Estimating Relative Humidity Using GOES-16 and MODIS Satellites for an Urbanized Area

The primary objective of this study is to estimate the Relative Humidity (RH) in urban areas using remotely sensed satellite measurements. RH is necessary to approximate the Heat Index (HI) used to quantify the impact of extreme heat events and as an early warning indicator. Currently RH is calculated from weather forecast and in-situ measurements, both of which have several inconsistencies. Weather forecasting models employed by National Weather Service do not have any urban parameterizations and their estimates of urban thermal conditions is inaccurate. The in-situ measurements are incapable of representing the spatial variability in temperature and humidity over urban areas. Herein we use the Geostationary Operational Environmental Satellite R-Series (GOES16) from NOAA to estimate hourly RH values in the urban New York City area. GOES16 is fitted with an advanced imaging sensor and has faster coverage that provides the capacity to study environmental phenomena and parameters at near-real time scale. The RH is estimated using a linear regression model with input variables that include land surface temperature, total precipitable water, normalized difference vegetation index, digital elevation model, urban and rural surface properties, and temporal variability. The estimation model is developed by a stepwise function, which selects the significant variables and the corresponding coefficients. The RH approximation at high temporal and spatial resolution are validated with ground station measurements inside and outside the city area. Preliminary, results show that the coefficient of multiple determination is 0.70 and the mean absolute error is 6.86 percent.