This repository contains four zipped data files which contain (i) the spatial distribution of aapa mire complexes (‘aapa mires’) and their wettest flark-dominated parts (‘wet aapa mires’) situated in the aapa mire and palsa mire zones of Finland, as selected for the study by Heikkinen et al. (in review), (ii) values for the six bioclimatic variables (growing degree days, mean January and July temperature, annual precipitation, and May and July water balance) averaged for the years 1981–2010, and developed for the studied aapa mires and wet aapa mires using a 50 x 50 m lattice system, and (iii) values for the same six bioclimatic variables developed for future climates and the two types of study mires, based on the global climate models for 2040–2069 and two Representative Concentration Pathways (RCP4.5 and RCP8.5), and (iv) values of climate velocity metrics calculated for the six bioclimatic variables and the two types of study mires. These data provide the essential data employed in conducting the analysis in the following work: Risto K. Heikkinen1, Kaisu Aapala1, Niko Leikola1 and Juha Aalto2: Exposure of boreal aapa mires to climate change, in review. 1 Biodiversity Centre, Finnish Environment Institute, Latokartanonkaari 11, FI-00790 Helsinki, Finland 2 Finnish Meteorological Institute, Weather and climate change impact research, Helsinki, Finland The data files are embedded in four compressed zip files (one of them including a geodatabase folder with files) which include several ArcGIS compatible tiff-raster or shape files. The names and contents of the four zipped files are as follows: (1) mires.zip – includes shape files describing the location and spatial configuration of the aapa mires (‘Aapa_mires.shp’) and the wet aapa mires (‘Wet_aapa_mires.shp’) included in the study, and the borders of different mire zones in Finland (‘Mire_zones.shp’); (2) climate_data_aapa_mires.zip – includes 18 tiff raster files showing the values of the six bioclimatic variables in the studied aapa mires within the 50 x 50 m resolution grid. The data in this zipped file include climate data averaged for the years 1981 – 2010 and for the future time slice of 2040–2069 and two Representative Concentration Pathways (RCP4.5 and RCP8.5); (3) climate_data_wet_aapa_mires.zip – includes 18 tiff raster files showing the values of the six bioclimatic variables in the studied wet aapa mires within the 50 x 50 m resolution grid. Similarly as in (2), the data in this zipped file include climate data averaged for the years 1981 – 2010 and for the future time slice of 2040–2069 and two Representative Concentration Pathways (RCP4.5 and RCP8.5); (4) velocity_data_for_mires.zip – includes zipped geodatabase folder velocity_open_mires.gdb which, in turn, includes spatial ArcGIS surfaces for the climate change velocity metric calculated for all the six bioclimatic variables, and the two types of mires and the two RCPs. In the zipped files (2) and (3), first part of the names of the included files refer to one of the six bioclimatic variables as follows: GDD5 – growing degree days, PREC – annual precipitation, TEMP_Jan – mean January temperature, TEMP_July – mean July temperature, WAB_May – May water balance, WAB_July – July water balance; and the remaining part of the name indicates the time period, type of the RCP and that of the mire. It should be noted that these data are embargoed until the end of the SUMI project for which they were developed, i.e. 1.1.2023. The coordinate system for the data files is: ETRS-TM35FIN (EPSG: 3067) (or YKJ Finland/Finnish Uniform Coordinate System (EPSG: 2393)). Summarization of the key settings of the study is provided below. A detailed treatment is included in the manuscript Heikkinen et al. (in review). Once the manuscript is accepted for publication an updated link will be provided. Study system: Aapa mires are waterlogged, peat-accumulating EU Habitats Directive priority habitats whose ecological conditions and biodiversity values may be jeopardized by climate change. Aapa mires depend on the surface water flows from the surroundings which makes them sensitive to hydrological alterations and falling water tables caused by land use (ditching for peatland drainage) as well as climate change (Gong et al. 2012, Sallinen et al. 2019). This sensitivity of aapa mires and their biodiversity to increasing temperatures and decreasing water balance and precipitation can be of particular concern as they occur in northern hemisphere, in areas where the largest climatic changes are projected to take place (AMAP 2017, Väliranta et al. 2017. Kolari et al. 2021). In the study by Heikkinen et al. (in review), we assess the climate exposure of these