How to develop a thermal map with QGIS and paint.net

Moderators: Uros, OXO, JBr

Post Reply
User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

How to develop a thermal map with QGIS and paint.net

Post by HanSchmid » Tue Feb 26, 2019 11:33 am

Hi everybody,

as I noticed a lot of people have problems to develop a fine and realistic thermalmap. In the mountain region I discovered a lot of strange things which could be easily changed by some tricks with QGIS ©. These effects I will explain below with some schemes.

I will show how i try to develop a thermal map for the german middle range mountain and the alps.

1) Creating a basic underlay

You just have finished your terrain at an early step of the landscape. This picture shows only one tile of the landscape because i did change the colours of the whole landscape.
1-ElevationData-of-a-Landscape.png
A picture form QGIS© after you have finished creating your elevation model.

At the next step you choose raster-> analysis -> DEM (terrain models)...

This is a picture of the German version of QGIS©:
2-Menu.png
A picture of the way how to get to the hillshade menu.

After you have chosen DEM (terrain models)... you should see the following menu:
3-Shading-Menu.png
The shading menu and the recommend settings.

Now i will explain a litte bit:
The Condor 2 © Thermal Modell just defines regions where a thermal will be likely and where not. It is obvious that you will not find any thermals above a huge lake. Things that are not so obvious that you will not find any thermal above the shadow from a mountain. At mountainuous regions you will always find the thermal at sunny side of the mountain and you will although do not find any thermal at the valley. Talk to alpine pilots of you want to know how to find thermals there.
At the Alps the best mountain ridges are the ones where the sun has an azimute of 220 Degrees (South-West at the afternoon). At February the sun is a little lower than 45 degrees, at June it is higher. You should use the azimute and the height of the sun which are typical at 14 o'clock local time for your region.

Do not forget to add -tr 30 30! Or else you will not be able to handle the processed data!

At the next post i will show you what the result is and how you get this usable for a thermalmap.
You do not have the required permissions to view the files attached to this post.
Last edited by HanSchmid on Wed Feb 27, 2019 3:14 pm, edited 2 times in total.
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

Re: How to develop a thermal map with QGIS and paint.net

Post by HanSchmid » Tue Feb 26, 2019 12:11 pm

After you have processed the elevation data you get a hillshade model. It will look like this:
4-hillshade.png
This is a picture of QGIS after processing the hillshade.

You may easily see the white areas where the sun warms the hills and the dark sides where it will be cold. Why is this important? If you set the thermal activity to "high" and the thermal map is considired, you get although strong uplifts in the mountain regions. If the landscape creator has forgotten to define the sunny ridges, you will get strong thermals in the valleys and weak thermals in the mountains. This is physically impossible.

Now you print your thermal map with QGIS print composer. You do the same procedure as you would do with the flightplannermap, because the resolution of the thermalmap is the same as the one of the flightplanner map. One tile of the thermalmap has a resulotion of 256 x 256 px. My Landscape has 20 tiles, this means the thermalmap is a greyscale image with 5120 x 5120 px.
This is the menu of the print composer:
4-hillshade.png
Print composer menu of QGIS.

Now you should save your QGIS-project, because we do not need QGIS anymore.
You open paint.net and have to darken the image a lot. You do this via "Adjustments -> Levels".
This process is a trial and error. So my thermal map had been developped 3 month ago, i do not know anymore how i used so settings.

At first you should save your original thermalmap from QGIS and put the write protection on!
Then you open the image again, do "adjustments -> levels" and check via the colour-picker tool of Paint.net wether the values in the flat area and sunny mountians are correct. I have choosen the area at the lake of constance, because the lake is definitively 100% flat and the surrounding montains have definitifely the maximum value of absorbed sun. If you have chosen the wrong settings, you can close your image und re-open the original thermalmap.
If you got the right settings, you can save the thermal-map. I have chosen the name "thermal-map-corrected.bmp".

With a lot of test-flights I noticed following conditions for realistic thermals:
In flat-land the greyscale should be 130.
In mountains the greyscale should be 160.

This is the picture you should get:
6-Verify-with-paint.net.png
A picture of the basic thermal map at the lake of constance.

