I've been playing with some terrain transformations, and found that some wouldn't work. I'd like to see if anyone else has had the same difficulty, and whether there are others that do work besides the ones I've noted. The following list gives transformation paths I've tried. Ones that worked are highlighted with an asterisk (*). Interestingly, I haven't seen some work that are actually put in the game as defaults (such as Marsh -> Coast). That seems wrong.

Pollution transition changes (due to global warming):

*Jungle revert to base terrain
*Forest revert to base terrain
Plains -> Desert
Tundra -> Grassland -> Coast
Grassland -> Coast
Grassland -> Plains
Marsh -> Coast
Sea -> Coast
Coast -> Plains
Ocean -> Hills
*Hills -> Mountains
*Mountains -> Volcano

Please add to this list if you've found others that work. I'm also interested in whether there are limitation that I'm not aware of (such as water is not allowed to transform).

I've seen marsh->coast work in the past. In fact, I've seen nearly every combo work previously. Maybe this is a new bug?

I wonder if maybe there is a limit to how many combos can be coded into a scenario? The fact that only four are working in my mod made me question it. I've also wondered if maybe there are certain ones that are hardcoded to work a certain way. I've put in Tundra -> Grassland -> Coast (to simulate ice caps melting and the sea rising). But if Grassland is supposed to transition only to Plains (as it is set in the default game), perhaps changing it messes up the transformation chain.

Anyhow, I've gotten the sun to go orange in my game, and I'll keep nuking away until it is red. Maybe the game has a preference for certain transformations and uses them up before moving on to others. Perhaps after all the hills have become volcanoes, I'll finally see some Tundra turn to Grassland....

One of the vanilla civ patches changes global warming to target vegetation squares(forest and jungle) first before changing "real" terrain.So you can reforest tiles to absorb GW.

About editor combos...anything which converts land into water (and I think it is the same the other way round) doesn't seem to function.What works for me in a special test game was converting mountains and hills into grassland and I'm almost sure I see changing plains in to desert in an epic game, but it was a long time ago.

I like Alpha Centauri for it realistic climate model...a shame that it is not even possible to mod this for CivIII.

I've seen marsh->coast work in the past. In fact, I've seen nearly every combo work previously. Maybe this is a new bug?
I personaly see that it is appropreate to see the Marshes transform into Coasts when the Glacires (Sp?) melted

I personaly see that it is appropreate to see the Marshes transform into Coasts when the Glacires (Sp?) melted

So an inland marsh would create a coast tile all alone. While the game's engine allowed the neighbouring marsh tiles to stay marsh. Does not sound appropriate to me. And it is called glaciers, afaik, since you wondered

On topic:
My experience with global warming is limited at best. So far, I have had grassland changing into plains and forest into grassland. IIRC, that also happened in vanilla, so it came as no surprise.

quite weird how these more "random" results happen to you guys. I suppose this has nothing to do with versions - since I reckon Firaxis must have had other priorities than terrain on their minds

So an inland marsh would create a coast tile all alone. While the game's engine allowed the neighbouring marsh tiles to stay marsh. Does not sound appropriate to me. And it is called glaciers, afaik, since you wondered

It would be an inland lake. We have those on all maps. It seems though that you could also have Marsh revert to base tile if you chose.

One of the vanilla civ patches changes global warming to target vegetation squares(forest and jungle) first before changing "real" terrain.So you can reforest tiles to absorb GW.

That is very heartening to hear. I'll just stick with it and see if we get to the other combos after deforestation is complete. I gave my game file to my 4-year old daughter today and instructed her to nuke everyone until the sun turned red. Forty-five minutes later, she proudly informed me that she was done. Fast learner.

EDIT: I played it for 20 turns past the point where all hills and mountains had become volcanoes, and all forests and jungles had reverted to the base terrain. The sun is still red, and no global warming effects are taking place. So it looks like that's it for the changes--desertification is not going to kick in. I'll try it again tomorrow with the hill and mountain transformations turned off, and see if I can get other changes to occur.

@ Camber

Keep up your work, it may be not a big thing for most of the mods playing in the past, but I am still working on the concept for an Alpha Centauri mod and so im highly interested in whats possible about GW.

Thanks for the encouragement. I just tried it again, with the following transformations coded. The ones that worked are flagged.