habitats by developing velocity metrics for both the aapa mire complexes (‘aapa mires’) and their wettest flark-dominated parts (‘wet aapa mires’) in Finland. Aapa mire data: Occurrences of aapa mires were identified from the CORINE CLC2018 land cover data which is available in Finland as a 20 x 20 m resolution raster data, by focusing on the CORINE category 4121 (‘Peatbogs’) which includes various open mires occurring in aapa mire and palsa mire zones, as well as in raised bogs zones. We excluded open mires occurring in the raised bogs zone but included CORINE Peatbog occurrences both from the aapa mire and palsa mire zones. This opted for this decision because open mires in aapa and palsa mire zones share several matching ecological features, and because palsa mires may provide suitable habitats for aapa mire species under warming climate. The adjacent peatbog 20-m pixels in the aapa and palsa mire zones were merged and converted into contiguous peatland polygons. From these, polygons smaller than 10 ha in size were excluded because typically they show only limited number of ecological elements central to the representative aapa mires. These selected ≥10 ha peatland polygons formed the first study mire dataset, aapa mire complexes, or ‘aapa mires’ in short (i.e., the whole aapa mire ecosystem containing all embedded mire habitats therein). The second study mire dataset was constrained to include only the wettest parts of aapa mire complexes characterized by flarks, i.e., open water pools, referred here simply as ‘wet aapa mires’. These wet aapa mire occurrences are typically smaller than the whole aapa mire complexes and occur more sparsely in the landscape. Thus, the climatic exposure of wet aapa mires can be expected to be greater than that of aapa mire complexes. This will very likely cause elevated climate change adaptation challenges for habitat specialist species that require open water or permanently wet environments. The spatial data for the wet aapa mires were determined with the help of the topographic database developed by the National Land Survey of Finland (NLS), and the land cover class ‘Swamps classified as difficult, dangerous and impossible to cross’ therein. Climate data: In the first phase, monthly average air temperature data for 1981–2010 were constructed at the 50 x 50 m spatial resolution across Finland, as described in Aalto et al. (2017) and Heikkinen et al. (2020, 2021). This was done by modelling the weather station data from 313 Fennoscandian stations together with variables of geographical location, local topography and water cover. Monthly precipitation data were developed by fitting kriging interpolation method to the data on 343 rain gauges, and the data on geographical location, topography and proximity to the sea. Based on the monthly temperature and precipitation data, six bioclimatic variables describing key ecological winter- and summer-time conditions for aapa mire ecosystems were calculated (cf. Parviainen and Luoto 2007, Ruuhijärvi 1988, Rydin and Jeglum 2006): (1) annual temperature sum above the base temperature of 5 °C (growing degree days, GDD5), (2) mean January temperature, (3) mean July temperature, (4) monthly climatic water balance calculated for May and (5) for July, and (6) annual precipitation sum. The two climatic water balance variables were calculated as the difference between the May - or July - total precipitation sum and the potential evapotranspiration (PET) in the corresponding month following Skov and Svenning (2004). In the second step, the data based on an ensemble of 23 global climate models from the Coupled Model Intercomparison Project (CMIP5) archives (Taylor et al. 2012) were employed to develop future climate surfaces averaged for the years 2040–2069 and the two Representative Concentration Pathways (RCP4.5 and RCP8.5). The monthly air temperature and precipitation data in these climate surfaces were interpolated to match the 50 × 50 m grid, then the change predicted by the GCMs was added to the 1981–2010 climate data, and finally, the values for the six bioclimatic variables were recalculated for the 50-m resolution grid across the whole Finland. In the third step, all the developed climate surface datasets were intersected by the spatial datasets of the two differently delimited aapa mire networks, i.e. ‘aapa mires’ and ‘wet aapa mires’. This allowed calculation of the climate change velocity metrics separately for the two types of aapa mires, namely, for both mire datasets by measuring the distance between climatically similar 50-m grid cells in the present and future climates by considering only locations with either (i) aapa mires, or (ii) wet aapa mires. Thus, matrix