Klimaskovfond

The goal of Klimaskovfond is to generate the needed data sets to get analize the Klimaskovfond score system

Overview of Forest Patches

We have a total of 533,500 forest patches, with sizes ranging from 0.01 to 1,228.918 hectares. The distribution of forest patch areas is visualized below:

Next, let’s focus on forest patches larger than 2 hectares.

Out of the total forest patches, only 30,697 correspond to patches larger than 2 hectares. The distribution of these larger patches is visualized below:

Species Area Relationship Analysis

To determine the optimal forest patch size for biodiversity preservation, we conducted a Species Area Relationship analysis using the sars R package. The analysis utilized GBIF data on species presences within forest polygons from 1999 to 2023, resolving synonyms.

The distribution of species across kingdoms is summarized below:

Kingdom	Percentage
Plantae	52.53
Animalia	28.11
Fungi	19.09
Protozoa	0.25
Chromista	0.02
Bacteria	0.01

Additionally, we present the top 10 classes with the highest proportions:

kingdom	phylum	class	Percentage	Cumulative_Percentage
Plantae	Tracheophyta	Magnoliopsida	33.60	33.60
Animalia	Chordata	Aves	17.12	50.72
Fungi	Basidiomycota	Agaricomycetes	15.41	66.13
Plantae	Tracheophyta	Liliopsida	13.98	80.11
Animalia	Arthropoda	Insecta	8.27	88.38
Plantae	Tracheophyta	Polypodiopsida	1.82	90.20
Plantae	Bryophyta	Bryopsida	1.60	91.80
Animalia	Chordata	Mammalia	1.05	92.85
Fungi	Ascomycota	Sordariomycetes	1.01	93.86
Plantae	Tracheophyta	Pinopsida	0.92	94.78

Species Area Relationship for Different Groups

Now, we perform a species area relationship analysis for three groups: plants and animals together, plants only, and animals only.

Plants and Animals

Here we can see the table of selected models:

Model	Weight	AIC	R2	R2a	Shape	Asymptote	CumWeight
ratio	1	239029.7	0.430	0.430	convex up	TRUE	1
asymp	0	239054.0	0.430	0.429	convex up	TRUE	1
p1	0	239066.5	0.429	0.429	convex up	FALSE	1
weibull4	0	239078.7	0.429	0.429	convex up	FALSE	1
p2	0	239079.2	0.429	0.429	sigmoid	FALSE	1
weibull3	0	239079.4	0.429	0.429	convex up	FALSE	1
heleg	0	239081.3	0.429	0.429	convex up	FALSE	1
betap	0	239083.8	0.429	0.429	convex up	FALSE	1
powerR	0	239087.6	0.429	0.429	convex up	FALSE	1
power	0	239093.5	0.428	0.428	convex up	FALSE	1
epm1	0	239095.3	0.428	0.428	convex up	FALSE	1
epm2	0	239095.4	0.428	0.428	convex up	FALSE	1
koba	0	239589.8	0.415	0.415	convex up	FALSE	1
monod	0	239791.6	0.409	0.409	convex up	TRUE	1
gompertz	0	239845.9	0.407	0.407	sigmoid	TRUE	1
negexpo	0	240025.4	0.402	0.402	convex up	TRUE	1
linear	0	240385.6	0.392	0.392	linear	FALSE	1
logistic	0	240447.2	0.390	0.390	sigmoid	TRUE	1
loga	0	243228.7	0.302	0.302	convex up	FALSE	1
chapman	0	250693.8	0.000	0.000	linear	TRUE	1

and the plot of the relationship

Plants only

Now we fit the same model but for plants only

Here we can see the table of selected models:

Model	Weight	AIC	R2	R2a	Shape	Asymptote	CumWeight
p2	0.958	173087.1	0.339	0.339	sigmoid	FALSE	0.958
powerR	0.042	173093.3	0.339	0.339	convex up	FALSE	1.000
ratio	0.000	173112.6	0.338	0.338	convex up	TRUE	1.000
epm1	0.000	173124.7	0.338	0.338	convex up	FALSE	1.000
epm2	0.000	173132.3	0.338	0.337	convex up	FALSE	1.000
power	0.000	173224.9	0.334	0.334	convex up	FALSE	1.000
p1	0.000	173226.9	0.334	0.334	convex up	FALSE	1.000
heleg	0.000	173226.9	0.334	0.334	convex up	FALSE	1.000
weibull3	0.000	173226.9	0.334	0.334	convex up	FALSE	1.000
weibull4	0.000	173228.9	0.334	0.334	convex up	FALSE	1.000
betap	0.000	173295.5	0.331	0.331	convex up	FALSE	1.000
asymp	0.000	173356.6	0.328	0.328	convex up	FALSE	1.000
gompertz	0.000	173431.7	0.325	0.325	sigmoid	TRUE	1.000
logistic	0.000	173674.7	0.315	0.315	sigmoid	TRUE	1.000
linear	0.000	173969.8	0.302	0.302	linear	FALSE	1.000
koba	0.000	174106.3	0.296	0.296	convex up	FALSE	1.000
loga	0.000	174980.7	0.257	0.257	convex up	FALSE	1.000
monod	0.000	174693.1	0.270	0.270	convex up	TRUE	1.000
negexpo	0.000	175213.5	0.247	0.247	convex up	TRUE	1.000
chapman	0.000	179804.3	0.000	0.000	sigmoid	TRUE	1.000

