Short movies of GOES AOD data over a set of regional domains
Static images of GOES AOD showing an aggregate over the day
We have the choice of displaying AOD values with SpatialPoints or(and?) Rasters. Both have strengths and weaknesses:
SpatialPoints
PROS:
Precise: Each point describes a reading at one specific location.
No need to calculate spatially aggregated readings.
CONS:
Different bboxes can make points overlap or spread out since their size isn't related to geography. The user will have to test whether whether points are too small to see, or so large that they cover each other.
Points might not be fixed to a static location, which makes temporal aggregation much more difficult. I think local_jon/local_R/spatial_aggregation.R describes a case study of this. Looks like it just describes sample location inconsistencies between G16 and G17 scans, which is forgivable. It also looks like sample locations are consistent between scans, so I guess this isn't a con.
Rasters
PROS:
Raster cells are directly related to geography, so different bboxes will not cause cells to overlap or separate.
Grid cells are static(?), which makes temporal aggregation much easier.
CONS
Imprecise: cells summarize many readings in a region into a single value. Outliers are averaged out and and it's possible for a small number of readings to generalize an entire region.
Takes time to calculate spatially aggregated readings. Higher grid resolution -> exponentially more time to compute.
Low resolution rasters are imprecise and ugly. High resolution rasters can be as precise as SpatialPlots, but take forever to compute/render.
What to do...
Misc:
[X] Check to see if it's possible for goesaodc_plotSpatialPoints() to aggregate multiple scans by reading up on the local_jon/local_R/spatial_aggregation.R problem. Do sample point coords remain consistent over multiple scans?
[X] Existing visualization functions and examples use base plot(), raster::plot(), and, rasterVis::levelPlot(). Ask if the package should transition to ggplot or stick with what it currenty uses. Okay to use AirFirePlots for plotting rasters. Goal is to adopt ggplot for most other plots, but can use base plot for now if it's easier and faster.
High level functions:
[x] goesaodc_plotScanPoints(): A function that draws a points-plot of a satellite scan(s) specified by either sat+time info or a file. If multiple scans
[x] goesaodc_plotScanRaster() A function that draws a raster plot of a satellite scan(s) specified by either sat+time info, or a file.
[x] goesaodc_animateScanPoints: Executable script that generates a video from scan point plots.
[x] goesaodc_animateScanRaster: Executable script that generates a video from scan raster plots.
Vignettes
[x] A vignette that demonstrates how to use package plotting functions using example fires.
The goal for V0.4 is to provide:
We have the choice of displaying AOD values with SpatialPoints or(and?) Rasters. Both have strengths and weaknesses:
SpatialPoints
PROS:
CONS:
local_jon/local_R/spatial_aggregation.R
describes a case study of this. Looks like it just describes sample location inconsistencies between G16 and G17 scans, which is forgivable. It also looks like sample locations are consistent between scans, so I guess this isn't a con.Rasters
PROS:
CONS
What to do...
Misc:
goesaodc_plotSpatialPoints()
to aggregate multiple scans by reading up on thelocal_jon/local_R/spatial_aggregation.R
problem. Do sample point coords remain consistent over multiple scans?plot()
,raster::plot()
, and,rasterVis::levelPlot()
. Ask if the package should transition to ggplot or stick with what it currenty uses. Okay to use AirFirePlots for plotting rasters. Goal is to adopt ggplot for most other plots, but can use base plot for now if it's easier and faster.High level functions:
goesaodc_plotScanPoints()
: A function that draws a points-plot of a satellite scan(s) specified by either sat+time info or a file. If multiple scansgoesaodc_plotScanRaster()
A function that draws a raster plot of a satellite scan(s) specified by either sat+time info, or a file.goesaodc_animateScanPoints
: Executable script that generates a video from scan point plots.goesaodc_animateScanRaster
: Executable script that generates a video from scan raster plots.Vignettes