How did i get the values?
Thermals in flat areas can not be closer than 2.5 times of the height of the thermal. Means if you have a cloud top at 3000 m, the closest thermal has to be at least 2.5 x 3000 m = 7500 m away. In the mountainous regions the thermals are a lot closer, due to the valleys not having any thermals.

Congratulation! You have a basic and rough thermal map, which can use at your landscape!

The next post i will show you how to implement fields, water and quarry to have although in flat regions a nice thermalmap.
You do not have the required permissions to view the files attached to this post.
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

janjansen
Posts: 1090
Joined: Wed Feb 24, 2016 9:26 pm

Re: How to develop a thermal map with QGIS and paint.net

Post by janjansen » Tue Feb 26, 2019 12:18 pm

please, do correct me if Im wrong, as i could very well be wrong, but doesnt condor thermal model already automatically use mountain slope and solar strength to determine thermal probability outside the flat area's? Im not even sure the thermal map is considered on mountain slopes. Have you tried if it makes a difference?

User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

Re: How to develop a thermal map with QGIS and paint.net

Post by HanSchmid » Tue Feb 26, 2019 12:33 pm

2) Making the basic thermalmap really good

While you are trying to generate the textures from Open-Street-Map, you use a mask with cities, water, quarries and so on.
After you have created a texture-tile sucessfully, you save your paint.net-project (.pdn) with the layers and put write-protection on (so you do not accidently change the texture!).

The first thing I have done is to create a thermalmap from the field map. This is again a trial-and-error process, so i did not want to have thermals too close or too far away. This I have done for the three different kinds of fields I would like to have at my Landscape. After that i created a mask called "thermal-map.pdn" with some layers of values i would like to have.
I recommend:
1) A first layer for water bodies which are black (0,0,0) or real dark.
2) A second layer for hotspots. This are typically cities (not inside valleys of mountain regions!) and quarries. This is 160,160,160 like the mountains.
3) A third layer for real bad conditions. At my landscape I do not use them, typical would be a swampland or a large river body. This is 40,40,40.
4) A fourth layer where i add the field map.

How to create a Field map: You just resize your texture to 256 x 256 px, make it black and white, and use the levels again. I have done this for 3 types of fields, because at my landscape i have to make a difference between fields inside mountain valleys (really bad conditions) and fields outside the mountains (good conditions). If you have a completely flat scenery, you can use only field maps for outside mountian regions.

This is a picture if my thermal-map mask:
7-thermalmap-mask.png
Please notice the different kind of layers.

After you have your mask for the thermal map and the first texture-tile, you can resize your texture-tile to 256x256 px and use the magic wand tool to delete the parts from the mask which you do not need. This is good described at the landscape reference guide.
It takes quite long to create the mask, but using the mask is really quick. I do not need more than 5 Minutes to create One tile for a thermal map.

At this point i would like to quote from the Terragen Landscape Guide. Developping a Landscape is for Years. It is not possible to create a landscape in a few weeks.

At last you see a picture of a generated thermal map, which I am going to use (if I ever get finished :lol: )
8-ThermalMap.png
Original ThermalMap - Tile from the Bavarian Scenery. You may notice that there are only small differences, but it is worth the result.


If you have questions, i try to answer during one week, because i have a lot of things to do.

with best greetings from southern bavaria,

Han Schmid, B.Sc. Meteorology
You do not have the required permissions to view the files attached to this post.
Last edited by HanSchmid on Wed Feb 27, 2019 2:12 pm, edited 1 time in total.
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

Re: How to develop a thermal map with QGIS and paint.net

Post by HanSchmid » Tue Feb 26, 2019 12:44 pm

janjansen wrote:
Tue Feb 26, 2019 12:18 pm
please, do correct me if Im wrong, as i could very well be wrong, but doesnt condor thermal model already automatically use mountain slope and solar strength to determine thermal probability outside the flat area's? Im not even sure the thermal map is considered on mountain slopes. Have you tried if it makes a difference?
Hi Janjansen,

i have tried a lot at the Scenery CW-A, which is my home area.

My results where following:
Acitivity: High
Flats activity: Very low

This leads to realistic conditions like we have here at high pressure. There is no thermal in the valleys and a lot thermal in the mountains.

