Geo Realism: Discussion on a new SDK based map generator

[Laskaris]

@Laskaris

Well lets imagine that its like Spore of Sim Earth. Where you can roast or freeze you planet in various ways. At some point in he game you will have god-like powers and could terraform however you wanted. Thus knowing a realistic approach to how it would chnage would be useful for that aspect of the game.

Well, with god-like powers, you could:

1. Make a plot cooler or warmer
2. Make a plot drier or wetter
3. Change the composition of the soil at a molecular level

Which is all you need to do whatever you want in terms of terrain, really.

Some of the most agriculturally productive regions that occur naturally on Earth are mollisols, i.e. the black earth soils that exist in temperate grassland regions. So, if I were a terraformer, I would try to first get there, and then through making the soil even more fertile with my god-like powers, I would get "Lush" terrain.

Polar Desert or Ice -> increase temperature and precipitation -> Tundra -> increase temperature -> Prairie -> increase soil fertility -> Black Earth Grassland -> increase soil fertility -> Lush

Boreal Forest -> increase temperature -> Seasonal Forest -> decrease precipitation -> Prairie -> the rest of the chain same as above

Tropical Rainforest -> decrease precipitation -> Evergreen Forest -> decrease precipitation and temperature -> Seasonal Forest -> decrease precipitation -> Prairie -> the rest of the chain same as above

(Savanna) Grassland -> decrease temperature -> Chaparall -> decrease temperature -> Prairie -> the rest of the chain same as above

Desert -> increase precipitation -> Scrub -> increase precipitation -> Prairie -> the rest of the chain same as above

This is just a rough idea I came up with in five minutes. But basically, I think it could go something like this, realistically - as far as we can be "realistic" about technology of the far future!

 

[primem0ver]

I would also like to add here that "Lush" as it stands is not the ideal terrain. It is an ideal terrain for food. Currently the stats for black soil and volcanic terrain will outdo the C2C stats for Lush. So do we want to make Lush more useful in other stats as well?

 

[Faustmouse]

I would also like to add here that "Lush" as it stands is not the ideal terrain. It is an ideal terrain for food. Currently the stats for black soil and volcanic terrain will outdo the C2C stats for Lush. So do we want to make Lush more useful in other stats as well?

I think it also should give more some more commerce than it does now. About more hammers, I'm not sure. Maybe add a new forrest that could only be planted on Lush and make it very good in production.



Some questions:

1) how hard is it to make soil more fertile after all?

2) With "unlimited" energy (Fusion Power), would it be possible to build something like a giant radiator in tundra and such and heat it to a moderare temperature?

3) Do you plan to give tropical, normal and polar cost and ocean different stats? If so, do you want to make them terraformable? I think this might have BIG influence on the global climate.

4) Someone mentioned change peaks and hills to flat terrain. I think I've read a while ago that this is actually in progress in japan? So it wouldn't be that far away in the techtree.

 

[Hydromancerx]

I would also like to add here that "Lush" as it stands is not the ideal terrain. It is an ideal terrain for food. Currently the stats for black soil and volcanic terrain will outdo the C2C stats for Lush. So do we want to make Lush more useful in other stats as well?

Possibly if if its not going to be a naturally occurring terrain anymore. Because I had to make it balanced with the other terrains. But if you have to work up to it, then it should have some realty good benefits.

 

[strategyonly]

Here is something about winter storms, for something like that??

http://forums.civfanatics.com/showpost.php?p=12173558&postcount=1567

 

[ls612]

Here is something about winter storms, for something like that??

http://forums.civfanatics.com/showpost.php?p=12173558&postcount=1567

Looks to me like concept code for a work in progress mod. I'm not certain what exactly he will add in that mod.

 

[strategyonly]

Can we pls get an update on this, thx . . . SO

 

[primem0ver]

As mentioned here, I am hoping to get the climate simulator done in time for v30.

 

[primem0ver]

Folks, I just thought I would update you. I have reached a problem in the development of the GeoRealism Engine; one which I have tried for more than 2 days to figure out. Unfortunately I have not found a solution.

I am trying to find a way to establish (calculate) the base temperature for a plot that works for both summer and winter (highs and lows). I can use a fairly simple equation for one or the other but there is no equation that I have been able to find online that can account for both.

