Alaska and Hawaii

[kgjenkins]

Process Alaska and Hawaii, making the AK and HI rasters available as separate GeoTIFFs (no need to have all those empty ocean pixels), but add the vector polygons to the US48 shapefile, geojson, and topojson files.

[cm325]

Yes! we need these too!

[kgjenkins]

It looks like these JPEGs would be the best source to work from:

http://planthardiness.ars.usda.gov/PHZMWeb/Images/300DPI/SIMP_AK.jpg
http://planthardiness.ars.usda.gov/PHZMWeb/Images/300DPI/SIMP_HI.jpg

They would need to be georeferenced, and then converted to a GeoTiff of zone values, based on the hue of the original JPEG. Then generate polygons for each zone, and save as shp, geojson, and topojson.

My original conversion of the US48 involved many experiments of different methods, and I should have fully documented the final methods used. But I think AK and HI will be a simpler process, since there are no borders or coastline lines in the state-based JPEGs.

@cm325 do you want to try to tackle this?

[cm325]

Hi,

Ok, I started on this. Georeferencing worked just fine in QGIS, but I'm not finding a nice way to automatically turn the rgba values into discrete zones. Do you remember what you used?

Thanks!

[kgjenkins]

@cm325 Thanks for working on this!

For the US48, I extracted the hue values from the original map image pixels (in order to eliminate variations in the pixels caused by the underlying elevation hillshade), and then reclassified those to the low value of the corresponding temperature range. I used Manifold at that time.

Using QGIS, you should be able to get the hue values using the GRASS i.rgb.his function, and then reclassify those values to the temperature range low value using GRASS r.reclass.

[cm325]

Yeah, That's the exact issue I'm running into, and I'm trying to avoid manually identifying every image shade within a given range just to put that in the reclassification. I'll try that rgb function, maybe that's the ticket! Thanks!

…

On Thu, Dec 8, 2016 at 9:33 AM, Keith Jenkins ***@***.***> wrote: @cm325 <https://github.com/cm325> Thanks for working on this! For the US48, I extracted the hue values from the original map image pixels (in order to eliminate variations in the pixels caused by the underlying elevation hillshade), and then reclassified those to the low value of the corresponding temperature range. I used Manifold at that time. Using QGIS, you should be able to get the hue values using the GRASS i.rdb.his function <https://grass.osgeo.org/grass70/manuals/i.rgb.his.html>, and then reclassify those values to the temperature range low value using GRASS r.reclass <https://grass.osgeo.org/grass70/manuals/r.reclass.html>. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#3 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABaeDeXktOV0DXx177wMNayZNr3NbRkSks5rGBVOgaJpZM4G320c> .

[kgjenkins]

FYI, I just fixed my typo -- it's i.rgb.his (not "rdb")

Good luck, and hopefully there won't be any JPEG artifacts that complicate things...

[cm325]

I was not confused :)
fingers crossed!