NAVO MCSST
mcsst_1.ncl: NAVO MCSST data is in
byte format, but fortunately, the NCL built-in function
fbindirread can handle this type. It is a simple process to expand
this to float data and apply the constants that transform the data into
actual SST's.
sst = new((/1024,2048/),"float")
sst = bytedata*xslope+yint: will convert the byte data to a float and calculate the true SST values.
See script for how lat/lon coordinate variables are created.
mcsst_2.ncl: Same script as example
1. Just demonstrates the non-interpolated data. I found that in areas
with sparse data (e.g. the Caribbean), the interpolated values were
greatly different from the observed values. If you plan to use this
data for feature identification, verify your results with the observed
values.
mcsst_3.ncl: Demonstrates zooming in
on a subregion. We can turn off raster mode now, since the area is
smaller.
For greater detail, we have changed to a higher resolution coastline by setting: mpDataBaseVersion = "Ncarg4_1".
When you plot a sub region, you need to turn off the cyclic point which is automatically added to each plot. This is done by setting: gsnAddCyclic = False
To zoom in on a region, you need to set the following resources:
mpMinLonF
mpMinLatF
mpMaxLonF
mpMaxLatF
mcsst_4.ncl: Used a Fortran shared
object (composit.f) to create a
composite image of a day and night pass. Because
of the grid, a double loop of this size is extremely slow in an
interpreted computer language like NCL. (IDL and MatLab would have
similar issues). The external routine is simplistic, but significantly
increases speed. You will want to use
WRAPIT to compile the shared object.