areas providing unsuitable habitat for aapa mire biodiversity were excluded and for both types of mires the distance from the present-day mire cell was linked to the nearest corresponding mire cell with similar future climatic conditions. The climate data for the years 1981 – 2010 and the future time slice of 2040–2069 and the two Representative Concentration Pathways (RCP4.5 and RCP8.5), clipped to the networks of the two types of aapa mires for all the six bioclimatic variables are included in the following two zipped files: ‘climate_data_aapa_mires.zip’ and ‘climate_data_wet_aapa_mires.zip’. Climate change velocity metrics: The climate velocities for the six bioclimatic variables, developed separately for the two types of aapa mires and the two RCPs, were calculated with climate-analog method (see Brito-Morales et al. 2018). For these calculations, both the present-day and future climate data from the two RCP scenarios were converted from continuous values into categorical climate surfaces following Hamann et al. (2015). During these conversion processes, following categories and within-class ranges were used: GDD5, within-class range 50 °C; January and July temperatures, within-class range 0.5 °C; water balance of May and July, within-class range 2.5 mm; and annual precipitation, within-class range 25 mm. In the conversion process, the climate surfaces in each of the 50-m grid cells were reclassified into one of the 29 GDD5, 27 January temperature, 22 July temperature, 21 May water balance, 22 July water balance, and 19 annual precipitation categories. Using the reclassified climate surfaces, the minimum distances between mire grid cells with similar present-day and future climates for the six variables were determined with the Euclidean distance function in ArcGIS. In the final step of calculating the velocity metrics, the mire-to-mire distances were divided by the number of years between the two points in time (see Brito-Morales et al., 2018; Heikkinen et al., 2020). The derived velocity metrics for the six bioclimatic variables yielded six individual estimates of climate exposure for the two types of study mires, illustrating the magnitude of climate displacement that the local mire species communities are projected to experience (Hamann et al. 2015, Brito-Morales et al., 2018). In our study, for each contiguous aapa mire and wet aapa mire, the mean velocity value for the climate variables were calculated as the average of the 50-m grid cells included in it. The data on the 50-m resolution velocities for the six bioclimatic variables and the two types of aapa mires and the two RCPs are included in the zip file ‘velocity_data_for_mires.zip’. References Aalto, J., Riihimäki, H., Meineri, E., Hylander, K., Luoto, M. (2017) Revealing topoclimatic heterogeneity using meteorological station data. International Journal of Climatology 37, 544-556. AMAP (2017) Snow, Water, Ice and Permafrost in the Arctic (SWIPA) 2017. Arctic Monitoring and Assessment Programme (AMAP), Oslo, Norway. Brito-Morales, I., García Molinos, J., Schoeman, D.S., Burrows, M.T., Poloczanska, E.S., Brown, C.J., Ferrier, S., Harwood, T.D., Klein, C.J., McDonald-Madden, E., Moore, P.J., Pandolfi, J.M., Watson, J.E.M., Wenger, A.S., Richardson, A.J. (2018) Climate Velocity Can Inform Conservation in a Warming World. Trends in Ecology & Evolution 33, 441-457. Gong, J., Wang, K., Kellomäki, S., Zhang, C., Martikainen, P.J., Shurpali, N. (2012) Modeling water table changes in boreal peatlands of Finland under changing climate conditions. Ecological Modelling 244, 65-78. Hamann, A., Roberts, D.R., Barber, Q.E., Carroll, C., Nielsen, S.E. (2015) Velocity of climate change algorithms for guiding conservation and management. Global Change Biology 21, 997-1004. Heikkinen, R.K., Kartano, L., Leikola, N., Aalto, J., Aapala, K., Kuusela, S., Virkkala, R. (2021) High-latitude EU Habitats Directive species at risk due to climate change and land use. Global Ecology and Conservation 28, e01664. Heikkinen, R.K., Leikola, N., Aalto, J., Aapala, K., Kuusela, S., Luoto, M., Virkkala, R. (2020) Fine-grained climate velocities reveal vulnerability of protected areas to climate change. Scientific Reports 10. Kolari, T.H.M., Sallinen, A., Wolff, F., Kumpula, T., Tolonen, K., Tahvanainen, T. (2021) Ongoing Fen–Bog Transition in a Boreal Aapa Mire Inferred from Repeated Field Sampling, Aerial Images, and Landsat Data. Ecosystems. Parviainen, M., Luoto, M. (2007) Climate envelopes of mire complex types in fennoscandia. Geografiska Annaler: Series A, Physical Geography 89, 137-151. Ruuhijärvi, R., (1988) Mire vegetation. Atlas of Finland 141-143. Biogeography, nature conservation. . National Board of Survey and Geographical Society of Finland, Helsinki, pp. 2-4. Rydin, H., Jeglum, J. (2006) The biology of peatlands. Oxford University Press, Oxford. Sallinen, A., Tuominen, S., Kumpula, T., Tahvanainen, T. (2019) Undrained peatland areas disturbed by surrounding drainage: a large scale GIS analysis in Finland with a special focus on aapa mires. Mires and Peat 24, 1-22. Skov, F., Svenning, J.