Revert to base terrain:
*Jungle
*Forest
*Marsh (apparently the Marsh -> Coast that Firaxis put in doesn't function. But Marsh to base terrain does.)
Also coded but not functioning:
Tundra -> Grassland
Grassland -> Plains
Plains -> Desert
Ocean -> Coast
Coast -> Plains
Ocean -> Hills

I didn't have the Hills -> Mountain -> Volcano path active, so we know that wasn't interfering with Desertification.

Incidently, Pfeffersack, what transformations did you have in Alpha Centauri? I imagine some kind of terraforming effect, where you were bringing green to a desolate landscape as you built more and more terraforming improvements. Too bad we can't have nuclear waste give a food bonus; we could use them as oxygenators or seed dispersers.

EDIT: This is off topic, but I've found some other things in my testing:
a. Barbarians won't use nukes. Also, nukes cannot be made to do defensive bombing (when they are in a stack of attacked units).
b. The AI won't use ships that build roads in water.
c. If ships are given the Build Roads job, Harbors don't work. Instead the game waits for sea lanes to connect trade routes.
d. Nukes that are detonated in the water (with no visible polluting effect) still affect global warming.

EDIT2: I tried it again with the following transformations coded. None of them worked. This was with a red sun (I'm launching 60 nukes on turn 1 to jump-start global warming).
Tundra -> Grassland
Grassland -> Plains
Plains -> Desert
Ocean -> Coast
Coast -> Plains
Ocean -> Hills
So I think that weakens the theory that vegetation transformations are taking priority over regular terrain changes. With no vegetation transformations coded, and hills and volcanoes left out, nothing happened at all!

EDIT3: I removed all water transformations, and all vegetation transformations. The remaining ones are listed below. Again, none worked. (This was with a red sun after firing 60 nukes. I also have 20 cities with a population over 13 in the scenario.)
Tundra -> Grassland
Grassland -> Plains
Plains -> Desert

I have a new hypothesis. So far, the only transformations that have worked, are ones that did not require the base terrain of surrounding tiles to change. My theory is that Grassland -> Hills would work, and so would Hills -> Volcano, and Grassland -> Volcano. I also hypothesize that Grassland -> Marsh, Forest, or Jungle would work, and that Plains -> Forest and Tundra -> Forest would work. See the pattern? None of these require a neighboring tile to change its base terrain when the transformation occurs. Whereas others I have tried (that failed), such as Grassland -> Plains, do require that surrounding tiles shift to accomodate the change. I think that this is the underlying problem. If this is true, then I should also be able to do the following transformations: Marsh -> Hills, Marsh -> Volcano, Jungle -> Hills, and Jungle -> Volcano. This is because the underlying terrain for Marsh and Jungle seems to invariably be Grassland.

EDIT4: Success! The following terrain transformations were coded in my last game, and all worked:
*Grassland -> Marsh
*Plains -> Forest
*Hills -> Jungle
*Marsh -> Hills
*Jungle -> Volcano
*Volcano -> Grassland
*Tundra -> Forest
*Mountain -> Marsh

What does this mean? It confirms my hypothesis that terrain transformations are only supported if they do not affect the surrounding base terrain. So, Tundra, Plains, and Desert could not change to anything else, although Tundra and Plains could have Forest laid over them. Also, the following list shows six tiles that are interchangeable: you can code them in any order of transformations, even if they loop in a circular progression (!):

Grassland
Marsh
Jungle
Hills
Mountain
Volcano

This is due to the fact that all six of these tiles share the same base color at the edges (deep green aka grassland), so surrounding base tiles are not affected by the transformations. Note that if Forest is coded to transform to something (other than base terrain), it probably won't work. This is hypothesized because Forest can occur on 3 different types of base terrain, some of which aren't interchangeable with the above mentioned Grassland family.

Which brings us back to some hearsay.


One of the vanilla civ patches changes global warming to target vegetation squares(forest and jungle) first before changing "real" terrain.So you can reforest tiles to absorb GW.

If what Pfeffersack said is true, this patch to which he refers may be the culprit--it may have forbidden transformations that affect the base terrain of surrounding tiles.

Does this mean that global warming doesn't work? Absolutely not! Although I'm disappointed that I can't make the seas boil and dry up into deserts and mountains, it does suggest a new mod. While global warming cannot cause the climate to get hotter and dryer, it can be used to make the climate hotter and wetter. I suggest that the following terrain transformations would make a logical and unified progression for global warming:

Tundra -> Forest
Plains -> Forest
Grassland -> Marsh
Marsh -> Jungle
Volcano -> Mountain
Mountain -> Hills
Hills -> Grassland

The last 3 on the list would represent erosion--mountains being worn down by torrential rains caused by severe climate change. If you want to make the whole progression cyclical, you can make new volcanoes erupt out of the jungle. That would yield the following progression:

Grassland -> Marsh -> Jungle -> Volcano -> Mountain -> Hills -> Grassland, etc.