and the plot of the model

Animals only

Now we fit the same model but for Animals only

Here we can see the table of selected models:

Model	Weight	AIC	R2	R2a	Shape	Asymptote	CumWeight
weibull4	1	176603.6	0.299	0.299	convex up	TRUE	1
ratio	0	176646.0	0.297	0.297	convex up	TRUE	1
p1	0	176651.3	0.297	0.296	convex up	FALSE	1
weibull3	0	176659.7	0.296	0.296	convex up	TRUE	1
betap	0	176662.7	0.296	0.296	convex up	TRUE	1
heleg	0	176662.8	0.296	0.296	convex up	FALSE	1
epm1	0	176678.6	0.295	0.295	convex up	FALSE	1
epm2	0	176682.0	0.295	0.295	sigmoid	FALSE	1
power	0	176682.4	0.295	0.295	convex up	FALSE	1
p2	0	176684.3	0.295	0.295	convex up	FALSE	1
powerR	0	176684.4	0.295	0.295	convex up	FALSE	1
koba	0	176756.5	0.292	0.292	convex up	FALSE	1
monod	0	176782.0	0.291	0.291	convex up	TRUE	1
negexpo	0	176810.6	0.290	0.289	convex up	TRUE	1
asymp	0	176840.1	0.288	0.288	convex up	FALSE	1
gompertz	0	177038.7	0.280	0.280	sigmoid	TRUE	1
linear	0	177081.5	0.278	0.278	linear	FALSE	1
logistic	0	177368.9	0.265	0.265	sigmoid	TRUE	1
loga	0	178774.3	0.199	0.199	convex up	FALSE	1
chapman	0	182390.6	0.000	0.000	sigmoid	TRUE	1

and the plot of the model

Buffer and intersection generation

Tasks needed

• Existing nature (A), add a 5 ha buffer, add jakob’s Forest dataset
• A = Existing nature
• B = Jakobs Forest or Fredskov
• Buffer ~5 Ha (225m) around B butcannot be part of A, and has to be a part of agriculture
• For every buffer patch we calculate the contiguity with B and A patches and calculate total area and proportion of forest within A and B and add Lavbund proportion

Read A and plot

Existing Nature (A) and Forest (B) Datasets

We begin by loading the existing nature dataset (A) and the dataset representing deciduous forest areas from Jakob Assmann’s work, which we’ll refer to as B.

Agricultural Land Dataset

Next, we acquire the dataset that represents agricultural land in Denmark. This information is crucial, as we want to ensure that the newly created forested areas are restricted to the existing agricultural land.

Generating a Buffer around Deciduous Forest

We generate a 225-meter buffer around the deciduous forest areas, approximating a 5-hectare squared region.

This script creates a buffer around deciduous forest areas, ensuring it conforms to the specified conditions and constraints. The resulting buffer is then saved as a Cloud-Optimized GeoTIFF (COG) file named “Buffer_all_225.tif.”

We can now visualize all this categories

Now in order to calculate areas and adjacencies the raster will be transformed into polygons

Now to actually calculate the values we will first unite resolve for A and B join them in the largest possible polygons

Now we filter form the joint polygons only the ones with areas higher than 200 ha, 100 Ha, 50 Ha and 25 Ha

Now we go one by one and we generate the potential forest content

200 ha

And now we transform this into polygons

We now we process the potential forest to add total area considering AB (Total_Area), the area of the potential forest is included (Ha), we caclulate the area of forest considering the adjacent A and B areas, Forest_Area

#> Error in PotentialForest$Total_Ha[i] <- PotentialForest$Ha[i] + AB_200[Temp$to_id,  : 
#>   replacement has length zero

100 ha

And now we transform this into polygons