In reply to the postings on the previous thread:

LGJ is right about changing the entire border. That is probably too much. But testing every individual square would take too long. So we could run the tests for groups of about five to seven adjacent squares.

He also has a point about the change being rather severe. If we use only a few vegetation types, it would be a big jump. I think that we should have more vegetation types, such as:

Wasteland (Nothing at all)
Desert (cacti, desert wildlife)
Scrubland (marginal agriculture possible)
Savannah (reasonably fertile)
Plains (even more fertile)

Trees can grow on any of the last three terrain types. Unless you have good technology, you usually can't tell what kind of terrain deforestation will result in. The three levels of trees are:

Sparse Forests (Trees mixed with the base vegetation type. Here you will know what deforestation will get.)
Forest (Basic Trees)
Old-Growth Forest

This list was for vegetation in temperate zones. They would have a corresponding type for tropical and arctic zones. For example, a Tropical Rainforest is the equivalent of an Old-Growth Forest.

This vegetation type is mostly independent of the geology, or physical shape of the land. Forests can grow equally well on hills and flatland.

However, there is one case when the geology has a big impact on the vegetation. That is when the water table comes above the surface of the ground. In that case, you get swamps, bogs, etc.

These are just preliminary thoughts. If they sound good, I could add more details and make a more complete proposal.

I still think that randomness is the way to go when determining the effects of pollution. The theories about the effects of atmospheric pollution are constantly changing. Five years ago, few people would have thought global warming would lead to an ice age. Five years from now, there will probably be a different theory.

On randomness do to pollution I'm all for it, except we should have some sort of trend because in the game the scientists will be discovering this and trying to find ways to prevent it.

One other thing is that we know that these gasses do cause warming. There have been theories about it causing ice agaes since near its infancy of the greenhouse effect theory. Anyway because volcanoes produce the same type of pollutants we also have to remember that they also affect things, not just humans.

Also with just random things happening you won't have any long-term implications to worry about, just short term. This gives the message to the player that whatever they do the idivisual effects won't last long and can be corrected (FE if a swamp is created you can drain the swamp).

You've got some really good ideas about the different types of terrain and their changes. Blocks are probably a good way to handle those changes, but I don't know what a good size would be, 5x2 sounds o.k. though.

There was some discussion in the old map generator model thread about half way down, it was mostly mining, but could be used for the deforestation rules also. There was also a little discussion about having more terrain types, about a quarter way down. Mark had mentioned putting that off until later because it would be an easy add-on.

Richard & LGJ
Global Warming: Well, i really didn't mean for this to get in here but anyway... I think we can all be happy with random problems arrising through global warming, we've established that.

So how about we use the death rate increases & terrain changes for the random stuff. And track the avg temp. for the disaster and disease increases. With these linked to the avg. temp, we would keep the player from being able to fix all the problems associated with warming.

Note that some places we call mountains, like the Appalachians, are considered to be hilly terrain at high elevation for the purpose of this model.

Sunken terrain is different. It would not have the wateland-plains spread like the other terrain types. It's vegetation is limited to four levels of forest cover, from No Trees to Old Growth.

High elevations would be considered arctic terrain, while some coastlines in temperate regions would be considered tropical terain.

So, each land type is a combination of geology, climate zone, and vegetation and/or forest cover. If we want, we can give a name to each combination. For example, sunken terrain in a temperate region with an old-growth forest would be called a Cypress Swamp, and flat or rolling arctic scrubland would be called Tundra.

I mentioned this on the disaster thread, but it also belongs here. Any stress to the atmosphere, like global warming, should result in those wind currents being regenerated at random. This would have a big impact on a lot of terrain, and would make climate change interesting.

I would also suggest that water currents be included. These would also be randomly generated at the start of the game and whenever the global temperature changes. The effect of water currents would be to change the temperature of some areas. Great Britian is at the same altitide as Newfoundland, but it is a lot warmer because the Gulf Stream brings warm air up from lower latitudes.

Also, I think we should use six climate zones like Toubado_Koomi proposed in the map thread. That would be more realistic than only three climates.

