4. Download & measure the depth rasters¶

Everything so far was free (index only). When you want the actual water-depth GeoTIFFs, .download() fetches and crops only the events you selected — and returns the catalogue itself, now carrying its files, so the result stays actionable. When you run this interactively you'll see a download progress bar.

This tutorial downloads real rasters, so it needs network access (it does not run in the offline test suite).

In [1]:

from pathlib import Path

import euroflood as ef

ef.settings.output_dir = Path("out")  # write crops here (relative, tidy)

cat = ef.floods("Zutphen, Netherlands", shape="bbox")  # bbox ROI (tutorial 02)

from pathlib import Path import euroflood as ef ef.settings.output_dir = Path("out") # write crops here (relative, tidy) cat = ef.floods("Zutphen, Netherlands", shape="bbox") # bbox ROI (tutorial 02)

Download a couple of events. Downloads are cached — re-running reuses the existing crops (nothing is re-fetched) unless you pass force=True.

In [2]:

dl = cat.head(2).download()
dl  # repr shows "… · ⬇ 2/2 downloaded"

dl = cat.head(2).download() dl # repr shows "… · ⬇ 2/2 downloaded"

Out[2]:

EuroFlood catalogue: 2 flood events · 2019-03-04 … 2024-02-05 · 7.3 km² total · ⬇ 2/2 downloaded

	collection	event_id	date	year	end_date	cluster_id	filename	download_url	area_km2	geometry	path
0	historic	4049036842	2019-03-04	2019	2019-04-08	333	WD_MERGE_2019-03-04---2019-04-08_duration_28_d...	https://jeodpp.jrc.ec.europa.eu/ftp/jrc-openda...	6.889	POLYGON ((6.14039 52.08873, 6.27499 52.08873, ...	out/flood_2019-03-04_id4049036842_8277ba1c.tif
1	historic	1407437849	2024-02-05	2024	2024-03-11	345	WD_MERGE_2024-02-05---2024-03-11_duration_28_d...	https://jeodpp.jrc.ec.europa.eu/ftp/jrc-openda...	0.438	POLYGON ((6.14039 52.08873, 6.27499 52.08873, ...	out/flood_2024-02-05_id1407437849_8277ba1c.tif

In [3]:

dl.files

dl.files

Out[3]:

[PosixPath('out/flood_2019-03-04_id4049036842_8277ba1c.tif'),
 PosixPath('out/flood_2024-02-05_id1407437849_8277ba1c.tif')]

Render a downloaded depth map¶

.plot(depth=True) renders the downloaded rasters. When a selection spans several events it shows their per-pixel maximum; a guardrail caps auto-downloads at 4 rasters unless you pass limit= or filter first.

In [4]:

dl.plot(depth=True)

dl.plot(depth=True)

Out[4]:

<Axes: title={'center': 'Flood depth — max of 2 events'}>

Work with a single raster: DepthRaster¶

.depths() returns lightweight DepthRaster objects (also ef.open_depth(path)). Each can read() its array, plot() / explore() itself, and save() to a .png or .html:

In [5]:

depth = dl.depths()[0]
array, _transform, crs, _nodata = depth.read()
print(
    "shape:",
    array.shape,
    "| CRS:",
    crs,
    "| max (m):",
    round(float(array.max()) * 0.01, 2),
)

depth = dl.depths()[0] array, _transform, crs, _nodata = depth.read() print( "shape:", array.shape, "| CRS:", crs, "| max (m):", round(float(array.max()) * 0.01, 2), )

shape: (451, 739) | CRS: EPSG:4326 | max (m): 5.28

In [6]:

depth.explore(tiles="grayscale")  # a single event's depth as an interactive overlay

depth.explore(tiles="grayscale") # a single event's depth as an interactive overlay

Out[6]:

Make this Notebook Trusted to load map: File -> Trust Notebook

depth.save("zutphen_depth.png")    # static image (by extension)
depth.save("zutphen_depth.html")   # interactive map

Measure them — no network, reads the cached rasters¶

.stats() is a per-event table (max / mean / p95 depth in metres, flooded area in km², water volume); .summary() is the aggregate envelope across all downloaded events. EFAS depth is in centimetres and is scaled to metres automatically (pass scale=1.0 to keep raw units).

In [7]:

dl.stats()

dl.stats()

Out[7]:

	event_id	date	wet_pixels	max_depth_m	mean_depth_m	p95_depth_m	flooded_area_km2	volume_m3	volume_Mm3
0	4049036842	2019-03-04	12463	5.57	0.830	2.45	4.9852	4138560.0	4.1386
1	1407437849	2024-02-05	582	0.38	0.176	0.31	0.2328	41068.0	0.0411

In [8]:

dl.summary()

dl.summary()

Out[8]:

{'wet_pixels': 8206,
 'max_depth_m': 5.13,
 'mean_depth_m': 0.827,
 'p95_depth_m': 2.32,
 'flooded_area_km2': 4.959,
 'volume_m3': 4100334.1,
 'volume_Mm3': 4.1003,
 'n_events': 2}

Downloads persist across sessions¶

A later query of the same area (to the same output dir) automatically finds the crops already on disk and pre-fills its path column — so you don't need to .download() again:

In [9]:

again = ef.floods("Zutphen, Netherlands", shape="bbox")  # same ROI -> finds the crops
f"{len(again.files)}/{len(again)} already downloaded"

again = ef.floods("Zutphen, Netherlands", shape="bbox") # same ROI -> finds the crops f"{len(again.files)}/{len(again)} already downloaded"

Out[9]:

'8/30 already downloaded'

In [ ]: