Data from: Plant community-specific greening patterns predict population size increases in a temperate herbivore
Data files
Aug 26, 2024 version files 3.14 MB
-
AllDates.csv
-
AllHabitats_WithDates2022.csv
-
HirtaPop_simplified.csv
-
MxDay1.csv
-
NAO.csv
-
PredictedMonthyl.csv
-
README.md
-
TrainGapsFilled.csv
-
Traits.csv
-
WeatherForAnalysis.csv
Abstract
Climate change driven impacts on vegetation productivity have been shown to drive mammalian herbivore population dynamics in arctic and alpine environments. However, there is less evidence for temperate systems. To address this, we examined the contribution of increasing plant biomass in different vegetation communities (measured by NDVI, normalised difference vegetation index) and winter weather on the observed long-term upward trend in the population of the Soay sheep of Hirta, St Kilda, UK. We found that biomass had increased in all vegetation communities present and done so increased fastest in vegetation types preferred by the sheep. Specifically, those communities with high Specific Leaf Area and Ellenberg’s N, low Leaf Dry Matter Content. Peak summer NDVI and either winter average wind speed or winter North Atlantic Oscillation data added to the variance explained by a simple density dependence model of yearly sheep population growth rates. The highest explanatory power was found for preferred vegetation types including maritime cliff communities dominated by Plantago species, but also for both inaccessible (Rumex acetosa-dominated) or unpreferred (Eriophorum vaginatum- or Sphagnum-dominated) communities where seasonal variation more closely reflects productivity due to minimal grazing. Although the climate is getting windier and wetter, it is also getting warmer allowing increased plant productivity and this appears to be behind the long-term increases in the Soay sheep population. Our study indicates that analysing key vegetation communities may reveal these links better than using landscape-level averages, and that oceanic-temperate systems may show similar climate-driven herbivore population trends to those reported in arctic and alpine systems.
README: Data from: Plant community-specific greening patterns predict population size increases in a temperate herbivore
https://doi.org/10.5061/dryad.ttdz08m72
Description of the data and file structure
The data used in the analysis is a mixture of remote sensing data (Landsat NDVI), meteorological data from weather stations at the study site and from the nearest long-term station, whole island population counts of the Soay sheep on Hirta, vegetation maps, vegetation composition data and plant trait data from databases.
Files and variables
If using the data from the Soay Sheep Project [meteorological data or sheep population data] it is advised that you contact the authors as the project data is continually being added to and checked. The files here are only appropriate for use with this particular analysis.
1. Prediction of Hirta weather prior to installation of weather stations
WeatherAnalysis.R
WeatherForAnalysis.csv
[Column headings for WeatherForAnalysis.csv]
Date: dd/mm/yyyy
Time: years since 01/01/1980
[Three island weather stations Brianans, Quarry and Signals with the following columnheadings]
AirAvgWind: Mean daily wind speed (m s-1)
AirTempAvg: Mean daily air temperature (°C)
AirTempMax: Maximum daily air temperature (°C)
AirTempMin: Minimum daily air temperature (°C)
PrecipTotal: Total daily rainfall (mm)
WindDir: Mean wind direction (degrees)
windMax: Maximum daily wind speed (m s-1)
[Stornoway airport meteorological data column headings]
StornowayRainDaily: Total daily rainfall (mm)
StornowayTempDailyMax: Maximum daily air temperature (°C)
StornowayTempDailyMin: Minimum daily air temperature (°C)
StornowayWindDailyDirDeg: Mean wind direction (degrees)
StornowayWindDailyDirRads: Mean wind direction (radians)
StornowayWindDailyMax: Maximum daily wind speed (knots)
StornowayWindDailySpd: Mean daily wind speed (knots)
NA is used for missing data
2. Accessing Google Earth Engine Landsat Data
GEE_code.txt
[Multipoint locations for the different vegetation types for use in Google Earth Engine]
3. Analysis of NDVI change for each vegetation type
TimeSeries_slopes.R
AllHabitats_WithDates2022.csv [NDVI data per habitat]
AllDates.csv [All potential dates for Landsat image]
Traits.csv [Community Weighted Mean Traits for each vegetation type]
[Column headingsAllHabitats_WithDates2022.csv]
Vegetation: Vegetation type
Day: Day of month (dd)
Month: Month of year (mmm)
Year: Year (yyyy)
Date: Date (dd/mm/yyyy)
Time: Years since 01/01/1985 (numeric)
NDVI: Normalised Difference Vegetation Index (numeric)
CleanedNDVI(>0.15): NDVI values with value less than 0.15 removed and replaced by NA.