I don't expect perfection. I expect these values to be off since weather simulation has not occurred yet, but I was hoping to find a way to a way to at least get a ballpark figure that works for all general points on the globe. Yet, I cannot seem to do so. None of the combinations of variables I have tried so far have been successful. Until this matter is resolved, it will be difficult for me to continue with the engine since I am down to the nit-picky details to make sure everything is in order.

 

[God-Emperor]

Is there some particular reason you don't just use the one that's good for getting the high to get the high and the one that's good for getting the low to get the low? Why try to use 1 equation to do 2 things when you already have 2 equations to do them?

 

[primem0ver]

Is there some particular reason you don't just use the one that's good for getting the high to get the high and the one that's good for getting the low to get the low? Why try to use 1 equation to do 2 things when you already have 2 equations to do them?

I wish it were that simple and I do invite your ideas to get me thinking. I would consider this if it was. However, I wrote separate a separate program to try and figure out a solution (shown below). Unfortunately the equation itself doesn't know if the outcome will be high or low a priori (beforehand). I must come up with a solution that calculates the temperature of a plot with reasonable accuracy for every month of the year.

The only reason I know what the high and low temperatures are is that the program calculates the supposed temperature for every day of the year and two variables keep track of the highest and lowest values that result, marking the julian date and resulting calendar date they occur on:

The four large statistics boxes contain the variables that will change depending on a location. The other up-down box values are the variables for which I am trying to find the right setting that works for all calculations.

This particular test shows a nearly perfect set of results for the Death Valley Desert in California (both hot and cold). However, the same up-down box values for another location will not be correct. By changing the necessary variables, I can sometimes get two locations to work. But then a third will be completely off for either hot, cold, or both.

EDIT: Case in point. Below are the results for an equatorial location in a tropical rainforest. These results are good as well. And notice I only changed one up-down variable: the "COOL" one. This would be doable by making the COOL constant slightly variable, varying with humidity. This gives me two data points that work with this set of variables and constants. However, using this set for Chicago and the resulting temperatures are way too hot in the summer and way too cold in the winter...

This also shows why I cannot use two equations. Notice that the hottest days are now in the spring and fall and the coldest days are in both winter and summer. (Only one result is shown since I only use two variables). This is correct since the hottest month is when the sun is highest in the sky. On the equator this happens around March 21st and Sept 21st (spring and fall equinoxes).

 

[Thunderbrd]

Death Valley's annual high is nowhere near realistic though... should be around 120f

Perhaps one thing that's throwing the model a bit is that there would be some temperature diffusion from one plot to another? Additionally, you probably have a fair model except that weather patterns will spike the results a bit more severely in a real world region. Thus you see somewhat more extreme results, so what you'd have to include would be a random % modifier variable taking place on each day that would represent the varying weather patterns.

(Maybe you've considered this though so I'm just throwing out ideas...)

 

[primem0ver]

Death Valley's annual high is nowhere near realistic though... should be around 120f

Perhaps one thing that's throwing the model a bit is that there would be some temperature diffusion from one plot to another? Additionally, you probably have a fair model except that weather patterns will spike the results a bit more severely in a real world region. Thus you see somewhat more extreme results, so what you'd have to include would be a random % modifier variable taking place on each day that would represent the varying weather patterns.

(Maybe you've considered this though so I'm just throwing out ideas...)

It is almost realistic. I am looking for average daily temperature. Not the highest temperature reached during the day versus the coldest temperature at night. The highest average daily temperature at Cow Creek Station is 102.1 Fahrenheit/38.9 Celsius (July). The lowest daily average temperature at Cow Creek Station is what I have wrong... so you are right I do need to play with the values a bit. I was remembering it during testing as 4.4 degrees Celsius but it is actually 11.4 (January).

 

[ls612]

Folks, I just thought I would update you. I have reached a problem in the development of the GeoRealism Engine; one which I have tried for more than 2 days to figure out. Unfortunately I have not found a solution.

I am trying to find a way to establish (calculate) the base temperature for a plot that works for both summer and winter (highs and lows). I can use a fairly simple equation for one or the other but there is no equation that I have been able to find online that can account for both.

