In situ aircraft measurements as validation for MODIS cloud products

Clouds influence climate through their ability to reflect radiation. The depth of the cloud and the cloud droplet number concentration determine the efficiency of the reflection. Stratus clouds which cover vast areas of the planet to the west of continents have a high susceptibility to changes in cloud properties due to aerosols and can play a large role in climate change. Satellites can monitor and provide information regarding these clouds and their changes. NASA deploys the large network of field data known as Aerosol Robotic Network (AERONET) for validation of MODIS derived atmospheric aerosol properties; however, a network for validation of derived MODIS cloud properties is nonexistent. The greater optical thickness of clouds limits cloud property validation through surface based observations. However, in situ aircraft measurements during field campaigns provide a platform in which the same properties from the MODIS cloud products can be calculated using the cloud droplet distribution from cloud droplet probes. Here, two stratus cloud field campaigns, Physics Of Stratocumulus Tops (POST) and the MArine Stratus Experiment (MASE), are used as validation for MODIS cloud products including cloud effective radius (r_e) and cloud optical thickness (COT). These campaigns were both off of the central California coast during summertime; MASE was July, 2005; POST was July-August, 2008. Cloud droplet probe measurements from vertical soundings through the cloud were used to calculate r_e and COT. The sounding that was closest in time to the MODIS satellite swath was used for the comparison with the satellite data. The satellite data was averaged for every 1km pixel along the flight path of the sounding. Comparisons of COT showed good correlation coefficients (R) for both MASE (R=0.89) and POST (R=0.93) and were close to the 1:1 line; although MASE had a steeper slope (k=1.21) verses POST (k=1.03). Also, R of r_e of 2.1µm retrievals for POST (R=0.95) were good but MASE (R=0.74) were a little weaker. However, Figure 1a shows that MODIS retrievals of r_e were overestimating the in situ measurements in both projects as the data is above the 1:1 line. MODIS r_e showed less of an overestimate when compared to the maximum in situ r_e, however the correlations were weaker (R=0.63 for MASE and R=0.41 for POST). Figure 1b shows the different MODIS r_e retrievals for POST. The 2.1µm data is the same retrieval as in Figure 1a. The 3.7µm r_e retrieval from POST also overestimates in situ r_e; whereas the 1.6µm r_e retrieval appears to be less of an overestimate however it has a reduced number of available data which includes the largest r_e data point. While MODIS is able to resolve the variability of r_e, as noted from the R, r_e is overestimated in most cases involving these stratus clouds. COT, however, is not such an overestimate and MODIS is also able to resolve the variability of COT. This lack of closure for r_e limits our effectiveness in using satellite data to improve our understanding of cloud processes and their effects on climate.