-C. (2004) Potential impact of climatic change on the distribution of forest herbs in Europe. Ecography 27, 366-380. Taylor, K.E., Stouffer, R.J., Meehl, G.A. (2012) An Overview of CMIP5 and the Experiment Design. Bulletin of the American meteorological Society 93, 485-498. Väliranta, M., Salojärvi, N., Vuorsalo, A., Juutinen, S., Korhola, A., Luoto, M., Tuittila, E.-S. (2017) Holocene fen–bog transitions, current status in Finland and future perspectives. The Holocene 27, 752-764.
Protected areas (PAs) are crucial in conserving biodiversity under climate change. In boreal regions, trends of biologically relevant climate variables (i.e., bioclimate) in PAs have remained unquantified. We investigated the changes and variability of 11 key bioclimatic variables across Finland during the period 1961-2020 based on gridded climatology. Our results suggest significant changes in annual mean and growing season temperatures over the entire study area, whereas, e.g., annual precipitation sum and April-September water balance have increased especially in the central and northern parts of Finland. We found substantial variation in bioclimatic changes over the 631 studied PAs; in the northern boreal zone (NB) the number of snow-covered days has decreased on average by 5.9 days between 1961-1990 and 1991-2020, while in the southern boreal zone (SB) the corresponding decrease has been 16.1 days. The number of frost days in spring with absent snow cover has decreased in the NB (on average -0.9 days) while increasing in the SB (0.5 days), reflecting the changing exposure of biota to frost. The observed increases in accumulation of heat in the SB and more frequent rain-on-snow events in the NB can affect drought tolerance and winter survival of species, respectively. Principal component analysis suggested that the main dimensions of bioclimate change in PAs vary across vegetation zones; for example, in the SB the changes are related to annual and growing season temperatures, whereas in the middle boreal zone the changes are linked to altered moisture and snow conditions. Our results highlight the substantial spatial variation in bioclimatic trends and climate vulnerability across the PAs and vegetation zones. These findings provide a basis for the understanding of the multifaceted changes the boreal PA network is facing and help to develop and direct conservation and management.
This dataset contains files that show the climate change velocity metrics calculated for three climate variables across Finland. The climate velocities were used to study the magnitude of projected climatic changes in a nation-wide Natura 2000 protected area (PA) network (Heikkinen et al., 2020). Using fine-resolution climate data that describes the present-day and future topoclimates and their spatio-temporal variation, the study explored the rate of climatic changes in protected areas on an ecologically relevant, but yet poorly explored scale. The velocities for the three climate variables were developed in the following work, where in-depth description of the different steps in velocity metrics calculation and a number of visualisations of their spatial variation across Finland are provided: Risto K. Heikkinen 1, Niko Leikola 1, Juha Aalto 2,3, Kaisu Aapala 1, Saija Kuusela 1, Miska Luoto 2 & Raimo Virkkala 1 2020: Fine-grained climate velocities reveal vulnerability of protected areas to climate change. Scientific Reports 10:1678. https://doi.org/10.1038/s41598-020-58638-8 1 Finnish Environment Institute, Biodiversity Centre, Latokartanonkaari 11, FI-00790 Helsinki, Finland 2 Department of Geosciences and Geography, University of Helsinki, FI-00014, Helsinki, Finland 3 Finnish Meteorological Institute, FI-00101, Helsinki, Finland The dataset includes GIS compatible geotiff files describing the nine spatial climate velocity surfaces calculated across the whole of Finland at 50 m × 50 m spatial resolution. These nine different velocity surfaces consist of velocity metric values measured for each 50-m grid cell separately for the three different climate variables and in relation to the three different future climate scenarios (RCP2.6, RCP4.5 and RCP8.5). The baseline climate data for the study were the monthly temperature and precipitation data averaged for the period from 1981 to 2010 