My results where following:
Acitivity: High
Flats activity: High

This leads to unrealistic conditions, especially at mountains with forests near a city (for example Innsbruck). I guess the thermal-map says there a little thermals above forest and good thermals above cities. Then have a strong thermal above Innsbruck and nearly none at the mountain slopes. This is not correct and i have only seen that kind of thermal at southern france, where the sun shines nearly direct from above. I have nearly 200 hours of flying inside the mountains.

Realistic conditions in cold air, where the acitivty is high at the flat-lands, are real strong updrafts at the mountains with forest and only a little thermal above Innsbruck. If you fly from Innsbruck to Kufstein, and try to find thermals above the city in real-life and not above the mountain slope, you land at Kufstein. Under this conditions I do not enjoy flying, because the thermals above the forests drive me crazy.

At the Simulator you have to fly above the cities, because there the thermal map says there would be good thermals. I have never seen that kind of thermal in real-life. This is not an issue of the Thermal-developping-system of Condor 2, that is better than any meteorologist could develop a model, this is an issue if a good thermal map.


Greetings,

Han
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

User avatar
EDB
Posts: 1575
Joined: Mon Oct 31, 2011 1:17 am
Location: The Netherlands, Europe Continent, Earth Planet, Milky Way Galaxy, Virgo Super Cluster

Re: How to develop a thermal map with QGIS and paint.net

Post by EDB » Tue Feb 26, 2019 12:53 pm

Condor needs an albedo thermalmap. Not a shadowmap.
The "shadowmap" is dynamically done by Condor internally. The thermalmap is static.

For good results the flatland activity slider can become quite important I found out.

Xavier
Posts: 622
Joined: Fri Jun 20, 2008 2:22 pm
Location: France

Re: How to develop a thermal map with QGIS and paint.net

Post by Xavier » Tue Feb 26, 2019 1:08 pm

Hi,

The simplest way to make a fair thermal map is to use the standard flightmapper map made with QGIs according to the Landscape guide texturing chapter.

First its already a the right dimensions and 32 bits data.

Second you can isolate rapidly the different areas by colour selection or other ways for the different settings you want:

Forest, fields, snow, rocks etc..., and then give them a black and white value for each area.

The difficulties with photo satellite sceneries remain with shadows and other artefacts.

The Thermal map works well only with a fair value of Flats activity (ie; Normal and High) and it enhances the ridge lift activity in Mountains.

Example:
From that part of a standard flight mapper map. It is a spring time scenery with an average minimal snow altitude corresponding to end of May.

Ph1.JPG

you can obtain this in half an hour, and then you still have some work for the known spots, and other modification with brightness and contrast, until you get an accurate thermal map:

Ph2.JPG
Not so difficult!
You do not have the required permissions to view the files attached to this post.
*****- Xavier - (XDL - VR FAN) *****

User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

Excursion: Mountain- and Valley Breeze

Post by HanSchmid » Wed Feb 27, 2019 1:10 pm

Hi Xavier,

thank you for showing how to generate faster a thermal map. I did not know that you could do that a lot faster via QGIS. By a given time i will try to correct the thread above and mark the things I changed so other reader of this forum do not get confused.

You are 100 % right that Condor 2 needs an Albedo map. As a said above the thermal developping system of Condor 2 is 100 % correct (like the newton axiomes, there is nothing about to discuss).

In the mountain regions you get other effects which Condor 2 just does not calculate. That does not make the thermal modelling wrong, there is just a part missing. What i will explain next is highly advanced meteorology, a lot of meteorologist do not understand how it works, but every glider pilot know the effects. You can google the alpine pumping, there you get a lot of information.

What i mean is the Mountain- and Valley Breeze, the effect of mountain volumina in the atmosphere and the effect of lower radiation extinction at higher altitudes (you may get easy a sunburn while walking at the top of the mountain, the radiation is much higher). So i have seen you have shown a map from Switzerland you may know the "Maloja" at the Inn-Valley and the breeze at southern france (Vinon, Gap Tallard, etc.).

So this is not a forum for meteorology i will just show some schemes. This is just a suggestion. I may be wrong.

1) Here i have made a map of the known valley-mountain circulation at Slovenia.
The red arrows show the main valley breeze, the yellow arrows the side valley breeze. I have written the typical windspeed at 15.00 local time. You may easily proove that these valley winds are not calculated while setting the wind to 0 km/h and flying around Lesce-Bled. In my opinion it is not necessary to calcute these winds.
1-Valley-Breeze.png
Valley Breeze between Ljubljana and Jesenice at Slovenia.

