A New Shallow Cumulus Cloud Detection Algorithm from Satellite and Ground-Based Data

Project Overview

Accurate simulations of boundary-layer cloud processes remain one of the outstanding challenges in earth-system modeling. Process-level understanding using observations is essential to validate and improve model simulations of such processes. This feasibility study is to test if a new satellite detection algorithm of continental shallow cumulus (ShCu) clouds developed using ground-based observations at the Southern Great Plains (SGP) is applicable to another region (Southeastern United States, SEUS) with more prevailing ShCu populations. Following the algorithm established at the SGP, we firstly generate surface reflectance maps at the SEUS using geostationary operational environmental satellite (GOES) reflectance data and identify a ShCu cloudy pixel when the GOES reflectance exceeds the clear-sky surface reflectance by a fixed reflectance-detection threshold of ShCu, 0.045. The cloud fractions (CFs) for ShCu cases from this satellite detection algorithm are then compared against the CFs by the ground-based ceilometer. Multiple factors may contribute to the discrepancy between these two CFs, which include 1) the different observed areas due to the point view of ceilometer and the spatial coverage of satellite pixels, 2) the movement of clouds because of wind, 3) the satellite parallax issue due to satellite slanted viewing direction towards the SEUS, and 4) the systemic difference between two instruments for CF validation, e.g., ceilometer used at the SEUS and stereo cameras used at the SGP. We provide solutions to reduce the uncertainties related to each of the factors step by step to facilitate an apple-to-apple comparison and validation. After such careful considerations of uncertainties, we conclude that it is feasible to apply the detection algorithm of continental ShCu clouds to the new region of the SEUS with the same detection threshold of 0.045. Through this algorithm, hourly CFs at the SEUS can be faithfully reproduced from satellite-observed reflectance data in comparison with the ground-based observations. A continental ShCu-clouds dataset produced using this algorithm can be greatly useful for cloud-morphology studies, which will help in evaluating and improving parameterizations of subgrid-scale cloud processes in climate models (e.g., DOE's Energy Exascale Earth System Model, E3SM). This dataset will greatly benefit studies of the land-surface impact on convection, which is one of the science foci of the LLNL Atmospheric System Research (ASR) project.

Mission Impact

From this feasibility study, a unique observational dataset for ShCu clouds can be generated based on our satellite detection algorithm. It can be used to help validate and improve DOE's E3SM performance in its simulating cloud radiative effects, which aligns well with the core competencies in the LLNL climate program.

This feasibility study extends our research areas to the SEUS and prepares us for ASR funding opportunities. Using our algorithm at the ARM SGP site, we studied the impact of land cover on ShCu cloud formation. The new algorithm can be applied to the SEUS to study land-atmosphere interactions, which is one of the science foci of the LLNL DOE ASR project.

This feasibility study prepares us for a DOE ARM mobile facility (AMF3) deployment at the SEUS. The AMF3 deployment will start in early 2023 and last 5-10 years. This deployment promotes many emerging science foci (e.g., convection transition, land-atmosphere interactions, and aerosol chemical processes) from the DOE's ARM, ASR, and terrestrial-ecosystem science (TES) programs. Our algorithm can be used at the SEUS over a large domain to generate cloud-occurrence probability maps, for example, to indicate the preferred locations to put instruments (e.g., ceilometers, stereo cameras, radars, and lidars) to observe more golden convection cases, and further influence the instrument site selection during the upcoming AMF3 deployment during intensive observational periods.

This ShCu detection algorithm will enable new capabilities to study an emerging competitive topic, convection transition, at the SEUS. Existing tracking algorithms for deep convection miss capabilities to detect the full life cycle of ShCu over vast regions and trace the deep precipitating clouds back to their seeds (ShCu). This feasibility study may enable us to compete for DOE BER new-proposal calls.

Publications, Presentations, and Patents

Tian, J. et al. 2022. "How Does Land Cover and Its Heterogeneity Length Scales Affect the Formation of Summertime Shallow Cumulus Clouds in Observations from the US Southern Great Plains?" Geophysical Research Letters 49 (7): 1-12 (2022); doi.org/10.1029/2021gl097070.

Tian, J. et al. 2021. "Summertime Continental Shallow Cumulus Cloud Detection Using GOES-16 Satellite and Ground-Based Stereo Cameras at the DOE ARM Southern Great Plains Site." Remote Sensing 13, no. 12: 2309 (2021); doi.org/10.3390/rs13122309.