Skip to content

FLODIS observed flood footprints ↔ impacts — introduction#

FLODIS (Mester, Frieler & Schewe, Potsdam Institute for Climate Impact Research; Scientific Data 10, 482, 2023) is the observed hazard-footprint → impact bridge: it links recorded flood impacts — human displacements, fatalities and economic damages — to the satellite flood footprints that caused them. It is the global companion to the European hanze backend, and the matched-and-enriched sibling of the raw impact tables in emdat.

earthlens ships a single flodis backend. This page orients it. For the hands-on walkthrough see Usage; the rendered API is the Reference page.

FLODIS matches three impact sources to one footprint source, in space and time:

  • impacts — EM-DAT fatalities and economic damages (CRED/UCLouvain), and IDMC human displacements.
  • footprints — the Global Flood Database (GFD; Tellman et al. 2021), a satellite-observed inventory of historic flood extents.
  • exposure — for each matched event, the affected population (GHSL + GPW), GDP, and critical-infrastructure counts (power, health, transport, telecommunication, water) intersected with the footprint.

What you get#

flodis is a tabular backend: download() returns a pandas.DataFrame, one row per matched event, carrying FLODIS's documented columns. dataset= selects which of the two tables you fetch:

  • dataset="damages" (the default) → the EM-DAT deaths/damages table (697 rows), keyed on the EM-DAT disasterno. Columns include total_deaths, no_affected_EMDAT, total_damages_(000_USD), the GFD-match block (GFD_matches, GFD_duration, matching_type), and the per-event exposure sums (pop_affected_sum_GHSL/GPW, GDP_affected_sum, the infrastructure counts).
  • dataset="displacement" → the IDMC displacement table (335 rows), keyed on the GADM GID_1 / GID_2 admin codes. Columns include displacements, num_provinces / num_districts, and the same GFD-match and exposure blocks.

Know these three things before you use it#

1. It is a pinned Zenodo release, and it is not current. earthlens pins the exact Zenodo record 8123096 (CC-BY-4.0) rather than a moving GitHub branch, so a request is reproducible; the files are byte-identical to the frozen upstream repository. FLODIS covers 2000–2018 — the overlap window of the three input archives — so it is a historical record, not a live feed. Both CSVs are small (≈395 KB and ≈185 KB); earthlens downloads them directly and caches them under your output directory.

2. earthlens fetches the impact tables, not the footprints. FLODIS carries the keys to the footprints, not the geometry. earthlens does not re-implement GDIS or the Global Flood Database — it exposes the join keys so you can attach the footprints from the layers earthlens already ships: the GDIS disaster geometry from the emdat backend (joined on disasterno) and the GFD flood extents from the gee backend (GLOBAL_FLOOD_DB/MODIS_EVENTS/V1). Usage shows the join.

3. The two tables are keyed differently. damages is keyed on the EM-DAT disasterno; displacement is keyed on the GADM GID_1 / GID_2. Filter both by country= (ISO3, e.g. "MOZ") and a [start, end] year window; filter the displacement table additionally by gid= (a GADM code). The tables carry no per-row coordinates — a bounding box is not a filter axis here, because the geometry lives in the emdat / gee layers you join to.

Licence & attribution#

FLODIS is CC-BY-4.0 (Zenodo record 8123096). Cite Mester, Frieler & Schewe (2023) and credit the upstream sources — IDMC, EM-DAT (CRED/UCLouvain) and the Global Flood Database (Tellman et al. 2021) — in any derived work. The backend logs the citation on each download().