Unfortunately, Plains and Tundra would be stuck with Forest, and Desert and Floodplain would not change.

I've incorporated this research into a new mod, called nucleartropics. For maximum tropification, go here! (http://forums.civfanatics.com/showthread.php?p=1884729#post1884729)

Terrific and impressive work, Camber!!! :goodjob: :king:

Running my own tests a few months ago I was disappointed because I haven't seen any pattern with whats possible and whats not possible.
Your research has uncovered the inner mechanics of climate transformation -
I think this would be an nice article for the War Academy.

The effect of GW in Alpha Centauri is mainly a rising sea level, which transforms landtiles in sea tiles.Other transformations can be undertaken by the player (altering humidity, creating mountains or flattening them or lowering sea level by starting the solar screen).

Sadly the land/water transformation isn't possible with Civ III, but the cycle you have discovered is a nice way to simulate a climate change process.
It has the potential to make the effects of GW really strong - imagine the danger of new erupting volcanoes near cities.It will surely make it in my AC mod :)

Rising sea level is something I would like to see. Too bad it isn't possible here. I did go ahead and try using the progression that I described, but the volcanoes were too common. I also wanted something that wasn't randomly cyclical. So I ended up with the following transformations:

Volcanoes -> Mountains -> Hills (the path of erosion)
Tundra -> Forest
Plains -> Forest (the paths of forestation)
Grassland -> Marsh -> Jungle (the path of junglification)
Forest -> Volcanoes -> Mountains -> Hills (my source of new volcanoes)

Because only forested Grassland tiles will actually respond to the last transformation, it makes the creation of volcanoes less widespread. I mainly wanted there to be a "resting state" for all of the paths, so that there would be a sense of progression. In my mod where I had the grassland family of tiles in a complete cycle, it got way too random-looking. The nice thing was that it was total cataclysm -- the map was pretty unrecognizeable after 200 turns. But it also meant that you couldn't count on anything staying the same. The above list of transformations gave me a nice balance between change and stability.

Sorry to double post, but I had a new issue in this thread.

While going through the script.txt file, I found the following entries for global warming:

#GLOBALWARMING_FOREST_JUNGLE
#map_center 0
Global Warming" has destroyed $TERRAIN0!

#GLOBALWARMING
#map_center 0
Global Warming has turned $TERRAIN0 into $TERRAIN1!

#GLOBALWARMING_RISING_COASTLINES
Coastlines are rising! Our scientist blame it on Global Warming...

I have seen the first two scripts appear in-game, but not the third. In fact, my tests have shown that land -> water transformations aren't possible. I found this text in both the vanilla and expansion text files, but C3C is the only one that has a land -> water transformation coded in the game file (conquests.biq has Marsh -> Coast coded, but it doesn't work in my tests).

Is there anyone out there who has seen the following message in a game?: "Coastlines are rising! Our scientist blame it on Global Warming..." I would really like to see this work!

Update: Forested tiles can be set to turn into any tile on the Grassland-compatable list, even if they are forested Tundra or Plains. I discovered this when trying to make only forested Grassland switch to Hills--instead, I got hills popping out of Tundra and Plains also, anywhere there had been a forest.

Rising sea level is something I would like to see. Too bad it isn't possible here. I did go ahead and try using the progression that I described, but the volcanoes were too common. I also wanted something that wasn't randomly cyclical. So I ended up with the following transformations:

Volcanoes -> Mountains -> Hills (the path of erosion)
Tundra -> Forest
Plains -> Forest (the paths of forestation)
Grassland -> Marsh -> Jungle (the path of junglification)
Forest -> Volcanoes -> Mountains -> Hills (my source of new volcanoes)

Good work Camber.

I like your recommendations, except for:
Volcanoes -> Mountains -> Hills (the path of erosion)

I'm thinking that a mountain would not be able to erode down to hills in a period of 6000 years (duration of the game). It would take millions of years even with constant rain and wind.

