Windspeeds at 175m in Germany, COSMO-REA6 and NEWA 20 Feb 2026

0-cosmo-newa-Germany.md

Windspeeds at 175m in Germany, COSMO-REA6 and NEWA

Keywords: wind-speed COSMO-REA6 NEWA Germany neweuropeanwindatlas reanalysis open-data python

I'm interested in windspeeds at 175m, the turbine height of many new wind turbines in Germany, Two sources of windspeed data in EUrope are

1. https://opendata.DWD.de/.../COSMO_REA6: hourly, 1995 to 2018, 6 x 6 km grid
2. NEWA, the New European Wind Atlas, 3 x 3 km grid.

What are good places to compare the two ? There are 8000-odd new windmills in Germany, running since 2024 or planned, under Marktstammdatenregister Gesamtdatenexport_20260101. Comparing at these sites is more interesting than looking everywhere -- importance weighting.

Plot NEWA vs. COSMO-REA6

So what ?

Experts will probably say "of course, dummy, everybody knows that". Is there a 0-order explanation ? Spotchecks of measured windspeeds at ~ 175m would be useful.

Notes, TL;DR

If anyone's interested, I can supply .csv file with NEWA and COSMO-REA7 windspeedslike this:

Newawind175 Cosmowind175 Cosmocap175 State     Lat     Lon  AGS                  Lkr     Place
       9.73        10.58        67.6    SH 54.4777 8.42394 1054        Nordfriesland Galmsbüll
	   ...
       5.02         3.57        13.2    BW 48.0106 7.90246 8311 Freiburg_im_Breisgau  Freiburg

Here Newawind175 is just the average of NEWA average windspeeds at 150m and 200m.

Wind data is extracted with homemade Python code (which is not to everyone's taste). I found it useful to

• subset big .nc files early (numpy .npz is good)
• map sites ──> the nearest Cosmo and NEWA grid points with KDtree
• little functions with print statements and type / shape annotations
• avoid software towers:

Comments welcome

cheers
— denis-bz-py t-online.de 2026-02-20 Feb

_{The purpose of computing is insight, not numbers.}

newa-wind.sh

#!/bin/bash
# newa-wind.sh [area=BY height=100-200]
#	download a bbox from neweuropeanwindatlas area= height= scale=
# or go to https://wps.neweuropeanwindatlas.eu/api/mesoscale-atlas/v1/docs
#	get-data-bbox, click "try it", fill in S N W E ..., execute, wait, click download
# 23jan

trap exit INT QUIT TERM
type bold > /dev/null 2> /dev/null  ||
	alias bold=echo


	# default args --
area=DE  # -> S N W S in the table below
height=100-150-200  # meso 50 75 100 150 200 250 500  micro 50 100 200
scale=meso  # meso 3 km / micro 50m

case $1 in
*=* )	export "$@" ;;  # area= etc.
?* )	exec bold "? $0  area=$area  height=$height  scale=$scale"  # -help
esac

set -o errexit -o nounset

#...............................................................................
	# add your area S N W E here --
case $area in
DE )	S=47.5  N=55  W=5.9 E=15.1 ;;
BY )  	S=47.5  N=50.5  W=9  E=13.5 ;;  # (113, 113)
BYne )	S=49.7  N=49.9  W=12.35  E=12.55 ;;  # max ws
* )		exec bold "Error: add your area to $0"
esac

case $scale in
meso )	var=wind_speed_mean ;;
micro )	var=wind_speed
		bold "micro: check for 0 in values"
		;;
* )		exec bold "Error: $0 scale=$scale"
esac

H="` echo $height |
	sed 's - \&height= g' `"  # 100-200 -> 100&height=200

#...............................................................................
out=$area-newa-$scale-${height}m.nc
bold "curl > $out"

	# download
	# curl progress on terminal: stays 0
gnutime \
\curl --output $out  --remote-time  --show-error \
	-H 'accept: application/x-netcdf' \
"https://wps.neweuropeanwindatlas.eu/api/${scale}scale-atlas/v1/get-data-bbox?\
southBoundLatitude=$S&\
northBoundLatitude=$N&\
westBoundLongitude=$W&\
eastBoundLongitude=$E&\
height=$H&\
variable=$var"

bold `file $out`  # "data" or ASCIi with eg "Bad Gateway"
# ncinfo $out