2) At the next step i made a scheme of the situation at Jesenice between the Karawanken and the Triglav.
You may notice that i have added an albedo map. The mountains have a bad albedo due to forest, at rocks there is a litte better albedo. The worst albedo is at the Triglav mountain (due to snow) and at the Sava river, where no thermals will form. The best albedo is at the City of Jesenice.
2-Topography.png
Topographic map of the Karawanken. I already added the albedo of the surface.

3) At next i will show the expected thermals with the mountain-valley-breeze which you can easily check if you make holiday at Lesce-Bled. I may be wrong, i never was at Lesce-Bled. I just talked to other pilots who have flown there.

The sun warms the mountain slope at the Karawanken and at the Triglav Mountain. An upslope wind will develop. (Small rosa arrows). At the opposite of the slope there is cooling, so the cold air will flow downslope.
At some good points, e.g. the rocks at the mountain slope there will be good thermal early in the day. These is represented with green arrows. These thermal will develop a cumulus cloud over the mountain. Due to this situation, a valley wind will blow from Ljubljana to the Karawanken. Also there will be downdraft inside of the valley. At noon these downdraft will lead to stabilisation of the airmass. At the City the Albedo is quite good so there will be also a thermal. Due to the stabilisation of the airmass this thermal will not reach any height, at most cases there is weak blue thermal. This thermal is represented with red and yellow arrows. The warm air is transported via the valley wind to the Karawanken. There it will lead to an upslope wind.
3-Reality.png
This is the expected situation of the thermal system at the Slovenian Alps.

At the next 2 Post (please wait with the answer until i have written it) i will show how i guees (!) the condor 2 thermal modelling works. I may be wrong.
You do not have the required permissions to view the files attached to this post.
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

RE: How to develop a thermal map - Case Study at Lesce

Post by HanSchmid » Wed Feb 27, 2019 1:30 pm

Test flight with following settings: Activity Normal, Flats Activity Low
91 - Activity Normal - Flats Activity Low.jpg
How i understood the usage of the thermal map with this settings the thermal map is not considered at all. As JanJansan mentioned correctly, Condor 2 does calculate the thermals at the mountains without the thermal map. This is a typical situation which can expect at a warm high pressure weather situation. I used the same colours in the scheme as for the thermal helpers. Notice the green thermals at the mountain and the red / yellow ones which will form over the higher mountains.

Test flight with following settings: Activity Normal, Flats Activity High
92 - Actitvity Normal - Flats Activity High.jpg
If you set the Flats activity to high, the thermal map is full considered. That means, the Condor 2 does calculate the mountain thermal and the albedo. You may notice this looking at the forest at the Karanwanken. There are less thermals, because i guess of the worse albedo of forest. At the next step i will go to extreme (means no fair settings).

Test flight with following settings: Activity High, Flats Activity High
93 - Activity High - Flats Activity High.jpg
As you may notice, there is nearly none thermal at the mountains. I have never seen that while flying in the alps, that does not mean, that it does not exist. It may also be, that this effect is wanted. There are also green thermals at the mountain slope, also they are much weaker. The red thermal develops over the city of Jesenice, under real conditions the valley wind would transport the warm air either to the Karawanken or to the Triglav.

At the next step i will show how i understood the thermal developping system with the used scheme. Please wait with answer, it will be the last post.
You do not have the required permissions to view the files attached to this post.
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

User avatar
HanSchmid
Posts: 18
Joined: Tue Aug 28, 2018 8:46 am

Re: How to develop a thermal map - Scheme

Post by HanSchmid » Wed Feb 27, 2019 2:04 pm

The Screenshots above have lead me to the following conclusion:

