.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0004\" display=\"inline\" location=\"graphic/wrcr27456-math-0004.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container> maps for 15 lakes for 12 consecutive months around Europe. Then, we use this data set (<span data-altimg=\"/cms/asset/b768599c-b07f-4709-8c0a-4f35a9a8bb43/wrcr27456-math-0005.png\"></span><mjx-container ctxtmenu_counter=\"122\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0005.png\"><mjx-semantics><mjx-mrow><mjx-mo data-semantic- data-semantic-role=\"equality\" data-semantic-speech=\"tilde\" data-semantic-type=\"relation\"><mjx-c></mjx-c></mjx-mo></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0005\" display=\"inline\" location=\"graphic/wrcr27456-math-0005.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mo data-semantic-=\"\" data-semantic-role=\"equality\" data-semantic-speech=\"tilde\" data-semantic-type=\"relation\">∼</mo></mrow>${\\sim} $</annotation></semantics></math></mjx-assistive-mml></mjx-container>1,000 Sentinel-2 images) to train a Generative Adversarial Network (GAN) to recognize spatiotemporal patterns. To accomplish this task, pix2pix algorithm is employed, matching consecutive past and current Chl-<span data-altimg=\"/cms/asset/8c64a785-16c3-49b9-9052-6cc5a9f04ef0/wrcr27456-math-0006.png\"></span><mjx-container ctxtmenu_counter=\"123\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0006.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0006\" display=\"inline\" location=\"graphic/wrcr27456-math-0006.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container> maps to future Chl-<span data-altimg=\"/cms/asset/2f8f0758-1901-4aaf-ac2c-9dd207001e7b/wrcr27456-math-0007.png\"></span><mjx-container ctxtmenu_counter=\"124\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0007.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0007\" display=\"inline\" location=\"graphic/wrcr27456-math-0007.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container> maps. This model has been applied to 3 water bodies around Europe that are not included in the 15-lakes training data set and has been found to perform accurately, achieving high Pearson and Spearman correlations and low RMSE values. Overall, the model can be used to make Chl-<span data-altimg=\"/cms/asset/c6005afe-c349-4b01-8d61-65d6a73d5860/wrcr27456-math-0008.png\"></span><mjx-container ctxtmenu_counter=\"125\" ctxtmenu_oldtabindex=\"1\" jax=\"CHTML\" role=\"application\" sre-explorer- style=\"font-size: 103%; position: relative;\" tabindex=\"0\"><mjx-math aria-hidden=\"true\" location=\"graphic/wrcr27456-math-0008.png\"><mjx-semantics><mjx-mrow><mjx-mi data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic- data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\"><mjx-c></mjx-c></mjx-mi></mjx-mrow></mjx-semantics></mjx-math><mjx-assistive-mml display=\"inline\" unselectable=\"on\"><math altimg=\"urn:x-wiley:00431397:media:wrcr27456:wrcr27456-math-0008\" display=\"inline\" location=\"graphic/wrcr27456-math-0008.png\" xmlns=\"http://www.w3.org/1998/Math/MathML\"><semantics><mrow><mi data-semantic-=\"\" data-semantic-annotation=\"clearspeak:simple\" data-semantic-font=\"italic\" data-semantic-role=\"greekletter\" data-semantic-speech=\"alpha\" data-semantic-type=\"identifier\">α</mi></mrow>$\\alpha $</annotation></semantics></math></mjx-assistive-mml></mjx-container> maps' predictions with low computational cost and without using any in situ data and without the requirement of training for every water body.","PeriodicalId":23799,"journal":{"name":"Water Resources Research","volume":null,"pages":null},"PeriodicalIF":4.6000,"publicationDate":"2024-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Water Resources Research","FirstCategoryId":"89","ListUrlMain":"https://doi.org/10.1029/2024wr037138","RegionNum":1,"RegionCategory":"地球科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ENVIRONMENTAL SCIENCES","Score":null,"Total":0}
引用次数: 0
Abstract
Satellite data are extensively used for water quality monitoring purposes, offering a significantly reduced cost compared to in situ data sampling. Using past measurements to predict future conditions remains a challenging task, because of the complexity of the natural phenomena that are involved, with great potential in terms of water resources management. This paper proposes a model that can be used to forecast Chlorophyll- (Chl-) values in water bodies, which are a common water quality indicator. The operation of the model lays on the fact that typically Chl- increases and decreases periodically. First, we apply C2RCC, which is a common atmospheric correction algorithm, to Sentinel-2 images to get Chl- maps for 15 lakes for 12 consecutive months around Europe. Then, we use this data set (1,000 Sentinel-2 images) to train a Generative Adversarial Network (GAN) to recognize spatiotemporal patterns. To accomplish this task, pix2pix algorithm is employed, matching consecutive past and current Chl- maps to future Chl- maps. This model has been applied to 3 water bodies around Europe that are not included in the 15-lakes training data set and has been found to perform accurately, achieving high Pearson and Spearman correlations and low RMSE values. Overall, the model can be used to make Chl- maps' predictions with low computational cost and without using any in situ data and without the requirement of training for every water body.
期刊介绍:
Water Resources Research (WRR) is an interdisciplinary journal that focuses on hydrology and water resources. It publishes original research in the natural and social sciences of water. It emphasizes the role of water in the Earth system, including physical, chemical, biological, and ecological processes in water resources research and management, including social, policy, and public health implications. It encompasses observational, experimental, theoretical, analytical, numerical, and data-driven approaches that advance the science of water and its management. Submissions are evaluated for their novelty, accuracy, significance, and broader implications of the findings.