The biggest problem areas, at least for me, are wetlands and aquatic ecosystems, and some of the environmental disasters. I am not totally pleased with these areas, but I am posting what I have to stimulate discussion. Maybe someone else can think of a way to improve these or other aspects of the model.

The model covers climate, vegetation, farming effectiveness, and some ecological disasters. Most of the calculations are done at the beginning of the game, so this shouldn’t take up too much processor time. I think that the initial vegetation generator should be a standard part of the game, but the routines for climate change during the game can be turned off for multiplayer games or slow computers.

My goal is not to mimic our planet. It is to generate a random Class M (earthlike) planet. This means that some things that occur on Earth might not happen exactly the same way here. I am only using general geological principles to try to create a realistic game environment that can run quickly in the background.

This ecosystem generator is designed run right after the map generator system that creates the physical shape of the land. Specifically, it is responsible for parts two and three of the posted map generator outline, which as far as I know have not been implemented yet.

I am assuming that part one of the map generator creates multiple types of terrain, something like: mountainous, hilly, rolling, and flat. I am also assuming that it tracks altitude. I haven’t seen recent documentation for all of this; I am working from the information on the clash website. If any of these assumptions are wrong, please let me know.

The first step is to generate climate zones. There are six climate zones per hemisphere, determined mainly by latitude. Each tile is assigned a base climate based on latitude:

Tropical: equator-12
Sub-Tropical:12-24
Warm Temperate: 24-40
Cool Temperate: 40-55
Sub-Arctic: 55-67
Arctic: 67-pole

As an option, there could be a bit of randomness to these numbers. It could also be possible for these to be changed in the game due to varying solar activity. This would simulate the “little ice age” that hit a few hundred years ago and disrupted, among other things, the Vikings’ colonization of Greenland.

These climate zones are modified by altitude and ocean currents. Every 100 meters above sea level is equivalent to a latitude of about 90 kilometers closer to the pole. After the climate has been adjusted for altitude, ocean current vectors are generated at random. These currents then influence the climate of nearby coastal regions. These areas are given the unadjusted climate of tiles a certain distance away in the direction that the current came from. This distance is determined by the size of the vector. So a strong current moving north in the southern hemisphere could cause lot of the nearby coastline to be changed from a warm temperate to a cool temperate climate.

The main effect of climate zones is to alter the effectiveness of farms, as warmer climates have a longer growing season. Climate will also determine what kinds of vegetation will grow in the tile, although in many cases this difference will only be cosmetic. Climate can have other uses, such as influencing the attrition of military units.

After the climate is generated, soil patches are assigned at random using a fairly broad brush. “Soil” is a term that includes every geological and chemical influence on plant growth in that area. It ranges from zero, meaning nothing at all (not even lichen) can grow, to eight, which represents exceptionally good growing conditions.

Soil is limited by climate and terrain. True arctic terrain, which is basically ice or permafrost, is limited to soils of zero to two. Sub-arctic terrain can only have soils from zero to five. Mountainous terrain limits the soil to a maximum of two and hilly terrain limits the soil to four, although terrain improvements like terracing can help out hilly terrain.

Areas beside the downstream end of large rivers are given a Soil increase of +1 to +3.

Note that some places that are called mountains, like the Appalachians, are considered to be hilly terrain at high elevation in this model.

After the soil is generated, the water rating of the terrain is assigned. “Water” is mostly rainfall, but also includes river flooding. It does not determine wetland formation or groundwater reserves; it is a measurement of how much water is naturally available to land plants. Irrigation and other terrain improvements do not change the base water rating; they provide an adjustment for farm purposes only. The water rating goes from zero to eight.

For this part of the model, the tiles are assigned water ratings in large random patches. Then, wind vectors are generated at random. If a wind current goes from the sea to the land, the coastal tiles will have a high water rating. If a wind current is blocked by high elevations, the downwind tiles will have a much lower water rating and the tiles on the windward side of the high elevation tiles will have a higher water rating. All of these ratings are then modified as follows:

All tiles in the Tropical climate zone are given a water rating of +2, all tiles in the Warm Temperate climate zone are given a water rating if –2, and all tiles in the Cool Temperate climate zone are given a water rating of +1. This is due to the large atmospheric convection cycles that would probably occur on any earthlike planet.

