Cloud Super-cooled Liquid Water Estimation From Satellite Data

John K. Roskovensky, Ivey Ivey, William Porch, Nick Beavis, Ryan Hermann, Sigmund Silber


An automated algorithm for estimating the potential cloud super-cooled liquid water (SLW) from satellite data was developed to perform cloud surveys for assessing potential precipitation enhancement over the state of New Mexico. The algorithm produces spatial cloud SLW column distributions by utilizing many of the Moderate Resolution Imaging Spectroradiometer (MODIS) derived cloud products in addition to parameterizations developed to estimate vertical cloud thickness and fractional cloud liquid water content. Vertically derived cloud SLW is integrated to produce column totals. Algorithm uncertainties are assumed to be large, within a factor of two over the state of New Mexico, due to parameter sensitivity and high MODIS cloud water path uncertainties.Sensitivity studeies identified differences up to 30% in SLW content from individual changes in the vertical cloud thickness estimation, moist adiabatic lapse rate, and cloud liquid fraction parameterization constants. Corrections of MODIS cloud water path artifacts also reduced the estimated SLW content by 10% to 50%. Comparisons of algorithm derived properties were made with other remotely sensed and in situ data. Algorithm cloud thicknesses were lower than co-registered CloudSat radar derived values by about 43% which likely led to an underestimation of liquid water content by more than 50%. Comparisons of the algorithm's vertical estimates of cloud SLW concentration to aircraft data during the Indirect and Semi-Direct Aerosol Campaign (ISDAC) in northern Alaska yielded consistent results. Ground based measurements from the Atmospheric Radiation Measurement North Slope Alaska site for seven days during ISDAC showed that the mean algorithm derived cloud base heights were nearly 1 km too high, but their adjusted temperatures were consistent with atmospheric sounding data. Algorithm derived SLW estimates were consistently within their uncertainty bounds in comparison with other measurements. Annual datasets from extreme dry and wet years over the state of New Mexico were examined to identify regions where cloud SLW appears prevalent and cloud seeding may be fruitful. Cloud SLW coincided well with high terrain especially in the Sangre de Cristo Mountains in the northern part of the state. Further analysis, focused on this specific region, found cloud SLW during the wet year to be 96% greater than in the dry year despite moderate (less than 20%) increases in all other cloud properties including cloud coverage, total cloud water, and cloud thickness.  National Center for Environmental Prediction reanalysis data were sampled during these periods as well and during the wet year the number of days in which the 500 mb level wind was predominantly westerly increased by 12% along with an 8% increase in low-level humidity. Most of the annual difference occurred in the winter months. The parameters with the highest correlation to cloud SLW were the mid-tropospheric 700 mb to 500 mb level relative humidity. This analysis indicates that regions of potential cloud SLW may be very sensitive to subtle changes in environmental and cloud conditions.

Full Text:



  • There are currently no refbacks.