I am going to implement (since this is what is working):
Tundra -> Forest -> Jungle
Plains -> Forest -> Jungle
Grassland -> Marsh -> Jungle

Edit: Is this right according to your latest update.

I've notice that jungle can also be changed to marsh.

I had it accidentally set jungle>marsh and marsh>jungle and I had a tile go back and forth from jungle to marsh and back to jungle after a hugh nuclear war.

That's right. The interchangeable tiles (Plains, Grassland, Hills, Forest, Mountain, Volcano, Jungle, and Marsh) can be set to transform into one another in any order. The only other change available besides the various combinations of these, that I'm aware of, is Tundra to Forest, which can then transform into any of the above tiles.

Are plains and grassland interchangable, I have been treating them as if there not. If they are I will have to rethink my order.

Currently:
Tundra>Forest>Jungle
Plains>Forest>Jungle
Grass>Marsh>Jungle

Maybe should be:
Tundra>Forest>Jungle
Plains>Grass>Marsh

I should add one other tidbit that I've found about global warming.

Terrain types can be set to revert to themselves. When this occurs, they lose some improvements. For example, I had Hills transform into Hills in a game, and they lost something (can't recall what, possibly mine or road). More research is needed on this.

I also tried making LM terrain transform to regular terrain of the same type (i.e., LM Hills to Hills) and it did not work; the LM Hills remained as LM Hills when hit with global warming--but lost some improvements.

Another observation is somewhat off-topic and also involves LM terrain. I noted that improvements must be enabled for both the LM and regular varieties of each terrain type, or else workers will not be able to make them. For example, I tried making irrigation available for regular grassland but not LM Grassland, and mines available for regular Mountains but not for LM Mountains. This resulted in irrigation and mines, respectively, being unavailable for worker jobs on all types of Grassland and Mountains. It is possible to have the food, shields, or commerce bonuses from improvements *vary* according to whether the terrain is LM or regular, so long as the value on both variations is 0, or 1 or more. So...I could make LM Mountains get +1 shield for mine and regular Mountains get +2 shields for the same improvement, or both get zero, but not have one get zero and another get 1 or more.

Are plains and grassland interchangable, I have been treating them as if there not. If they are I will have to rethink my order.


Yup. Everything on that list is interchangeable. It is much more simple than we originally thought. It makes sense though, since they all share a sort of dark green base palette. The Plains are the only ones that are a little unexpected if you look at it as a "base green" issue, but you'll note that they are only a lighter version of the same green.

EDIT: actually, it has been almost a year since I've looked at it. So if you could double check the Plains->Grassland transformation, I would appreciate it.

Any data on the rates of Global Warming and how Pollution or Nukes contributes to it?


Also, I've noticed in tests that once a Forested Tundra/plain is put into the 'green square' rotation, clearing it results in Grassland (So Forested Tundra->Jungle Cleared is Grassland)...Other notes: 'Clearing Wetlands' can be applied to any of the 'green terrains' except grassland.. so Hills, Mountains, and Volcanoes can be 'cleared' (I haven't seen what can happen if you 'chop' them instead) and will become Grassland

Tests from this idea

Clear Wetlands for Forest instead of Chop... Will this make Forested Tundra turn to Grassland when cleared?

Chop Forests for Jungle, Swamp, Hills, Mtn, Volcanoes...will you get the bonus? would a 'Chopped' Jungle that had Warmed from a Forested Tundra yield Tundra again?

I don't have answers for all of these, but I'll give the few I know.

I do know that when I gave 100 nukes to a test player and launched half of them into the sea, it made the global warming indicator go red. The player also had a couple of cities that were size 3, so I can't tell you what portion of the global warming was caused by the nukes. I can also tell you that at that point, on a small map, I had over 20 tiles transforming each turn (I had reveal map turned on). Some tests on the exact dynamics of these factors would be interesting.

It appears that clearing a forested Tundra or Plains tile will reveal Tundra or Plains underneath. But Hills, Jungle, Mountain, Volcano, and Wetlands all have Grasslands as the base underneath. Once a Forested Tundra or Forested Plains tile turns to one of these (Hills, Jungle, Mountain, Volcano, or Wetlands), I think you lose the Tundra or Plains properties of that tile permanently--the base is now Grasslands. Some confirmation tests would need to be made; I'm still a little fuzzy in my memory of whether or not Grasslands can be made to turn into Plains.

I had never heard before that you can assign the "clear wetlands" or "clear forest" jobs to Hill, Mountain, or Volcano terrain. But if it is true, it would undoubtedly result in a Grasslands tile, as this is the base for these types of terrain.

