Reduction in cloud cover after correcting finite resolution measurement biases

Published: 12 November 2020| Version 1 | DOI: 10.17632/k36f6tsvp5.1
Soumi Dutta, Larry Di Girolamo, Sagnik Dey, Yizhe Zhan, Catherine M. Moroney, Guangyu Zhao


The following datasets are used for the research entitled 'Reduction in cloud cover after correcting finite resolution measurement biases'. It's generally accepted that around 70% of the Earth's surface remain covered by clouds. We have found from this study that near-global (50N-50S) coverage of total cloud fraction is 47%, using resolution-corrected MISR cloud fraction data. To assess this result, view-angle corrected MODIS data (within MISR swath) has been used. Also, pixel-wise collocation of ASTER-MISR and MODIS-MISR have been done to get robust evaluation of the corrected MISR data. ASTER radiance values which is available at 15m resolution is used to calculate cloud fraction which is treated as 'true' data for the assessment process. Also, MODIS full-swath cloud fraction data (from GEWEX) has been analyzed to understand the differences after view-angle correction applied to MODIS. MISR resolution-corrected new cloud fraction dataset falls within ±5-8% of the benchmark ASTER data. Thus it provides improved estimates of near-global cloud coverage which is a step forward towards correct estimation of cloud feedback process. Data description: Figure 1: Data files for top panel (PRCCF, SECF, MOD35 within MISR swath) and bottom panel (PRCCF-SECF, PRCCF-MOD35 within MISR swath, SECF-MOD35 within MISR swath) of figure 1 are stored inside folder named ‘Figure 1’. Figure 2: Bar plot values for six regions are provided in file ‘bar plot values.doc’ inside the folder named ‘Figure 2’. Figure S1: The file ‘radiance.npy’ inside the folder ‘Figure S1’ contains radiance data for an ASTER image in DN (Digital Number) values. Figure S1(a) which is the radiance image, is plotted using those DN values. In case of Figure S1(b), a threshold (DN=31) has been applied to DN values contained in radiance.npy so that, DN<31 values are considered as clear pixels and above 31, all are cloudy pixels. Basially Figure S1(b) is the cloud mask. Figure S1(c) shows the histogram plot of the radiance values. Figure S2: Figure S2(a) shows seasonal CF climatology derived from MODIS-ST product of GEWEX database. Files for plotting figure S2(a) are inside folder S2(a). Figure S2(b) shows difference between MODIS-ST (full swath) and MOD35 (within MISR swath) for the period 2001-2009. Files for plotting figure S2(b) are inside folder S2(b).



University of Illinois at Urbana-Champaign, NASA Jet Propulsion Laboratory, Indian Institute of Technology Delhi


Atmospheric Science, Satellite Remote Sensing, Climate Data, Passive Remote Sensing