NA represents missing data
[Column headings for AllDates.csv]
AllDates: Dates of all Landsat fly overs (dd-mmm-yy)
Time: Years since 01/01/1985 (numeric)
Year: Year (yyyy)
NA represents missing data
[Column headings for Traits.csv]
VegType: Vegetation type
LDMC: Community Weighted Mean Leaf Dry Matter Content (mg g-1)
SLA: Community Weighted Mean Specific Leaf Area (mm2 mg-1)
F: Community Weighted Mean Ellenberg’s moisture indicator (no units)
N: Community Weighted Mean Ellenberg’s nitrogen indicator (no units)
NA represents missing data
4. Calculation of summer maximum NDVI per vegetation type
TimeSeries_MaxMin2.R
AllHabitats_WithDates2022.csv [NDVI data per habitat]
AllDates.csv [All potential dates for Landsat image]
Traits.csv [Community Weighted Mean Traits for each vegetation type]
[Column headings described above]
NA represents missing data
5. Linear modelling of drivers of population change
TimeSeries_maxmincorrelation_cleardays.R
MxDay1.csv [Output from TimeSeries_MaxMin2.R]
HirtaPop_simplified.csv [Sheep population data]
NAO.csv [North Atlantic Oscillation data]
PredictedMonthyl.csv [Output from WeatherAnalysis.R]
TrainGapsFilled.csv [Output from WeatherAnalysis.R]
[Column headings MxDay1.csv]
Index: Line number
Year1: Year (yyyy)
Species: Vegetation Type
NDVI: Predicted maximum summer NDVI
Time: Year including fraction (yyyy.yyy) for maximum NDVI
DecDay: Fraction of year from 01/01 of that year for maximum NDVI
ClearDays: Number of NDVI values for that combination of year and vegetation type
[Column headings HirtaPop_simplified.csv]
HirtaCountYear: Year of population count (yyyy)
Year+1: Year plus one
HirtaTotal: Number of sheep counted during island count
Delta: ln(N(t+1)/Nt) where Nt is the number of sheep in year t
[Column headings NAO.csv]
Year: Year (yyyy)
WinterNAO: Sum of NAO values for December to March
[Column headings PredictedMonthyl.csv]
Index: Row number
[For the three weather station Sg: Signals, Br: Brianans, Qu: Quarry]
AvWD: Monthly average wind speed (m s-1)
SgAvTG: Monthly mean temperature (°C)
SgTtRRadj: Monthly total rainfall (mm)
SgGDDadj: Monthly growing degree days (threshold 5°C)
Year: Year (yyyy)
Month: Numeric 1-12
STAvWD: Stornoway monthly average wind speed (m s-1)
STAvTG: Stornoway monthly mean temperature (°C)
STTtRRadj: Stornoway monthly total rainfall (mm)
STGDDadj: Stornoway monthly growing degree days (threshold 5°C)
STNSWD: North-south wind direction component
STEWWD: East-west wind direction component
Time: Years since 01/01/1985
[Column headings TrainGapsFilled.csv]
Same column headings as PredictedMonthyl.csv
Code/software
All code runs in R version 4.4.0
Code is available at a GitHub repository: https://github.com/RobinPakeman/NDVIpaper.git
Access information
Other publicly accessible locations of the data:
- Github repository: https://github.com/RobinPakeman/NDVIpaper.git
Data was derived from the following sources:
- Landsat imagery via Google Earth Engine
- Meteorological data from the Meteorological Office via the Centre for Environmental Data Analysis
- Plant composition data from Gwynne, D., Milner, C. & Hornung, M. (1974) Vegetation and soils of Hirta. In: Island Survivors: the ecology of the Soay Sheep of St. Kilda. P.A. Jewell, C. Milner, & J. Morton Boyd, eds. London: Athlone Press.
- Plant trait data from LEDA - Kleyer, M., Bekker, R.M., Knevel, I.C., Bakker, J.P., Thompson, K., Sonnenschein, M., Poschlod, P., Van Groenendael, J.M., Klimeš, L., Klimešová, J. & Klotz, S.R.G.M. (2008) The LEDA Traitbase: a database of life‐history traits of the Northwest European flora. Journal of Ecology, 96, 1266-1274.
- Local meteorological and sheep population data from the Soay Sheep Project: https://soaysheep.bio.ed.ac.uk/
Methods
Please see the paper for details. The work described combined whole population census data, meteorological data, and Landsat NDVI data for different vegetation types.