Access

View meta data:

• Birkholz Level1 CSM01 August
• Birkholz Level1 CSM03 May-July
• Birkholz Level2 DWD CSM01 August
• Birkholz Level2 DWD CSM03 May-July
• Birkholz Level2 UZK CSM01 August
• Birkholz Level2 UZK CSM03 May-July
• Falkenberg Level1 CSM01 August
• Falkenberg Level1 CSM02 May - start of August
• Falkenberg Level2 DWD CSM01 August
• Falkenberg Level2 DWD CSM02 May - start of August
• Falkenberg Level2 UZK CSM01 August
• Falkenberg Level2 UZK CSM02 May - start of August
• Lindenberg Level1 CSM00 August
• Lindenberg Level1 CSM01 June-July
• Lindenberg Level2 DWD CSM00 August
• Lindenberg Level2 DWD CSM01 June-July
• Lindenberg Level2 UZK CSM00 August
• Lindenberg Level2 UZK CSM01 June-July

Download data:

• Get single Birkholz data files via HTTP/OPeNDAP | Get all Birkholz data files via wget script
• Get single Falkenberg data files via HTTP/OPeNDAP | Get all Falkenberg data files via wget script
• Get singel Lindenberg data files via HTTP/OPeNDAP | Get all Lindenberg data files via wget script

Description

The data set contains two wind products: (i) vertical profiles of the mean wind vector and (ii) vertical profiles of wind gusts. The data set is based on Doppler Lidar (DL) measurements and it is available for three different supersites (Falkenberg, Lindenberg, Birkholz) operated during the Field Experiment on Sub-mesoscale Spatio-Temporal Variability in Lindenberg (FESSTVaL) covering the period May 17, 2021, to August 31, 2021. For quality assurance purposes, an intercomparison experiment of the eight DL systems participating in FESSTVaL was organized in Falkenberg between July 16 and August 09, 2021. This implies that no DL wind_and_gust products are available for the Lindenberg and Birkholz sites during this period. The comparison experiment in Falkenberg included operation of DLs in the system configuration applied to determine gusts (see below) – thus for this site the wind_and_gust product is available also for the period July 29 to August 09, 2021. Note that two additional wind and gust products are available for which a separate description is given: (a) profiles based on ultrasonic anemometer measurements at the 99m tower at the Grenzschichtmessfeld (GM) Falkenberg, and (b) profiles from the same lidar measurements but derived using a different retrieval algorithm (level2uzk). DL profiles extend throughout the lower atmospheric boundary layer from 90m up to a maximum height typically above 1000 m dependent on the atmospheric backscatter conditions. The DL measurements were based on a conically Doppler lidar scanning geometry with high temporal resolution (~3.4s for one full scan, azimuth resolution of approx. ~33 deg) and a constant zenith angle of 28 deg. The realization of such a scanning strategy was possible via the continuous scan mode option of the DL system with a number of accumulated pulses per beam Npa = 3000. Two different types of Halo Photonics DL systems were used at the three sites during the campaign: (i) a Halo Photonics Streamline XR with a range gate length of 48 m and Halo Photonics Streamline with a range gate length of 30 m. For the non-XR systems the focus was set to 500 m, for XR systems it is set to infinity per default. For more details concerning the scan configuration see also Steinheuer et al. ( https://doi.org/10.5194/amt-15-3243-2022).

In total there are three DL data subsets provided (Level1, Level2dwd, Level2uzk):

The Level 1 data are provided by DWD using the dl_toolbox (https://github.com/mkay-atm/dl_toolbox) and include both the instantaneous DL measurements and related values (e.g. radial velocity and signal-to-noise ratio as function of range gate, time, and azimuth direction) and relevant information concerning the system’s specific parameters which are either fixed by the manufacturer (e.g. wavelength, pulse repetition frequency, pulse length) or can be configured by the user (e.g. range gate length, number of pulse accumulation, focus).

Level-2 data represent 10-min averages of the derived mean wind vector and of wind gust speeds. Usually, gusts are defined as a 3s moving average (WMO, https://library.wmo.int/doc_num.php?explnum_id=10616). We try to match this by calculating first for each scan (sampling time 3.4 s) a wind estimate and then we search for the maximum value (= gust) within a pre-defined 10min interval. The gust is provided if at least 50% of the individual scans within the 10 minutes have been processed.