modelled at a resolution of 50-m, based on which estimates for the annual temperature sum above 5 °C (growing degree days, GDD, °C), the mean January temperature (TJan, °C) and the annual climatic water balance (WAB, the difference between annual precipitation and potential evapotranspiration; mm) were calculated. Corresponding future climate surfaces were produced using an ensemble of 23 global climate models for the years 2070–2099 (Taylor et al. 2012) and the three RCPs. The data for the three climate variables for 1981–2010 and under the three RCPs will be made available in separately via METIS - FMI's Research Data repository service (Aalto et al., in prep.). The climate velocity surfaces included in the present data repository were developed using climate-analog approach (Hamann et al. 2015; Batllori et al. 2017; Brito-Morales et al. 2018), whereby velocity metrics for the 50-m grid cells were measured based on the distance between climatically similar cells under the baseline and the future climates, calculated separately for the three climate variables. In Heikkinen et al. (2020), the spatial data for the Natura 2000 protected areas were used to assess their exposure to climate change. The full data on N2K areas can be downloaded from the following link: https://ckan.ymparisto.fi/dataset/%7BED80465E-135B-4391-AA8A-FE2038FB224D%7D. However, note that the N2K areas including multiple physically separate patches were treated as separate polygons in Heikkinen et al. (2020), and a minimum size requirement of 2 hectares were requested. Moreover, the digital elevation model (DEM) data for Finland (which were dissected to Natura 2000 polygons to examine their elevational variation and its relationships to topoclimatic variation) can be downloaded from the following link: https://ckan.ymparisto.fi/en/dataset/dem25_astergdem25. The coordinate system for the climate velocity data files is: ETRS-TM35FIN (EPSG: 3067) (or YKJ Finland/Finnish Uniform Coordinate System (EPSG: 2393)). Summary of the key settings and elements of the study are provided below. A detailed treatment is provided in Heikkinen et al. (2020). Code to the files (four files per each velocity layer: *.tif, *.tfw. *.ovr and *.tif.aux.xml) in the dataset: (a) Velocity of GDD with respect to RCP2.6 future climate (Fig 2a in Heikkinen et al. 2020). Name of the file: GDDRCP26.* (b) Velocity of GDD with respect to RCP4.5 future climate (Fig. 2b in Heikkinen et al. 2020). Name of the file: GDDRCP45.* (c) Velocity of GDD with respect to RCP8.5 future climate (Fig. 2c in Heikkinen et al. 2020). Name of the file: GDDRCP85.* (d) Velocity of mean January temperature with respect to RCP2.6 future climate (Fig. 2d in Heikkinen et al. 2020). Name of the file: TJanRCP26.* (e) Velocity of mean January temperature with respect to RCP4.5 future climate (Fig. 2e in Heikkinen et al. 2020). Name of the file: TJanRCP45.* (f) Velocity of mean January temperature with respect to RCP8.5 future climate (Fig. 2f in Heikkinen et al. 2020). Name of the file: TJanRCP85.* (g) Velocity of climatic water balance with respect to RCP2.6 future climate (Fig. 2g in Heikkinen et al. 2020). Name of the file: WABRCP26.* (h) Velocity of climatic water balance with respect to RCP4.5 future climate (Fig. 2h in Heikkinen et al. 2020). Name of the file: WABRCP45.* (i) Velocity of climatic water balance with respect to RCP8.5 future climate (Fig. 2i in Heikkinen et al. 2020). Name of the file: WABRCP85.* Note that velocity surfaces e and f include disappearing climate conditions. Summary of the study: Climate velocity is a generic metric which provides useful information for climate-wise conservation planning to identify regions and protected areas where climate conditions are changing most rapidly, exposing them to high rates of climate displacement (Batllori et al. 2017), causing potential carry-over impacts to community structure and ecosystem functions (Ackerly et al. 2010). Climate velocity has been typically used to assess the climatic risks for species and their populations, but velocity metrics can also be used to identify protected areas which face overall difficulties in retaining ecological conditions that promote present-day biodiversity. Earlier climate velocity assessments have focussed on the domains of the mesoclimate (resolutions of 1–100 km) or macroclimate (>100 km scales), and fine-grained (<100 m) local climatic conditions created by variation in topography ('topoclimate'; Ackerly et al. 2010; 2020) have largely been overlooked (Heikkinen et al. 2020). This omission may lead to biased exposure assessments especially in rugged terrain (Dobrowski et al. 2013; Franklin et al. 2013), as well as a limited ability