|Paper title||FLOMPY: Floodwater mapping and extraction of flood-affected agricultural fields|
|Form of presentation||Poster|
FLOodwater Mapping PYthon toolbox (FLOMPY) is an automatic, free and open-source python toolbox for the mapping of floodwater. An enhancement of FLOMPY related to the mapping of agricultural damaged regions from floods is presented. FLOMPY requires only a specified time of interest related to the flood event and geographical boundaries. The products of the FLOMPY consists of a) a binary mask of floodwater b) Delineated agricultural fields and c) damaged cultivated agricultural fields.
For the production of the binary mask of floodwater, the toolbox exploits the high spatial (10m) and temporal (6 days per orbit over Europe) resolution of Sentinel-1 GRD. Τhe delineation of the crop fields is based on an automated extraction algorithm using pre-flood Sentinel-2 multitemporal (optical) data. Sentinel-2 dataset were considered due to their high spatial (10m) and temporal (~5 days) resolution. In order to extract the damaged cultivated agricultural field information, vegetation and soil moisture information were used. In particular, for each delineated crop field, multitemporal vegetation and soil moisture indices from Sentinel-2 dataset were calculated. Then, according to the temporal behaviour of the indices each crop field was classified as “cultivated” or “not-cultivated”.
In this study, we present one case study related to the “Ianos” Mediterranean tropical-like cyclone over an agricultural area in central Greece. The “Ianos” cyclone took place from 14th to 19th of September 2020 and caused a lot of damage over several places in central Greece. We focus on an agricultural area of 325 km2 near Palamas where a lot of casualties were reported. The binary mask of the floodwater is extracted by exploiting Sentinel-1 intensity time series using FLOMPY`s functionalities. Delineated agricultural fields are extracted using a 3-month pre-flood Sentinel-2 dataset. The detection of flood-affected cultivated agricultural fields yielded satisfactory results based on a validation procedure using visual interpretation.
Floodwater, agricultural field, and flood-affected cultivated cropland maps can support a number of organizations related to agricultural insurance, food security, agricultural/water planning, natural disaster assessment and recovery planning. Overall, the end-user community can benefit by exploiting the proposed methodological pipeline by using the provided open-source toolbox.