Water affects the soil rating due to leaching and erosion:
Water 6: Soil –1
Water 7: Soil –2
Water 8: Soil –3

Water and Soil ratings are not allowed to below zero or above the terrain limitations. If they do, they are simply assigned ratings of the minimum or maximum value allowed.

After Soil, Water, and Climate are assigned, the natural vegetation is generated. Each vegetation type can grow only in certain conditions, but most tiles will be able to support multiple types of vegetation.

In an effort to simplify things, I have come up with the “ecological province.” An ecological province is any patch of tiles that has the same climate, soil, and water. These should generally be patches of three to thirty tiles. An ecological province is assumed to be a region with almost identical environmental conditions.

Each ecological province is assigned one vegetation type out of all the vegetation types it could support. This base vegetation will not change, although the borders of the provinces could shift.

I do not have a complete list of vegetation types because that would take a good bit of time and I am fairly certain that the model will need to be revised, which would mean the list could become obsolete. But I do have samples:

Taiga
Requires: Sub-Arctic or Cool Temperate, Water 4-8, Soil 2-6
Effects: Provides low-quality softwood lumber. Reduces movement rate of units.

Temperate Grassland
Requires: Warm or Cool Temperate, Water 2-5, Soil 4-8
Effects: Soil +2. Can only be farmed if Farm Tools is at Level 30 or better (Heavy Plow)

Farming effectiveness is based primarily on the Soil and Water values of an ecological province. Each one of these is a limiting factor, so the smallest value is the basis for the crop yields. This number is then altered by climate and technology.

For this calculation, irrigation improves the Water value and fertilization improves the Soil value. In some cases, irrigation is a waste of time, and the player would be told this. However, the Water value should usually be the limiting factor, at least before the land is exhausted.

Aquatic ecosystems like wetlands, lakes, and oceans are determined by a different system. Water is irrelevant, and the ecology is determined by climate and a single Water Quality factor that goes from zero to eight. This factor has nothing to do with how suitable the water is for humans to drink or use; it describes how well the water can support life.

Wetlands are generated on land tiles, and they can have different vegetation types like other terrain. There are three types: Basin, Riverine, and Coastal.

Basin wetlands form on flat or rolling terrain where the water table rises above the land. If we use groundwater, then places with low elevation and very high groundwater reserves become these wetlands. If we don’t track groundwater, then random low elevation ecological provinces become basin wetlands.

Riverine wetlands occur alongside rivers in flat terrain. They have a chance of occurring in a low elevation ecological province that a river is passing through. Ecologically they are similar to Basin wetlands, and have the same vegetation types.

Coastal wetlands and estuaries form on flat coastal areas, often at the mouths of rivers. Because of their high salt concentrations, they have a different set of vegetation.

Wetlands can increase disease, but they decrease the chances of flooding, decrease the effects of pollution, and restore groundwater levels and purify the water (if we track this). All wetlands greatly slow the movement rate of units and are impassable to most mechanized units. Wetlands can be drained to make normal terrain, which is usually fairly good farmland. When this happens, the terrain is assigned a Soil and Water rating.

Lake and Ocean tiles also have ecology types based on Water Quality. The main effect of this will be fishing yields, but hopefully there will be undersea exploration and colonies later in the game. This will cause interactions with underwater terrain similar to the normal land interactions (movement, farm productivity, terrain improvements, etc.)

After the terrain is generated at the start of the game, it can undergo changes as time progresses. Some of these will be small and mild, but others will have a big impact on almost everyone.

The standard ecological change or ecological disaster progression is: Check, Hotlist, Countdown, Change. Checks are performed for many different reasons, and if the check is positive the area is put on the hotlist. Depending on your civ’s technology level, you can be warned about an area on the hotlist. This could allow you to take action to prevent the problem. All areas on the hotlist are monitored every turn to see if a change will really be triggered. If a change is shown to occur, the area is put on the Countdown list. You can be still warned about these events, but you can no longer prevent them. Sometimes something will happen every turn on the Countdown list, causing the change to be phased in. Other times, the change happens all at once after a certain number of turns.

