Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name



Earth & Environmental Sciences


Reza Khanbilvardi

Subject Categories

Remote Sensing


aerosol; GOES satellite; particle radius; UHI; Urban Heat Island; urban; warm cloud


This study uses ten years of Geostationary Operational Environmental Satellite (GOES) data (1999-2009) to assess urban effects of anthropogenic aerosols and Urban Heat Island (UHI) on cloud-top radiative properties of optically thick clouds. GOES images at noon local time of channels 1, 2, and 4 are used to calculate visible albedo, and shortwave infrared reflectance at 3.9-µm. Cloud-top particle radius is inversely related with SWIRR. Albedo and cloud fraction are measured for cold and warm clouds, and SWIRR is used as a particle-radius proxy for warm clouds only. AERONET fine-mode particle retrievals (fAOD) from the CCNY station are used for the aerosol measure from 2002-2008. A daytime UHI is calculated using ground weather stations in and around New York City. Days are divided into high, medium, and low Aerosol, and UHI. The urban location at Central Park, in NYC is compared with two rural control locations 130 km to the north and south. Climatologies one hour downwind of these three starting points are created using ground station wind data for areas of 5, 10, 20, and 35-km radius.

In overall aerosol results, warm cloud SWIRR is variable but decreases with increasing aerosol. The decrease is significantly greater for areas downwind of the rural locations compared with urban (NYC). This urban-rural difference increases significantly with increasing fAOD for all seasons but winter. Cloud fraction for warm clouds varies inversely with SWIRR for seasons other than winter. Urban-rural differences in cloud fraction are increasingly negative with increasing urban fAOD, suggesting an urban aerosol effect of cloud dissipation. This could be explained by cloud sedimentation and evaporation-entrainment effects, or a semi-direct effect. In a visual inspection of a large number of mapped climatologies, SWIRR values of pixels are inversely correlated with albedo.

Summer season warm clouds have significant positive urban-rural SWIRR differences at all aerosol levels, and differences increase with higher fAOD and at smaller downwind-area radii. Spring, summer and fall SWIRR differences increase with fAOD, and have an inverse relationship with cloud fraction urban-rural differences, with the only exception of the spring high aerosol category. Cold-cloud albedo in the summer season increases with increasing CCNY aerosol level, with the rural values becoming increasingly higher than urban.

In UHI outcomes, lower urban-rural albedo and cloud fraction with higher UHI is expected, and is found in larger downwind areas. Urban cloud enhancement may be present in cases of deviations, or in trends toward a deviation from this pattern in data drawn more exclusively from the urban and rural starting points. High UHI urban-rural differences are higher than expected at the smaller downwind areas for summer and fall albedo, and for winter and fall cloud fractoin. Higher UHI days have higher warm-cloud SWIRR in both urban and rural areas in spring and fall. SWIRR urban-rural differences are high for all of the summer season, and increase with UHI in spring and fall. An urban effect may be at play in winter cloud fraction outcomes, in which the urban-rural difference is consistently positive. Urban cloud fraction is generally lower downwind of the NYC location for other seasons.

Low, optically thick clouds are fewer, and/or are fewer downwind of more narrowly defined study areas in urban compared with rural locations across both datasets for seasons other than winter. These clouds have a negative radiative forcing; that is, their net effect is cooling, with implications for the global radiative balance. Short-term warming can be felt directly by urban and suburban residents. Refinements in predicted cloud cover is important to urban planners both regarding peak energy loads and water resources. Comparisons of these outcomes with climate prediction models could be useful in their incorporation of urban effects.