- The Level2dwd dataset v01 contains results based on the DWD processing. This processing is based on the data processing described in Päschke and Detring (2024, https://doi.org/10.5194/amt-17-3187-2024). Their publication deals with the analysis of Doppler Lidar TKE measurements. As the data processing has similar challenges (due to the small number of pulses per ray), the data handling for the DL gust mode was adapted to this. The new data et (version v01) is generated with this updated processing. A separate publication is in preparation and will be added to this description, for first information see Detring, C. et al., 2022, https://doi.org/10.5194/ems2022-184).

- The Level2uzk dataset contains results based on the Uni Cologne processing (see Steinheuer et al., https://doi.org/10.5194/amt-15-3243-2022; and see https://github.com/JSteinheuer/DWL_retrieval).

All data are organized in daily files. The original measurements cover the lower 2600m (non-XR systems) / 4200m (XR systems) above ground level. However, depending on the signal quality and the results of the product’s quality assurance, the availability of reliable data in this lower part of the atmospheric boundary layer may be limited to lower heights. Due to configuration changes, system updates or location changes, there are single days for which the data are either missing or not complete. Operational decisions on the measurement strategy during FESSTVaL and the availability of the instruments led to different measurement tasks of the DL systems during the experiment which implies that the data products at each site originate from different systems during sub-periods of the experiment..

Limitations: Due to the short sampling time per ray, regular time synchronization vs. a reference at prescribed intervals occasionally resulted in a jump back of the time stamp assigned to each vector of radial velocity data. We did not correct that since we wanted to keep the original level-1 data as they were provided from the instrument.

- The quality control procedures implemented for the level2dwd product include an assessment of the signal-to-noise ratio of the backscattered lidar signal (snr), various statistical tests to remove outliers (acf, r2), and completeness tests concerning the availability of both single beam data per scan and single-scan wind values per 10-minute interval (n_good_data, n_good_circulation). Each lidar-based value is accompanied by a quality flag (qwind) where 0 = bad, and 1 = good.

- The quality control procedures implemented for the level2uzk product omits an assessment of the signal-to-noise ratio of the backscattered lidar signal (snr), but are instead based on statistical coherence. This involves fitting a wind vector to the radial observations and iteratively eliminating outliers that are inconsistent with the fit. The iteration stops when the fit have small uncertainties (wind is returned) or too many rejected observations (no wind is returned). The amount of included observations and the quality of the fit is combinend to a covariance matrix for the wind vector describing the uncertainty of each estimate. Please check Steinheuer et al., https://doi.org/10.5194/amt-15-3243-2022 for a detailed description.

Reliability of both the derived mean wind and the gust wind speeds has been assessed by comparison with the sonic wind and gust product data at a reference level of 90 m for a several-months data set. Rmsd values are in the order of 0.3 m/s for the mean wind speed, and 0.7 m/s for the maximum gust speed, respectively.

• Instruments: Source: ground based remote sensing, Doppler lidar Streamline / Streamline XR (Halo Photonics Ltd.) :
  sups_rao_dlidCSM00 : S/N #44 (XR)
  sups_rao_dlidCSM01 : S/N #78
  sups_rao_dlidCSM02 : S/N #177
  sups_rao_dlidCSM03 : S/N #172
  fval_kit_dlidCSM01: S/N #161 (XR)
  
• Locations:
  Birkholz: 52.200 degrees North, 14.192 degrees East, 70 meters above mean sea level
  Falkenberg: 52.167 degrees North, 14.123 degrees East, 73 meters above mean sea level
  Lindenberg: 52.209 degrees North, 14.122 degrees East, 115 meters above mean sea level
  
• Vertical resolution:
  Level 2 data: 26.48 m and 42.48 m (XR systems)
  Level 1 data have a range resolution of 30 m (48 m for XR systems) along line-of-sight with a zenith angle of 28 deg)
• Time resolution: Level 1 data: approx. 0.33 s
  Level 2 data: 10 minutes
• File format: NETCDF4
• Time period (see single XML Files for information about the single data sets):br/> Start: 2021-05-18
  End: 2021-08-31
• Standard: SAMD v2.2, https://doi.org/10.25592/uhhfdm.9902
• Project: FESSTVAL

carola.detring (at) dwd.de
Doppler LIDAR DWD processing: frank.beyrich (at) dwd.de
Doppler LIDAR UzK processing: julian.steinheuer (at) uni-koeln.de