The simplest and most common change is a shift in the border between two ecological provinces. This would usually result in a few tiles switching vegetation. Checks for this would be performed on one border of about fifty ecological provinces every turn. If the check is positive, the border is hotlisted. There are other conditions, discussed later, that can cause borders to become hotlisted. If the border passes two checks in a row, then it shifts after a countdown of three turns. If it fails two checks in a row, it is removed from the hotlist.

The check is:

C = Rand / ( 1+ abs(s1-s2) + abs(w1-w2) )

Where s1 and s2 are the soil values of the two provinces and w1 and w2 are the water values of the two provinces. Thus, the borders of two provinces with similar geological characteristics are much more likely to shift then the borders of two dissimilar provinces.

The check is positive if the number is above some change number defined for each terrain boundary. Some changes are more likely then others; changes that represent natural succession of vegetation would have a lower change number.

I realize that this system would change the soil and water numbers of tiles in addition to their vegetation. This is intentional; it represents natural changes in water availability and soil characteristics. The changes are likely to be gradual, so it won’t be a huge shock.

If an ecological province grows too big, it is split into two different provinces. Each province remains unchanged, but they are tracked seperately.

Human activity will influence the terrain. At low levels of habitation and use, the province simply keeps its natural state. But at higher population levels or intensive use of the tile, changes will occur.

Deforestation is the simplest human-induced change. Whenever there is a significant change in human activity on a forest province, that province is put on the hotlist. If the terrain fails two deforestation checks in a row, the forest use is assumed to be sustainable and the terrain is not checked until there is another big change in land use. If two deforestation checks in a row are passed, then the province is put on the countdown list for some number of turns. The timber yield decreases every turn, and at the end of the countdown the terrain changes. The deforestation check depends on the level of land use, the technology levels of your civ, the type of forest, and a random number representing the forest’s ability to renew itself.

After deforestation, the province is put on the hotlist for terrain change. If it is left alone and it borders other forest provinces, the forest could regrow.

Extensive farming can lower the Soil value of the terrain. Soil loss is similar to deforestation; the check works exactly the same way. If the Soil Loss check fails, then the farming is assumed to be sustainable. If it passes, the province is put on the countdown list and the soil value slowly decreases. Technology advancements can prevent this, but often at the cost of pollution.

Pollution from industry, mines, and farms can also impact terrain. The check for this change would be similar to the logging and soil loss checks. There are several types of pollution:

Acid Rain occurs downwind of large industrial centers that produce atmospheric pollution. It is most likely to happen when there is a high Water value, especially when a wind vector crosses a mountain. If the check is passed, the province is put on the countdown list and stays there until the pollution is controlled. When a province is on the countdown list, the Soil or Water Quality rating decreases at some rate that is based on the severity of the pollution. This decrease changes the maximum value, so it is not possible to fix acid rain with fertilizers. After the pollution stops, the maximum Soil or Nutrients value slowly increases to what it once was.

Liquid and solid pollution of all other types has the same effect as acid rain, but the effect is more severe and is localized to the province the pollution originated in.

Human activity can have more subtle effects on the land. Some types of land use, like extensive grazing, can cause an entire ecological province to change form. This check would combine the border change and logging checks.

There are several big environmental changes that can happen naturally. Changes in the activity of the sun can change the climate zones, which would change farm effectiveness everywhere as well as altering a lot of the natural environment. Some places would become uninhabitable, while others would be better places to live. This would prompt a lot of migrations and other upheavals. It should happen at most two or three times per game, but I think that it would be a lot of fun to deal with.

This is not the place to discuss global warming or cooling, but I should note that this model allows global warming to have huge effects. When a global temperature change of a sufficient magnitude hits, the wind and water current vectors would be changed and the climate latitudes would be shifted. A new set of random vectors would be drawn, which would change a lot of things all over the world.

Fishing yields would be important to the economic model. As these are directly related to the Water Quality rating, anything that affects a good fishery would impact the amount of fish caught. This could result in anything from a mild annoyance to widespread starvation.