4-Condor.png
This would be a scheme of the thermal system with no wind, normal activity and normal flat activity. The sun warms the sunny side of the hills, there is an upslope wind. You can not see this one using the thermal helper, but you may easy check wether this one exists. At parts with good albedo there will form the green thermals. These are the strong mountain thermals which we all like to fly. You can see them using the thermal helper.
At the other side of the mountain there is a downslope wind. As previous mentioned this one develops due to shadow. The Condor 2 model does altough calculate this one, just try to fly form Lesce-Bled. You can not see this one with thermal helper.
Over the city of Jesenice there is the best albedo of the map. This is 100 % true. Over the city there develops a thermal, but due to a missing valley circulation it can rise and form a Cumulus Cloud. You may see this thermal with thermal helper, it is red and yellow.
This situation happens at noon. Then the valley breeze is not or only weak existend and the warm air rises. If you fly in the alps you may know the situation. You can easily check that via looking at satellite pictures. Over Munich at the Flatlands of the Alps you have at 12.00 o'clock huge Cumulus clouds. I guess at Ljubljana it will we be the same.

2) High activity and considering the full albedo of the thermal map.
5-Condor-high.png
If you go to extreme, then the Condor 2 thermal map overlays the thermal developping system at the mountains. Means the effect of the thermal map seems to be higher than the effect of the mountain slope. Than you have strong red/yellow thermal over Jesenice and weak or none red thermal over the forests of the Karawanken.

To to this situation i thought of increasing the thermal map at the mountains.

I did a lot of test flights at my own scenery and i came to the conclusion that the best is to add a hillshade overlay. I tested at the middle-range mountains with high activity where i have a lot of real-life flying experience and nearly know every thermal. Using the way how Xavier described i always got stronger thermals in the flat areas when i set the activity to normal or high.
At my scenery i did not want to have this effect this is why i added an hillshade overlay to the albedo map.

I am sorry for the long explanation. ^^ Like JanJansan I am the opinion that we can easily increase the qualitiy of a landscape. This is just a suggestion from me how we can do it.

Please note, there is huge difference between upslope/downslope wind system at a mountain and the valley breeze system. At southern France the difference is not known because the breeze overlays the upslope/downslope wind. At the northern alps where we have a strong northern valley breeze the breeze is often stronger than the upslope winds. Means if we go lower for example than the height of the Malaupe at Gap, we fall into the valley system and have to land. The upslope/downslope wind system is really good implemented, as i said above better then anyone else could. The valley breeze system is not implemented and i have actually no idea how you could implement it.

with best greetings,
Han Schmid
You do not have the required permissions to view the files attached to this post.
------------------------------------------------------------------------------------------------
Not active in summertime. JK - Han Schmid

User avatar
EDB
Posts: 1575
Joined: Mon Oct 31, 2011 1:17 am
Location: The Netherlands, Europe Continent, Earth Planet, Milky Way Galaxy, Virgo Super Cluster

Re: How to develop a thermal map with QGIS and paint.net

Post by EDB » Wed Feb 27, 2019 2:37 pm

You can ad some height info to the thermal map... That's what I always did to get the thermal-map more real (higher is usually dryer). And that is what Uros also gave as an option. Currently I haven't done that to see how the new Condor 2 weather system works. But Shadows (sun on hill thermals) are dynamically done by Condor.

The flatlands activity is there to correct for situations were the (higher) hills would suck up all the warm air and the flatlands would be blue. Which for example in Nephi would be unnatural (where the higher grounds are cooler and the flat dessert is hot). But if you use flatlands activity high in Slovenia2, it can give unnatural results. So task setters should be aware of that. In Slovenia2 the Forrest's have a quit high (too high?) albedo in the thermalmap, but still the Flatlands Activity High setting is quit noticeable. While in Slovenia2 Flatlands Activity Normal should be Max, for Nephi a new Flatlands Activity Very High setting might be useful... It all depends on the landscape used...

The influence of the flatlands activity also depends on the % of flatlands and hills in a landscape. That makes it more difficult to predict what the result is per landscape.

The problem I see now is that many tasks are set with flatlands activity high because when the setting is lowered, joining seams to take longer...

See also this task : http://www.condor-club.eu/comp/showtask/0/?id=12086
It has flatlands activity high and (general) activity high... and gave blue sky above smaller hills.

User avatar
OXO
Condor Team
Posts: 6360
Joined: Mon Feb 13, 2006 1:08 am
Location: France 42
Contact:

Re: How to develop a thermal map with QGIS and paint.net

Post by OXO » Fri Mar 01, 2019 9:54 am

Chris Wedgwood,
Condor Team

Post Reply