I don't expect perfection. I expect these values to be off since weather simulation has not occurred yet, but I was hoping to find a way to a way to at least get a ballpark figure that works for all general points on the globe. Yet, I cannot seem to do so. None of the combinations of variables I have tried so far have been successful. Until this matter is resolved, it will be difficult for me to continue with the engine since I am down to the nit-picky details to make sure everything is in order.

I am certainly not a climate or an earth scientist, so I can't really help with scientific answers. What I can suggest is that for plots that your model is handling anomolously you could have the engine catch the bad numbers and replace them with a statistical normalization of nearby plots with similar conditions that have reasonable numbers. It isn't perfect, not by a long shot, but for a V1 it could be workable.

 

[johny smith]

I am not a climatologist, but generally I thought the variance is higher the further inland. The wind would greatly effect the temperature depending on whether is a warm or cold wind. I do not know how you are taking account for mountains blocking fronts and etc for the equation.

 

[Taxman66]

Is there a known standard deviation of average tempature? Or perhaps one based on terrain or logestical/longitudital location?

I ask because maybe it would be easier to get the average tempature and if needed you can find the near-max and near-min based off the deviation.

 

[johny smith]

Just for generalization. There is a constant effect of solar radiation on the earth that could be measured in latitude for general temperature, and there is a standard for the distance from the oceans for difference in seasons. More land equals larger difference in winter and summer due to the oceans better regulating seasons. Anyway that was what my basic physical geography course taught me.

 

[primem0ver]

I appreciate you guys trying to help.

I am not a climatologist, but generally I thought the variance is higher the further inland. The wind would greatly effect the temperature depending on whether is a warm or cold wind. I do not know how you are taking account for mountains blocking fronts and etc for the equation.

This is true for the most part. Temperature variance is affected by several things but the biggest is probably humidity (which decreases as you go inland).

Is there a known standard deviation of average tempature? Or perhaps one based on terrain or logestical/longitudital location?

I ask because maybe it would be easier to get the average tempature and if needed you can find the near-max and near-min based off the deviation.

Yes. And my current model is based on this. However, the model does not hold when taking into account the difference between seasons. This is where I am running into a problem.

Just for generalization. There is a constant effect of solar radiation on the earth that could be measured in latitude for general temperature, and there is a standard for the distance from the oceans for difference in seasons. More land equals larger difference in winter and summer due to the oceans better regulating seasons. Anyway that was what my basic physical geography course taught me.

This is true. And my model also takes this into consideration. However, the variables that affect solar radiation and more importantly, how those variables affect the incoming solar radiation (and its bleedoff during the night) is not altogether clear in the articles I had read up to the time I posted my issue.

However, I have recently come across an article that goes into detail about those issues and I think it may help me come up with a solution. The article is specific to China but I think I can extrapolate. Keep your fingers crossed.

 

[johny smith]

This is true for the most part. Temperature variance is affected by several things but the biggest is probably humidity (which decreases as you go inland).

Are you factoring in humidity to the equation to determine temperature. Or just looking for an independent equation not affected by humidity?

This is true. And my model also takes this into consideration. However, the variables that affect solar radiation and more importantly, how those variables affect the incoming solar radiation (and its bleedoff during the night) is not altogether clear in the articles I had read up to the time I posted my issue.

Why does the bleed off matter at night? I mean are you going to have day and night? Cloud cover I thought would prevent bleed off more than anything else. Basically I would think wetter areas have less variation. It would be better to just go with humidity versus inland. I would think in and out solar radiation would be mostly effected by humidity.

I know there is a constant as well that is from the capturing of the solar radiation that changes near the equator and etc, but don't remember exactly right now how it works. I am looking for something to post that shows the effect.

If you are factoring in all of the various factors of course it will become more complex, but I thought something simple is better than nothing. If you already are at the point that you have many other factors it would be a shame just to generalization the seasonal temperature difference without including other factors.

 

[johny smith]

I found one image online.

Spoiler :

These are photos from my physical geography book.

Spoiler :

Spoiler :

Spoiler :

Spoiler :

Spoiler :

Anyway crappy quality just thought something quick that at least shows how my book claims the energy is used.