Fishing itself has an impact of the ocean. To determine if an area is being overfished, a check similar to the logging check would be done on aquatic provinces. If the province passes two Overfishing checks in a row, that province is put on the countdown list and the Water Quality rating of that area decreases at some rate. If the same rate of fishing is maintained, the decrease will accelerate until the fishery collapses. If fishing is stopped for a period of time, then the water quality can increase again as the ecosystem is restored to initial levels.

This change in Water Quality would not affect undersea farms or the base province value. It would only be a temporary adjustment for fishing purposes.

Note that in a free society, stopping fishing would not be an easy process. Also, There would be economic consequences if a large industry was stopped by the government or by the depletion of the fisheries.

Hunting is a problem. I think it should be tracked, as earlier nomadic societies could depend on it. But there isn't an easy way to track wildlife with the current system. Ideas would be appreciated.

In addition to damaging the farms and land, water and air pollution would have a chance of causing disease and unhappiness in the population.

I neglected to mention that water pollution would travel downstream and affect any wetlands provinces below the source of the pollutants. Cities downstream could also have the disease and unhappiness associated with the pollution. If a polluted river crossed the border between two civs, this could have diplomatic consequences. The same is true of acid rain and air pollution.

This model would allow warfare to create long-lasting effects on the land. They still can't farm on World War One battlefields because of the chemicals and shells in the ground. And then there is the Agent Orange option for reducing attrition due to guerilla warfare. I think it would be interesting to see if players would use a chemical like that if it were presented with a different name, or if it had a different purpose like allowing armored divisions to go through dense forests easier.

I forgot about the possibility of taking land from the sea like the Dutch did. If the player wanted to do this, there would have to be a check for the chance of success and the upkeep the land requires. If the land was generated, it would be assigned Soil and Water values. There would always be the chance of the land flooding again if the dikes were not properly maintained (or if they were bombed by a clever enemy). The dikes would have to be treatred as a terrain improvement.

Volcanoes and Earthquakes are the responsibility of the disaster model, not this one. But these would have an impact on the ecology.

The disaster that this model does affect is landslides. Defoliation or other improper use of hilly or mountainous terain would put a tile on the hotlist for landslides. If a landslide does occur, the towns and terrain improvements are all damaged or destroyed. Also, the soil rating of the tile would decrease by one or two.

Deforestation will often result in a lot of soil loss due to erosion. Deforestation would result in soil losses of zero or one on flat terrain, zero to two on rolling terrain, and one to two on hilly terrain. This soil loss would be a permanent change for the province. Also, the province would be assigned a lower change number for border changes, so the surrounding provinces would be more likely to move in to replace the former forest province.

Widespread construction in a province has a chance of reducing that province's soil value by the same amounts as deforestation. Construction in mountainous terrain would reduce the soil value by one to three.

This and other soil loss will affect water quality. Whenever soil loss occurs for any reason other than farming and border changes, any bodies of water in or adjacent to the province will have their Water Quality temporarily reduced by that same amount. So if deforestation reduced the soil of a province by one, and then a landslide reduced the soil value by two, the Water Quality rating of a river running through that province would be temporarily reduced by three.

When you say "permanent" do you mean for the rest of the game? If so I will have to say you are wrong there. Given enough time and left alone (or you can aid this processs with correct tech) a place can usually restore itself, some things more easily restored than others.

I'll try to think about how the model interfaces with the econ system. But please expect that it's going to take awhile, since I'm still really working out the fundamental nuts and bolts for demo 5. I'm only likely going to be able to discuss a few things here, if any, in the context of demo 5 econ model.

So as I see it, with your model, either the AI or the player will basically need to look at every square every turn in order to make the right decisions with respect to irrigation and fertilization. Even if I've mischaracterized this by a factor of 10 too high, I still think it's too much. IMO this is an unwise design decision. I think my economic model with a simple production function approach handles this in a much more efficient manner, if at some cost of inaccuracy.

If I've misunderstood the way you plan to handle it, please let me know.

Sorry it's taken so long to get back to you on the model.

First off, I had thought we wouldn't use such a detailed system for the climate, which is why my disaster model is very simplistic in regards to the climatic disasters. But I'm glad it looks like we will use your model, because it makes doing realistic disasters easier.

I'm glad to hear from you. Is there anything in the model that you think I need to change? If we can make the models work together well, there could be less redundancy in coding and calculations.