I do remember testing a while back in which I found that
The Plains was not part of the "Grassland group" and

Desert and Flood Plains had a cycle of their own (where they could Terraform to each other with Desert counting as the Base Terrain for Flood plains...although, as I remember, it didn't work for Floodplains that actually bordered on a river)

I'd imagine the Water tiles might also be part of a group (although I'm not certain what Coastal Tiles that are actually coastal would do if set to change into Ocean or Sea..whether they would resist it, do it , or start eating into coastlines while doing it)

Well, In my Tweaked Out mod, I had Forest go to Rain Forests (in my mod, jungles were re-named rain forests and started as 1 food, 1 shield and 1 commerce (to represent the scientific breakthroughs from studying rainforest plants)), and rain forests go to grassland. In addition, I had tundra change to grassland.

In the course of a long playtest game where I was playing as the Ottomans, I was running my country with my usual put greenpeace to shame game style (I'd added an additional pollution reducing city improvement for the late industrial era, and I'd build all pollution reducing buildings before building a single factory or hospital), but the AI went with their typical slash & burn policies, so the sun was orange by 2000, and I personally witnessed grasslands changing into plains, plains into deserts, one of my forests (and my small empire had about 90% of the world's forest acreage) turned into a Rain forest, and one of my Rain Forests turned into a grassland. I've never seen whether or not the Tundra to grassland worked or not.

Did some testing and here are results

Terrain change is only allowed in 3 ways [a series of Comprehensive tests in each I had starting nukes, 3x3 blocks of each terrain, including Forested Plains and Tundra, and would make All terrains globally warm to one type (except for the terrain itself which would warm to none)]

1. "Grassland group" (Grassland, Forest, Jungle, Marsh, Hill, Mountain, Volcano) The Base Terrain here is grassland and so these can be Chopped/Cleared into grassland, have Forest planted on them, and all Globally Warm to each other

2. Plains+Tundra are two parts of the Non-Grassland group that can be Forested, they Keep their Base Terrain. However, If they are forested, and the forest globally warms to another part of the grassland group, the base terrain changes. Also Forest can globally warm to either of them in which case the base terrain is changed...NOTE: even if the Base Terrain changes, the GRAPHICS of it don't change, so you can have a white grassland or a green tundra.

3. Desert, Floodplains, Desert is the Base Terrain for Floodplains (note this test was done without any rivers, so I think Rivers may Force the Desert to be Floodplain) but Desert CAN warm to Floodplain and non-river Floodplain can be Warmed, Cleared, or chopped to Desert

4. Water Tiles cannot terraform to or from anything (I didn't try 'Clearing/Chopping them' but no global warming transforms worked)


On actual rates
2 tests each had 100 pollution from cities (from a building), and a size 2000 map
1 had 2000 globally warmable squares Jungle<->Marsh interchangeable starting all jungle
The other had 25 Jungle and the rest of the map Plains (which were set to Warm to none)

It seems the chance of a square globally warming is determined for each square, because the block of 25 Jungle had 3 squares warm in ~200 turns, wheras the pure Jungle Map had about 121 including returns over 80 turns (about because that involved counting the global warming messages each turn)

This means that the chance of an individual square globally warming (not including Nuclear effects) is

(Total Pollution/Map Size)^2 *0.3

OR to put another way

Number of Squares that will globally warm / turn=
(Total Pollution^2)/Total Map Size * Fraction of squares on Map that will pollute * 0.3

(the 0.3 is an estimate it could probably be anywhere from 0.2 to 0.5 ..also in the first test I had 15 cities to be able to monitor all the Global Warming going on, but only one had the Polluting building)... Also the Sun was Yellow the whole time for both of those tests.

Great work. Thanks alot for the effort. This is very useful info.

Wow, it looks like we finally know basically how it works. Thanks Krikkitone :D

Bit more Testing it seems the number of Squares that Transform in a turn are

(Pollution/Total Map squares)^2 * Pollutable Squares * 0.3
OR
% chance of a square polluting = (Pollution/Total Map squares)^2 * 30
I'm pretty sure that this determines the Sun Color

Nukes are permanent (or at least the color of the sun caused by nukes won't change for ~300 turns)...on a size 2000 map, 16 Nukes turned the sun light orange, 23 turned it dark orange and ~150 more than that didn't change the color.