Meytar Sorek-Hamer, David Broday.
In Year 4 we performed an analysis of an updated version of MAIAC AOD at high resolution (1x1km) over Israel for the years 2002-2013, and continued to develop the mixed-effects model for predicting PM2.5 and PM10 using various AOD data. In general, the model revealed better results for the PM10 prediction (R2=0.8) than for PM2.5 prediction (R2=0.66). Furthermore, we used additional LU and met. variables (e.g. population density, distance from roads, PBL, RH) that slightly improved the model performance. As the initial descriptive analysis demonstrated some problems with the data, especially in relation to spatiotemporal availability over the study period, and as we found that in certain PM stations there is a good fit while at other stations there is no fit at all, we reported these problems to the MAIAC (NASA) team and are about to receive a new version of the satellite AOD product soon, to finalize our study. Moreover, we received the new collection 6 MODIS data over Israel and developed a PM prediction model based on the Deep Blue products for Israel and SJV, CA (USA). A manuscript describing this work has been submitted to Remote Science of Environment