

LOS ANGELES
LAWN REMOVAL
PROGRAM
DISCUSSION
NDVI
The results of the NDVIs for 2013-2016 were not as we expected. Instead of vegetation going down due to lawns being taken out and replaced with non-plants, the NDVI showed that vegetation went up over our study area as a whole. The change detection map for NDVI displays this increase in vegetative cover across our study area (displayed in red). However, the change detection map also shows a slight decrease over some of the 975 sample parcels (displayed in blue), reflecting decreased vegetation coverage presumably due to turf replacement. There was an increase of rain with 2013-2014 receiving 6.08 inches, 2014-2015 receiving 8.52 inches and 2015-2016 receiving 9.65 inches of rain (LA Almanac). Additionally, NDVI is heavily affected by trees and canopy coverage. Los Angeles initiated The Million Trees program to plant a million trees in LA. The trees planted each year could have easily affected our NDVI because the spatial resolution in Landsats 7 and 8 are so large. Based on the data we have, many residents replaced their grass with trees and bushes as well, which could have increased the overall NDVI. From out research, it does not appear that the city of Los Angeles required a specific type of replacement. In addition to 2016’s NDVI using Landsat 8 instead of Landsat 7 like 2013-2015, the image may also look different because of the Rey and Blue Cut fires around Los Angeles.
Thermal
Similarly, the results of the thermal data for 2013-2016 were not what we anticipated. We hypothesized that due that replacement of lawns with materials other than plants, our study area would reflect a steady increase in thermal brightness temperature. The change detection map in our results section shows that brightness temperature changed positively (marked by the points in red) over many of the 975 sample parcels, which is consistent with our hypothesis there would be an increase in temperature over the lawns that underwent replacement. The observed difference from 2013 to 2016 was an increase in the average brightness temperature of 18.2. Though seemingly in line with our hypothesis, the wide range of values between 2013/2014 and 2015/2016 may be due to the fact that Landsat 7 Enhanced Thematic Mapper (ETM+) data is collected at 60-meter spatial resolution and then arbitrarily resampled to 30-meter pixels, which is still far larger than most lawns within Los Angeles. Another explanation could be temperature variability across the observed days each year: the 2013 and 2016 images may simply have been captured on cooler days than the days captured for the images in 2014 and 2015.
Further Research
If this project were to be continued, it is important to be able to use more satellites that could produce higher quality and higher spatial resolution images. Within the scope of our project, we were unable to acquire better images than the ones Landsat 7 and 8 provided. If images with higher spatial resolution could be used, our data would be more accurate because the average size of a lawn is small. Additionally, the thermal analysis conducted uses resampled data in which there are an unknown amount of estimates involved.