FLOPROS flood-protection standards — a guided tour¶

This notebook teaches the flopros backend: how to fetch the FLOPROS global database of flood-protection standards (Scussolini et al., 2016) and explore it. The data is public (an NHESS-2016 supplement shapefile, CC-BY-3.0, no credentials).

FLOPROS is the defended-vs-undefended correction: each of ~4650 subnational polygons carries the flood protection in force as a return period in years — e.g. "defended against the 100-year flood". Clip a hazard map by the local standard and you separate protected from exposed areas.

The request axes¶

A request selects a protection layer and, optionally, a unit or a region:

Argument	Meaning	Values
layer	which protection layer(s)	merged_riverine (recommended), modelled_riverine, design_*/policy_* (min/max, riverine + coastal), or None for all
country	keep one unit	a name/geonunit (case-insensitive)
lat_lim / lon_lim	keep a region	a WGS84 bounding box
geometry	output shape	True -> FeatureCollection, False -> DataFrame

A protection value of 0 means no standard is recorded for that layer in that unit.

Setup¶

Imports and a single notebook-relative cache/output dir (out/, gitignored).

In [ ]:

import matplotlib.pyplot as plt

from earthlens.core import EarthLens

DATA = "out"

import matplotlib.pyplot as plt from earthlens.core import EarthLens DATA = "out"

Quickstart — map the merged protection standard¶

The shortest end-to-end call: the merged riverine layer for every subnational unit. The first call downloads the ~14 MB NHESS supplement zip and caches it under out/; every later call reuses it.

In [ ]:

fc = EarthLens(
    "flopros",
    layer="merged_riverine",
    path=DATA,
    cache_dir=DATA,
).download()

print(len(fc), "units |", list(fc.columns))

fc = EarthLens( "flopros", layer="merged_riverine", path=DATA, cache_dir=DATA, ).download() print(len(fc), "units |", list(fc.columns))

A choropleth of the merged_riverine column — the protection return period (years) per unit:

In [ ]:

ax = fc.plot(
    column="merged_riverine",
    cmap="YlGnBu",
    legend=True,
    legend_kwds={"label": "Protection standard (return period, yr)", "shrink": 0.6},
    figsize=(11, 5.5),
    edgecolor="grey",
    linewidth=0.1,
)
ax.set_title("FLOPROS merged riverine protection standard")
ax.set_axis_off()
plt.show()

ax = fc.plot( column="merged_riverine", cmap="YlGnBu", legend=True, legend_kwds={"label": "Protection standard (return period, yr)", "shrink": 0.6}, figsize=(11, 5.5), edgecolor="grey", linewidth=0.1, ) ax.set_title("FLOPROS merged riverine protection standard") ax.set_axis_off() plt.show()

The highest standards cluster in the wealthy, low-lying river basins of north-west Europe and East Asia — the pattern the ranked table below confirms.

In [ ]:

fc[["name", "geonunit", "merged_riverine"]].sort_values(
    "merged_riverine", ascending=False
).head(8).reset_index(drop=True)

fc[["name", "geonunit", "merged_riverine"]].sort_values( "merged_riverine", ascending=False ).head(8).reset_index(drop=True)

Compare the protection layers¶

Omit layer= to keep all FLOPROS layers at once — Modelled, Merged, and the Design / Policy min & max (riverine _riverine + coastal _coastal). This reuses the cached shapefile, so no new download.

In [ ]:

allfc = EarthLens("flopros", path=DATA, cache_dir=DATA).download()
layer_cols = [
    c for c in allfc.columns if c not in ("name", "geonunit", "type_en", "geometry")
]
layer_cols

allfc = EarthLens("flopros", path=DATA, cache_dir=DATA).download() layer_cols = [ c for c in allfc.columns if c not in ("name", "geonunit", "type_en", "geometry") ] layer_cols

The Design and Policy layers bracket the modelled estimate: for a single unit you can read the modelled standard against the design range of its actual defences.

In [ ]:

cols = [
    "name",
    "modelled_riverine",
    "merged_riverine",
    "design_min_riverine",
    "design_max_riverine",
]
allfc[allfc["merged_riverine"] > 0][cols].sort_values(
    "merged_riverine", ascending=False
).head(6).reset_index(drop=True)

cols = [ "name", "modelled_riverine", "merged_riverine", "design_min_riverine", "design_max_riverine", ] allfc[allfc["merged_riverine"] > 0][cols].sort_values( "merged_riverine", ascending=False ).head(6).reset_index(drop=True)

Filter by country or bounding box¶

Pass country= (matched on name or geonunit, case-insensitive) to keep one country's units, or lat_lim/lon_lim for a region. Both reuse the cache. geometry=False drops the geometry for a plain table.

In [ ]:

de = EarthLens(
    "flopros",
    country="Germany",
    geometry=False,
    path=DATA,
    cache_dir=DATA,
).download()
de[["name", "geonunit", "merged_riverine", "modelled_riverine"]].reset_index(drop=True)

de = EarthLens( "flopros", country="Germany", geometry=False, path=DATA, cache_dir=DATA, ).download() de[["name", "geonunit", "merged_riverine", "modelled_riverine"]].reset_index(drop=True)

Takeaway¶

• One call shape — layer x (country / bbox) x geometry — covers the whole product.
• Values are protection standards as return periods (years); 0 means no standard recorded.
• Use merged_riverine for a single representative number; the Design/Policy layers give the defended range.
• Join FLOPROS to a hazard map (e.g. from aqueduct or gee) to separate protected from exposed exposure.

See the FLOPROS reference for the full layer list and licence.