If you tell Mark that the climate model will help make the disaster model better, he might give me more clock cycles

quote: Originally posted by Richard Bruns on 05-26-2000 09:33 AM Irrigation, Ferrtilization, and other such things are determined by ecological province, not by individual squares. I figured that there would be at most 200 ecological provinces being farmed.

I have read your suggestion about possibly insisting on coincidence between economic and ecological provinces in the other thread, and I think its not an acceptable solution. IMO it would lead to an increased number of contorted, ugly, provinces, and a great deal of player frustration. Provinces already serve a dual role in organization of economic and governmental activities. I think we need to keep the economic and ecological provinces distinct.

comments:

1. I really like that you are going for the M class planet idea rather than mimicing Earth. I think it will really keep players on there toes with regards to the new way pollution will be handled.

2. I don't see how "true artic" terrain can have soil rating up to 2. IMO it should always be 0, I mean it's ice or permafrost going a few feet down.

3. Same for sub-artic terrain, but not always 0, maybe up to 2.

4. terracing should also help out mountainous terrain, IMO. The Aztecs used terracing in some of the highest (elevation) cities ever built, and these cities were quite large.

5. Why are the warm temp. zones given a -2 to water rating and cool temp. given a +1? I'm no geology major, but this seems strange to me, living in the south where it's always wet and humid.

6. I think you should include a groundwater variable, it shouldn't make much difference in CPU load.

7. I really like the idea of the sun influencing climate. After all it was the sun that caused that "little ice age" you spoke of. Just in case you didn't know, the sunspots occur in 11 year cycles, with decreasing activity causing less heat/light to reach the earth towards the end of the cycle.

8. You talk about how wind and ocean currents will change with pollution. Another thing that changes is more cloud cover, i.e. it gets cooler, with the increased warmth. This is why I was pushing, along with LGJ, for the warming=cooling theory.

9. Your ecology model also leads credibility to doing landslides now. Before, it was just pop. and infrastructure that was affected. Now it's also the land, which can have far more devastating effects.

10. About erosion, LGJ said it would eventually be replenished, but I feel that the replenishment would be far out of the Clash game. It would take thousands of years.

Thanks for the comments. I'll reply in order:

2. Soil of zero means that nothing at all can grow. Soil of one means that lichen and moss can grow, and maybe some small plants. Soil of two means that some grasses and shrubs could make it. In some of the northern parts of Canada and Russia, there is enough of this vegetation for some animals to graze. This could support a small group of people.

I just noticed that I would have allowed taiga to grow on arctic terrain. I should have set the Taiga soil limits from three to six. That is probably what caused your concern; I should have checked that more closely.

3. Siberia is sub-arctic terrain. That terrain supports huge patches of taiga. The blanket of snow insulates the ground and prevents permafrost from forming, so the soil doesn't get really bad. However, you are right that five is too high. Four would be a good upper limit.

4. When I wrote about mountains, I was thinking about places like the Rockies, Alps, and Himalayas. I had forgotten about the Andes. I'll have to think about my treatment of mountains and post a revision of that aspect of the model.

5. As the air over the tropics heats up and rises, it releases the water it is holding. These dry air masses then move toward the poles and fall back down around 30 degrees latitude. As the air sinks, it picks up the water and holds it. This is why almost all of the world's deserts are around this latitude. A similar convection current makes cool temperate areas more rainy.

There are exceptions, like China and the southeast United States. These are mostly due to ocean currents. A look at a world map will show that places in the warm temperate zone are usually deserts.

6. I would also like to do groundwater. I'll see if I can come up with a way to do it without taking up too many clock cycles. I will probably try do it by ecological province, but the problem is that groundwater would have to shift with the provinces. That doesn't make much sense. But it would make more sense than leaving it out entirely...

7. I know about sunspot cycles, but usually they don't have an appreciable effect on things. However, it would make sense for the model to run the solar change check every 11 years.

8. Cloud cover change due to pollution does happen, but clouds are not necessarily the primary factor in temperature change. There are a lot of unknown variables that might be different on a different world. I'd prefer to keep the temperatire changes randomized.