Atmospheric Radiation Measurement Climate Research Facility US Department of Energy

cldtype > Cloud Type ClassificationVAP Type(s) > Baseline • Evaluation

The Cloud Type (CLDTYPE) value-added product (VAP) provides an automated cloud type classification based on macrophysical quantities derived from vertically pointing lidar and radar. Up to 10 layers of clouds are classified into seven cloud types based on predetermined and site-specific thresholds of cloud top, base, and thickness. Inputs for the CLDTYPE VAP include lidar and radar cloud boundaries obtained from the Active Remotely Sensed Cloud Location (ARSCL) and surface meteorological systems (MET) data. Rain rates from MET are used to determine when radar signal attenuation precludes accurate cloud detection. Temporal resolution and vertical resolution for CLDTYPE are 1 minute and 30 meters respectively and match the resolution of ARSCL. The CLDTYPE classification is an initial step for further categorization of clouds. It was developed for use by the Shallow Cumulus VAP to identify potential periods of interest for the LASSO model and is intended to find clouds of interest for a variety of users.

The primary input for cloud classification in the CLDTYPE VAP is cloud boundaries from the ARSCL data product. ARSCL merges lidar and radar cloud tops and bases to produce a single composite of clouds, a cloud mask. The combination of lidar and radar retrievals provides complementary capabilities. While the lidar detects low- and most mid/high-level clouds, it can be limited by strong optical attenuation if a cloud layer has a high concentration of hydrometeors. Radar, on the other hand, can be effective at detecting mid/high clouds through the low lidar-attenuating layers but fails to detect cloud layers with small particles to which the lidar is more sensitive. To insure all clouds are included, if lidar cloud boundaries at 3.5 kilometers or below are available but not found in the ARSCL cloud mask, the boundaries are added in a preliminary step in the CLDTYPE VAP.

Cloud top, cloud base, and thickness of the cloud layers are calculated from the ARSCL data product. Prior to cloud classification, cloud layers are screened by a minimum required thickness and a minimum separation between layers. In the first step, cloud layer with thickness less than or equal to 120 meters are removed. In the second step, adjacent clouds layers that are separated by 120 meters or less are merged into a single cloud layer. These filtering steps are included to reduce noise. Each identified layer (up to 10 layers) is then assigned one of seven cloud types (1. Low clouds, 2. Congestus, 3. Deep Convection, 4. Altocumulus, 5. Altostratus, 6. Cirrostratus/Anvil, 7. Cirrus) on the basis of the top height, base height, and layer thickness.

The CLDTYPE VAP currently runs at ARM’s Southern Great Plains (SGP) Central Facility (C1). In the future, we may also run the CLDTYPE VAP at the North Slope of Alaska (NSA) Barrow site (C1), and ARM Mobile Facility (AMF) sites where ARSCL data are available; however, work may be required to define meaningful cloud boundary thresholds for cloud types at these sites.

Primary Derived Measurements


  • Fixed
  • AMF1
  • AMF2
  • AMF3


Riihimaki L, X Li, Z Hou, and L Berg. 2021. "Improving prediction of surface solar irradiance variability by integrating observed cloud characteristics and machine learning." Solar Energy, 225, 10.1016/j.solener.2021.07.047.


Goss HB, KS Dorsey, CB Ireland, MR Wasem, RA Stafford, and R Jundt. 2020. 2019 Atmospheric Radiation Measurement (ARM) Annual Report. Ed. by Kathryn Dorsey, ARM user facility. DOE/SC-ARM-19-032. 10.2172/1604869.


Yang W, A Marshak, and G Wen. 2019. "Cloud edge properties measured by the ARM shortwave spectrometer over ocean and land." Journal of Geophysical Research: Atmospheres, 124(15), doi: 10.1029/2019JD030622.
Research Highlight


Flynn D, Y Shi, K Lim, and L Riihimaki. 2017. Cloud Type Classification (cldtype) Value-Added Product. Ed. by Robert Stafford, ARM Research Facility. DOE/SC-ARM-TR-200. 10.2172/1377405.

View All Related Publications


Damao Zhang
Pacific Northwest National Laboratory

View All VAP Translators