class: top, left, inverse, title-slide .title[ # Impact of floods on property and agriculture ] .author[ ### Lampros Sp. Mouselimis ] .institute[ ### Monopteryx ] .date[ ### 2023-07-20
monopteryx-dashboard/
linkedin
Nooksack River flood, November 2021
] --- class:hide_logo
# Definition and causes of Floods <br> Floods are one of the many natural disasters affecting people and property around the world every year. Floods can be attributed to many factors such as, * heavy rainfall * rapid snow melt * storm surges <br> There are three main types of floods: * Flash floods are caused by a sudden, intense, and heavy rainfall. This type of rainfall can occur within minutes or hours and can be very dangerous because they give people little time to prepare. * River floods are caused by a river overflowing its banks. They can occur when there is heavy rainfall over a long period of time, or when snow melts rapidly. * Coastal floods are caused by a storm surge, which is a large wave of water that is pushed inland by a storm. Coastal floods can occur during hurricanes, tropical storms, and tsunamis. --- # Monitoring & Assessment of flood events As of July 2023, there are various satellite imagery providers which allow to monitor and assess flood events, for instance Commercial Providers: * ICEYE * Maxar * Planet Open Access Data Providers: * Copernicus Emergency Management Service * USGS (United States Geological Survey) .pull-left[ The table on the right summarizes the differences between commercial and open-access services. Someone would choose between the two groups depending on the needs and budget. For instance, if high revisit frequency and high resolution are required then a commercial provider would be the right choice, otherwise open-access providers can work for most of cases. ] .pull-right[ <img src="images/differences_imagery_providers.png" width="65%" style="display: block; margin: auto 0 auto auto;" /> ] --- # Post-assessment of a flood event Open-access remote sensing data (with a smaller revisit frequency compared to commercial) can be used to assess a flood event. In this use case, the "Pajaro flooding" which occurred in March 2023 and was [covered in media reports](https://www.latimes.com/california/story/2023-03-20/a-long-history-of-racism-set-the-stage-for-pajaro-flooding) will be analyzed using the "Copernicus Emergency Management Service" and specifically the "Global Flood Monitoring" (GFM) product. .pull-left[ * The Area of Interest (AOI) on the right was used for the post-assessment * Data for the "Pajaro River" AOI was downloaded from "2023-03-02" to "2023-03-31" * The pre-processing and visualization were performed using R programming and remote sensing R packages * The "Land Cover Product" and the "building footprints" of the AOI for the year 2022 were downloaded from the "Planetary Computer Data Catalog" * The Crop Sequence Boundaries (USDA - United States Department of Agriculture) were used to extract information related to the cropland of the AOI ] .pull-right[ <img src="images/aoi.png" width="55%" style="display: block; margin: auto;" /> ] --- class:hide_logo # Flood assessment of built-up areas The GFM product uses historical time series of Sentinel-1 Synthetic Aperture Radar (SAR) Level-1 Ground Range Detected (GRD) satellite imagery (from the European Space Agency - ESA) and flood mapping algorithms. From this product I used, * the flood ensemble values (flooded areas) * the uncertainty values (estimated uncertainty of flood mapping) .pull-left[ Then, the pre-processed GFM data was merged with the land cover data and the building footprints. * The image on the top-right shows the flooded areas in green color mainly occurred on "2023-03-14" (as the legend shows too) <br> * The image on the bottom-right shows the flooded area which is dominated by the yellow color (a major flood occurred on "2023-03-31") with pixels in the boundaries of different colors indicating that there were floods in the other 3 days as well ] .pull-right[ <img src="images/build_up/flood_date_build_up_2.png" width="37%" style="display: block; margin: auto;" /> <img src="images/build_up/flood_date_build_up_1.png" width="37%" style="display: block; margin: auto;" /> ] --- class:hide_logo # Flood assessment of built-up areas .pull-left[ The bar plot on top, which is based on the "flood-ensemble values", shows that in almost all four days the floods affected building footprints (left) and the majority of flooded buildings occurred on "2023-03-14" (right) <br><br><br><br><br><br> The uncertainty (or likelihood) heatmap on the bottom, shows the confidence of the flood algorithms to classify the flood pixels. The values range between 0 and 100 and values close to 100 indicate high confidence. Based on this heatmap we can say that pixels which include buildings and were classified as floods at 2023-03-02" are highly probable. ] .pull-right[ <img src="images/build_up/bar_plot_build_up.png" width="90%" style="display: block; margin: auto;" /> <img src="images/build_up/heatmap_build_up.png" width="80%" style="display: block; margin: auto;" /> ] --- class:hide_logo # Flood assessment of Cropland Floods can have a negative impact also on crops. The next image shows the extent of floods on cropland for the AOI which shows (based on the legend) that on "2023-03-26" the floods were more severe, <img src="images/cropland/flood_cropland_1.png" width="90%" style="display: block; margin: auto;" /> --- class:hide_logo # Flood assessment of Cropland The next bar-plot shows the number of crop fields affected (on the right) and justifies that at "2023-03-26" there were more flooded crop fields, <img src="images/cropland/bar_plot_crops.png" width="100%" style="display: block; margin: auto;" /> --- class:hide_logo # Flood assessment of Cropland The majority of these fields were near the river (marked with red in the next image) .pull-left[ <img src="images/cropland/flood_cropland_2_wo_line.png" width="100%" style="display: block; margin: auto;" /> ] .pull-right[ <img src="images/cropland/flood_cropland_2.png" width="90%" style="display: block; margin: auto;" /> ] --- class:hide_logo # Flood assessment of Cropland .pull-left[ * The uncertainty range of values shows high confidence levels at "2023-03-14" (top heatmap) <br><br><br><br><br><br><br><br><br><br><br><br> * There are also cases of partly flooded crop fields (bottom plot), * in blue: the crop field area * in red: the partly flooded area * in orange: the outside of the field area ] .pull-right[ <img src="images/cropland/heatmap_crops.png" width="100%" style="display: block; margin: auto;" /> <img src="images/cropland/proportion_field.png" width="30%" style="display: block; margin: auto;" /> ] --- class:hide_logo # Flood assessment of Cropland The following plot shows which crop types were affected by floods on thousand acres. This plot takes into consideration the uncertainty levels and the proportion of flooded crop fields as shown in the previous slide, <img src="images/cropland/crop_types_flood.png" width="105%" style="display: block; margin: auto;" /> --- # Professional Services <br> If you are looking for a professional to process remote sensing data and extract information using the R and Python programming languages don't hesitate to contact me <br> [https://monopteryx.netlify.app/contact/](https://monopteryx.netlify.app/contact/) <br><br><br><br> ### References: * [Copernicus Emergency Management Service](https://emergency.copernicus.eu/) * [Global Flood Monitoring (GFM)](https://extwiki.eodc.eu/GFM) * [Los Angeles Times report](https://www.latimes.com/california/story/2023-03-20/a-long-history-of-racism-set-the-stage-for-pajaro-flooding) * [R programming](https://www.r-project.org/) * [Planetary Computer Data Catalog](https://planetarycomputer.microsoft.com/catalog) * [Crop Sequence Boundaries (USDA - United States Department of Agriculture)](https://www.nass.usda.gov/Research_and_Science/Crop-Sequence-Boundaries/index.php)