to detect sites decoupled from the regional climate (Aalto et al. 2017; Lenoir et al. 2017). This study provided the first assessment of the climatic exposure risks across a national PA (Natura 2000) network based on very fine-grained velocities of three established drivers of high latitude biodiversity. The produce fine-grain climate velocity measures, 50-m resolution monthly temperature and precipitation data averaged for 1981–2010 were first developed, and based on it, the three bioclimatic variables (growing degree days, mean January temperature and annual climatic water balance) were calculated for the whole study domain. In the next phase, similar future climate surfaces were produced based on data from an ensemble of 23 global climate models, extracted from the CMIP5 archives for the years 2070–2099 and the three RCP scenarios (RCP2.6, RCP4.5 and RCP8.5)26. In the final step, climate velocities for each the 50 x 50 m grid cells were measured using climate-analog velocity method (Hamann et al. 2015) and based on the distance between climatically similar cells under the baseline and future climates. The results revealed notable spatial differences in the high velocity areas for the three bioclimatic variables, indicating contrasting exposure risks in protected areas situated in different areas. Moreover, comparisons of the 50-m baseline and future climate surfaces revealed a potential wholesale disappearance of current topoclimatic temperature conditions from almost all the studied PAs by the end of this century. Calculation of climate change velocity metrics for the three climate variables The overall process of calculation of climate velocities included three main steps. (1) In the first step, we developed high-resolution monthly average temperature and precipitation data averaged over the years 1981–2010 and across the study domain at a spatial resolution of 50 × 50 m. This was done by building topoclimatic models based on climate data sourced from 313 meteorological stations (European Climate Assessment and Dataset [ECA&D]) (Klok et al. 2009). Our station network and modelling domain covered the whole of Finland with an additional 100 km buffer. However, it was also extended to cover large parts of northern Sweden and Norway for areas >66.5°N, as well as selected adjacent areas in Russia (for details see Heikkinen et al. 2020). This was done to capture the present-day climate spaces in Finland which are projected to move in the future beyond the country borders but have analogous climate areas in neighbouring areas; this was done to avoid developing a large number of velocity values deemed as infinite or unknown in the data for Finland. The 50-m resolution average air temperature data were developed for the study domain using generalized additive modelling (GAM), as implemented in the R-package mgcv version 1.8–7 (R Development Core Team 2011; Wood 2011). In this modelling we utilised variables of geographical location (latitude and longitude, included as an anisotropic interaction), topography (elevation, potential incoming solar radiation, relative elevation) and water cover (sea and lake proximity), and subsequent leave-one-out cross-validation tests to assess model performance (for full process description, see Aalto et al. 2017; Heikkinen et al. 2020). The resulting topoclimate data effectively captured the physiographic effects of solar radiation and cold-air pooling. To produce gridded precipitation data, we applied global kriging interpolation to the data from 343 rain gauges from the ECA&D dataset. The interpolation was carried out using information on geographical location, topography (elevation and eastness index) and proximity to the sea and R package gstat. The eastness index was obtained from a sine-transforming aspect raster surface calculated from a 50 m × 50 m digital elevation model to capture the effect of prevailing westerly winds on the accumulated precipitation on windward slopes. The gridding was first run at a resolution of 500 × 500 m, whereafter gridded precipitation values were bilinearly interpolated into the same 50 × 50 m resolution as the air temperature data. Next, the three bioclimatic variables ((i) growing degree days (GDD, °C days) indicating the accumulated warmth during the growing season; (ii) mean January air temperature - TJan, °C; (iii) climatic water balance - WAB, mm) were calculated for each 50 x 50 grid cell from the high-resolution gridded 1981–2010 ('baseline') climate data. Earlier research has demonstrated the ecological relevance of these three complementary variables which provide estimations of winter cold, seasonal warmth and moisture availability (Sykes et al. 1996; Luoto