Replies: 6 comments 2 replies
-
Nice having some user reading the doc of the functions since it is clearly not updated... The parameter you have identified does not exist in the code. Rather look at the pervious / impervious definitions for UTRF and LCZ (respectively
Same for this one. However, for this one you will see that building is still not inside the definition. The reason is actually quite simple, from what I remember only surfaces not covered by buildings are considered as pervious or impervious (thus the sum of pervious + impervious + building + undefined = 1). We kept this definition of calculation. Then the building is an other part of the calculation (let's say it can be covered by a green roof...). |
Beta Was this translation helpful? Give feedback.
-
Thank you for your explanation.
Ok, I'm trying to sum existing fields (
Having the same approach with grid_indicators, I try to sum everything, and few units have a total of 1 : "WATER_FRACTION" + "BUILDING_FRACTION" + "HIGH_VEGETATION_FRACTION" + "LOW_VEGETATION_FRACTION" + "ROAD_FRACTION" + "IMPERVIOUS_FRACTION" I guess this is because of data model / availability. Thanks again ! |
Beta Was this translation helpful? Give feedback.
-
No need to reply to that one, I got the answer myself ! |
Beta Was this translation helpful? Give feedback.
-
Anyway share the answer for everyone and we can correct if not right ! =) |
Beta Was this translation helpful? Give feedback.
-
I suppose you say pervious, if so yes.
This is the right interpretation and yes no duplicate
We are waiting for your guess =)
I do not understand why summing lcz and urb impervious fraction. I forgot to remind that but rsu indicators should be "well defined" in the wiki (respectively https://github.com/orbisgis/geoclimate/wiki/RSU-indicators#impervious_fraction_lcz and https://github.com/orbisgis/geoclimate/wiki/RSU-indicators#impervious_fraction_urb). Even though everything is maybe not clear, good to have your opinion.
It is definitely, and as always in this case, I wonder why we did not think about that. We should definitely do that. I may set this issue as discussion and create a specific issue for that point.
Yes it is probably the main reason. An other reason would be that you also have the superimpositions which are not considered in the sum you have performed (high_vegetation and most of the other layers). The fact we leave the superimposition information as "raw data" might not help the user at the first reading but then it is a way to leave him be the one who choose what to do with the superimposed layers. Never know if it is the best option but it is the one we chose.
|
Beta Was this translation helpful? Give feedback.
-
Concerning impervious fractions in the BDTOPO, according to what I read in the code, I think you are right, the following layers are used: CONSTRUCTION_SURFACIQUE, PISTE_AERODROME, TERRAIN_SPORT, SURFACE_ROUTE, SURFACE_ACTIVITE (@ebocher correct if not right)
Actually these indicators are dedicated to different uses (LCZ are used for Local Climate Zones classification while URB are used for the Urban Typology by Random Forest - UTRF - classification). For now they are defined the same way but this definition may change if we get more detailed informations or a feeling that there is a need to better dissociate urban layers for the UTRF classification.
In the LCZ classification, pervious surfaces are needed so we define a way to have this fraction. In reality, I think that water and vegetation for example should not be summed under a same "pervious" label since they are really different, this is why we did not keep the pervious fraction indicator in the UTRF classification (vegetation and water fractions are used but individually).
Yes
Good idea, this is something which can be performed. I add this information to the issue about undefined_fraction. |
Beta Was this translation helpful? Give feedback.
-
Hello,
I'm bit surprised with some choices being made to compute pervious / impervious fraction. Looking at code source, I was able to understand some results which appeared a bit weired to me :
fractionTypePervious
. I don't really get why ? It does not appear that obvious to me. I would say that isolated trees are not pervious but others are. And working with BDTOPO (which does not have isolated tree), it should be by default considered as pervious.fractionTypeImpervious
, I can see that by default, building is not taken into account. I'm also wondering why !?I don't know if this can and should be changed by default and if it has any incidences on others indicators like LCZ/UTRF. If not, changing them could be nice. If it has, would it be possible to have 2 more alternatives attributes taking this into account :
["low_vegetation", "high_vegetation", "water"]
["road", "building"]
Maybe, I'm too fast, and your explanations will convinced me. I'm really curious.
Thanks !
Beta Was this translation helpful? Give feedback.
All reactions