NINAnor / ecosystemCondition

This repository is for documenting the design and calculation of indicators for ecosystem condition in Norway
https://ninanor.github.io/ecosystemCondition/
Creative Commons Attribution 4.0 International
0 stars 4 forks source link

Comments to the workflow #94

Closed anders-kolstad closed 1 year ago

anders-kolstad commented 1 year ago

Marianne Evju:

Relating to the list in the introduction:

Punkt 1: Import Nature type data data set (incl. GRUK) and ANO data set. GRUK-datasettet er i utgangspunktet strukturelt likt ANO-datasettet, men det brukes jo ikke, utover NiN-data, der det allerede inngår i Naturtypedatasettet. Dermed kan det kanskje kuttes ut her?

Punkt 4: spesifiser hvilke NiN-variabler som brukes? Strengt tatt så bør variablene være additive, ikke verste styrer. En lokalitet som har trinn 1 på slitasje (< 1/16) og trinn 1 på kjørespor (</16) , er fort over grenseverdien for trinn 2. Men det kan kanskje løses bedre i neste versjon – gitt også at variablene i NiN får mer meningsfulle trinndelinger. Det bør uansett påpekes.

Jeg skjønner ikke workflowen punktene 8-10, og hvordan naturtype/økosystemtilhørighet er ivaretatt.

anders-kolstad commented 1 year ago

Marianne Evju:

Relating to the start of chapter: Scaled indicator values

Jeg forstår ikke her om økosystem er tatt hensyn til når du skalerer opp til HIA (punkt 2) og hvordan punkt 3 eventuelt tar hensyn til dette. Poenget mitt: innenfor en gitt HIA i en gitt region så vil vi (sannsynligvis) forvente at areal uten slitasje vil være større i våtmark enn i naturlig åpne områder – dermed må informasjonen om økosystem beholdes når det skaleres opp til HIA-nivå. Jeg skjønner faktisk ikke hva punkt 3 betyr her, så et er vanskelig å vurdere hvordan det er tenkt.

anders-kolstad commented 1 year ago
anders-kolstad commented 1 year ago

The one-out all-out principle was used because I had the feeling that there is considerable double-counting going on in the field surveys. E.g. one sees a patch of erosion, and if there is reason to suspect both vehicle-caused damage (kjørespor) and other human-caused damage, then the patch of erosion is attributed equally and fully to both 7TK and 7SE. I judged this because of how often we have the same variable scores for both variables.

But after thinking more about this, I agree with Marianne in that an additive approach is less wrong. We must trust that the field workers have tried to seperate these things, and even though there must be a lot of overlapp between 7TK and 7SE at small spatial scales, a polygon with both variables set to 2 must be in a worse state than if one variable is set to 2 and the other to 1 or zero. I will fix this in the indicator. It will likely mean changing some figures ect downstrean as well.

anders-kolstad commented 1 year ago

Steps 8-10 in the workflow are not easy to explain at this stage, but I added another cross ref to help navigate. I also specified in point 5 that ecosystems are kept in separate data sets. This also anwers Mariannes second post - the three ecosystems are treated in parallel (separately) from the start.

anders-kolstad commented 1 year ago

I changed from a one-out all-out approach to an additive approach in the newest version.