Improve styling of new global landcover data

[clhenrick]

reference #12

[clhenrick]

@almccon I pulled out the RGB values from the VRT file for the previously used landcover data, matched them up with the legend you posted in #12 and this is what I got:

Urban and Built-Up Land = r: 239 g: 239 b: 239
Dryland Cropland and Pasture = r: 222 g: 224 b: 190
Irrigated Cropland and Pasture = r: 213 g: 224 b: 164
Mixed Dryland/Irrigated Cropland and Pasture = r: 0 g: 0 b: 0
Cropland/Grassland Mosaic = r: 213 g: 224 b: 164
Cropland/Woodland Mosaic = r: 9 g: 174 b: 132
Grassland = r: 222 g: 224 b: 190
Shrubland = r: 221 g: 224 b: 206
Mixed Shrubland/Grassland = r: 221 g: 224 b: 206
Savanna = r: 222 g: 224 b: 190
Deciduous Broadleaf Forest = r: 43 g: 174 b: 103
Deciduous Needleleaf Forest = r: 43 g: 174 b: 103
Evergreen Broadleaf Forest = r: 43 g: 174 b: 103
Evergreen Needleleaf Forest = r: 43 g: 174 b: 103
Mixed Forest = r: 43 g: 174 b: 103
Water Bodies = r: 222 g: 224 b: 190
Herbaceous Wetland = r: 0 g: 0 b: 0
Wooded Wetland = r: 0 g: 0 b: 0
Barren or Sparsely Vegetated = r: 239 g: 243 b: 217
Herbaceous Tundra = r: 0 g: 0 b: 0
Wooded Tundra = r: 43 g: 174 b: 103
Mixed Tundra = r: 222 g: 224 b: 190
Bare Ground Tundra = r: 0 g: 0 b: 0

Looks like some of these categories have 0,0,0 which makes sense because they're mostly absent in the U.S. I'd imagine? Thinking we may have to come up with new values for global coverage.

[clhenrick]

This color chart is from a Tom Harrison blog post and might be a good starting point:

Pulled out the color values from the image and they are here:

[clhenrick]

Here are some screen shots after applying the above color palette:

[clhenrick]

Here is the new landcover style viewed while running Tessera at zoom 9:

Vs. OG Terrain style:

Thoughts:

• Something about the bright yellowish color next to the deep green in the sierra nevada doesn't sit well with me. According to the new lancover data these are grassland areas.
• Perhaps changing to lighter green would provide for a better transition?
• Or did the OG Terrain tiles include park areas at lower zoom levels to get that much of a deep green coverage? Is that something we'd like to replicate?

[almccon]

Hmm. I don't think that OG Terrain is doing anything with park areas at that zoom. I think it's just a difference in classification between the two landcover sources.

I think we should nudge the Tom Patterson colors in the direction of the OG palette, even if it reduces the amount of variation. Especially the yellows of the grassland and the blue-green color of the evergreen forests should converge on the overall olive tones of OG.

[clhenrick]

@almccon sounds good, I'll give it a shot.

[clhenrick]

Altered the palette towards being a little more olive similar to the OG Terrain. Here's the shot of the Bay Area to the Sierra Nevada:

North America:

North Africa / Southern Europe:

I would've liked to grab these screen shots while running Tessera but it seems to be busted for me currently. Looks like something to do with the terrain-class.mss file being in a sub-directory?

[clhenrick]

Here's a screenshot while running Tessera:

[clhenrick]

• [x] make forests less blue so they don't clash with park color
• [x] brown out grasslands color (see transition to southern Sahara)

[clhenrick]

@almccon modified: evergreen forests > less blue; grasslands > more brown:

[clhenrick]

seems like it clashes less with parks at higher zooms now as well:

[almccon]

Yeah, looks really good. It might be a little too yellow overall... like I expect to see a little bit more green adjacent to the rainforests in Africa and Brazil. But as far as I'm concerned this is good enough to launch, so I'm happy to leave any further color tweaks as a nice-to-have.

[clhenrick]

hmmm, yeah I agree about it looking too yellow, I'll try tweaking the grasslands color once more.

[clhenrick]

Attempted tweaking the grasslands color a bit, but think the issue may also have to do with what areas the data classifies as grassland. When comparing the current style of the new data with the OG style and data it's very noticeable that areas were classified differently. The new data seems to apply the grassland classification to a much broader area then the old data.

If you don't approve of the current iteration I'll revert to the previous one from above.

[almccon]

Thwarted by those grasslands again! I think that was my original problem with this data. Let's revert.

[clhenrick]

reverted!

[almccon]

I couldn't resist one more tweak to bring in even more of the olive color from the OG style, plus some more greenish in the woody savanna adjacent to the jungles in Africa.

[clhenrick]

I like it!

On Friday, September 18, 2015, Alan McConchie notifications@github.com wrote:

I couldn't resist one more tweak to bring in even more of the olive color from the OG style, plus some more greenish in the woody savanna adjacent to the jungles in Africa.

[image: screen shot 2015-09-18 at 2 28 19 pm] https://cloud.githubusercontent.com/assets/1212178/9971587/df0efcf0-5e11-11e5-9c8f-a32174c3cf6b.png

— Reply to this email directly or view it on GitHub https://github.com/stamen/terrain-classic/issues/23#issuecomment-141573972 .

-Chris L Henrick

MFA, Design & Technology Parsons, The New School For Design

http://chrishenrick.com Github: @clhenrick https://github.com/clhenrick Twitter: @chrislhenrick https://twitter.com/chrislhenrick

“Whenever I see an adult on a bicycle, I no longer despair for the future of the human race.” — H.G. Wells