[Updated version]

This is an updated version from 2025-09-05 of the original data set that has been published on 2025-07-25.

• Data are now in cleaned and consolidated text file format. Before it was a very raw XML format with lots of junk text lines, also data from other sensors and with unclear time referencing (several time stamps could be present in a file without transparent information which time was relevant for the LNM data).
• Documentation has been added 

Introduction

As part of its synoptic meteorological weather data measuring network (automated SYNOP network of ground stations), the German Meteorological Service continuously operates about 150 laser-optical distrometers, the Laser Niederschlags Messer (LNM) from manufacturer Thies.

This device registers the size and vertical fallspeed component of each particle (raindrops, graupel, snowflakes, hail, or anything else) that passes trough a horizontally oriented laser light sheet of length 22.8 cm, with 2.0 cm and height 0.75 mm. 

From all particles registered during 1 minute, a 2-dimensional discrete distribution (numbers in 20x22 fallspeed-diameter-classes) is constructed and several precipitation characterizing integral parameters are derived, such as precipitation rate, estimated type of precipitation according to WMO ww tables, estimated equivalent radar reflectivity factor, particle size distribution, etc.

Relevant for the operational SYNOP data of DWD is only the estimated type of precipitation. However, the 2-D fallspeed-diameter distributions for each minute ("raw data") are also recorded, but not put into DWD's official longterm data bases.

Because these raw data may be highly relevant and interesting for the international scientific community in the fields of cloud physics, numerical weather prediction and radar meteorology, we provide via Zenodo a comparatively large raw data set for the years 2019 to 2024 of >= 150 stations throughout Germany.

The data set is accompagnied by additional parameters which are relevant for quality control and cross checks, namely 10-minute values of temperature, relative humidity, accumulated precipitation from rain gauges, cloud cover, wind speed and wind direction. All relevant station meta data are provided as well.

Documentation

Below, we provide the most important informations about the data set. It is accompagnied by some PDF documents in the compressed tar archive Documentation.tgz :

• Technical manual specifically for the DWD devices (in German language only):
  • LNM_manual_Thies_deu_kurz.pdf (short manual)
  • LNM_manual_Thies_deu.pdf      (detailed manual)
• English translation/description of the data format and the diameter-velocity class limits:
  • LNM_telegram_engl.pdf
  • LNM_classes.pdf
• English manual from the manufacturer website (see below) not specific for DWD devices:
  • 5.4110.xx.x00_laser_precipitation_monitor_eng_short.pdf (short manual)
  • 5.4110.xx.x00_laser_precipitation_monitor_eng.pdf        (detailed manual)
• README file

Instrument description

Short and detailed descriptions of the specific DWD-version of the LNM are given in German language, see above.

English manuals are not available for the DWD-specific version, but have been downloaded from the manufacturer website for a newer instrument version and are also provided in Documentation.tgz. Links to the manufacturer's website can be found below.

Data files and formats

Distrometer data 

• Yearly gzipped tar archives named lnm_thies_DWD_20XX.tgz
• One data file per day and per station in it's own directory. Directory structure in each archive:
  • ./<yyyy>/<mm>/<dd>/<wmoid>/<data-file> (yyyy=year, mm=month, dd=day, wmoid=5-digit WMO station ID)
• File name <V>_<wmoid>_<yyyy>-<mm>-<dd>.txt (V=1-digit instrument version, either "2" or "3"), e.g. 3_10474_2023-07-01.txt
• One line per one-minute measuring interval (nominally 1440 lines per daily file)
• Semicolon-separated format, many records per line
• Detailed documentation of each record in LNM_telegram_engl.pdf
• Definition of the diameter- and fall velocity classes in LNM_classes.pdf
• The time stamp in each line is the approximate end of the measuring minute in UTC. Due to technical issues with time synchronisation, there is an uncertainty of about 30 seconds. Therefore, we have rounded the time to the nearest full minute.
• Sometimes 1 to 5 minutes in a daily file are missing due to spurious data stream transmission problems which are not recoverable. Therefore not every file contains the full 1440 lines.
• Files from days where stations have only partly been operational have been removed, e.g., when a certain station has been shut down or installed during that day, or when there was a power-outage or other disruption of the station operations.

Additional synoptic data

• Yearly gzipped tar archives named synop10min_DWD_20XX.tgz
• One data file per day and per station in it's own directory. Directory structure in each archive:
  • ./<yyyy>/<mm>/<dd>/<wmoid>/<data-file> (yyyy=year, mm=month, dd=day, wmoid=5-digit WMO station ID)
• File name synop10min_<wmoid>_<yyyymmddhhmmss>_1.0days.txt, e.g. synop10min_10747_20200101000000_1.0days.dat
• 6 heaer lines preceded by "#"
• One line for each 10-minute interval
• Space-separated format in 7 columns
• Parameters in each column defined in header line No. 6:
       # end-time     T2M[°C]   RH2M[%]   P10min[mm]   CLC[oct]  FF[m/s]  DD[°]
  (end time of 10-min interval, 2-m temperature, 2-m relative humidity, accumulated precipitation from rain gauges, total cloud cover, wind speed and wind direction)

Station metadata

See the more or less self-explanatory text file station_metadata_LNM_20250713_utf8.txt.

• This file gives for each WMO-ID the name, geographic longitude, latitude, station height AMSL and the time span of its existence (begin time and end time).
• There are 181 stations altogether given in the file.
• However, not all the stations have been operational all the time. Some stations have been closed, moved, or newly installed during the 5 years, so that less stations were active at each particular time (~150).
• There might be more than one line per station in the metadata file. In this case the station has been moved to different location(s) and the time interval of each location is given in the last two columns (FROM_DATE and TO_DATE). 

Basic software

Todo: add python code snippets to read the data files and compute precipitation rate, particle size distributions and equivalent radar reflectivity factor.

Further informations about the instrument