et al. 2006; Huntley et al. 2007, 2008). Following Carter et al. (1991), GDD was calculated as the effective temperature sum above the base temperature of 5 °C as follows: GDD5 = ∑ni (Ti - Tb), if Ti -Tb > 5 where Ti denotes the mean temperature at day i, Tb represents the base temperature, and n is the length of the summation period. However, because the daily air temperature data was not available, here the GDD was estimated using monthly data as in Araújo & Luoto (2007). The WAB is the difference between the total annual precipitation sum and the potential evapotranspiration (PET), which was estimated from the monthly air temperatures following Skov and Svenning (2004): PET = 58.93 × Tabove 0°C / 12 (2) In the second step we developed data on future climates by using the climate projections from the ensemble of 23 global climate models (GCMs), derived from the Coupled Model Intercomparison Project phase 5 archives (Taylor et al. 2012). From these archives, we processed to predicted averaged changes in mean temperature and precipitation with respect to the baseline 1981–2010 for the years 2070–2099, and the three RCP scenarios (cf. Moss et al. 2010). As the Coupled Model Intercomparison Project phase 5 climate scenario data represent coarse-scale resolution data, we converted it to match our fine-resolution baseline climate data by interpolation. For this, the climate model data depicting the predicted change in mean temperatures and precipitation with respect to the baseline climate were bilinearly interpolated to the 50 × 50 m grid system, and the change predicted by the GCMs was added to the spatially detailed baseline climate data. After this, the bioclimatic variables were recalculated for each RCP scenario to allow the calculation of climate change velocities across the whole country and the Natura 2000 protected areas. (3) In the third step we developed climate change velocities for the three bioclimatic variables using the climate-analog approach (Hamann et al. 2015) where velocity is calculated by measuring the distance between present-day locations with certain climatic conditions and their future climate analogues, divided by the number of years between the two points in time. Thus, we calculated climate-analog velocities for the 50-m resolution grid climate data by measuring the distance between climatically similar grid cells for the present and future climates under RCP2.6, RCP4.5 and RCP8.5. Prior the actual climate-analog velocity measurements, the climate variable surfaces were converted from continuous values into classified variable surfaces. For this, we defined the boundary values for the variable classes so that the climatically matching grid cells had their within-class ranges as small as possible but, at the same time, avoided artefactual extreme precision. After a set of pilot reclassifications, the following within-class ranges were applied: GDD, within-class range 50 °C with 51 categories; TJan, within-class range 0.5 °C with 60 categories; WAB, within-class range 50 mm with 55 categories. Next, using the reclassified present-day and future climate surfaces the search of the minimum distances between grid cells with similar present-day and future GDD/TJan/WAB climates were executed. The search was carried out using the ArcGIS software (Desktop 10.5.1.) by employing the Euclidean distance function. The minimum distances measured for each 50-m grid cell were divided by the difference between the mean points in the two time slices, 1981–2010 and 2070–2099. The resulting 50-m resolution climate velocity surfaces for the three climate variables are provided in the zipped files included this data repository. In Heikkinen et al. (2020), these climate velocity data were employed in a series of subsequent analyses. For example, high-velocity areas ('velocity hotspots') of the three climate variables were visually compared with each other based on maps showing their 50-m resolution velocities across mainland Finland and the degree of overlap between the present-day range and projected future range of the three climate variables were investigated in each of the 5,068 Natura 2000 polygons included in the study. References Aalto, J., Riihimäki, H., Meineri, E., Hylander, K., Luoto, M., 2017. Revealing topoclimatic heterogeneity using meteorological station data. International Journal of Climatology 37, 544-556. Ackerly, D.D., Loarie, S.R., Cornwell, W.K., Weiss, S.B., Hamilton, H., Branciforte, R., Kraft, N.J.B., 2010. The geography of climate change: implications for conservation biogeography. Diversity and Distributions 16, 476-487. Ackerly, D.D., Kling, M.M., Clark, M.L., Papper, P., Oldfather, M.F., Flint, A.L., Flint, L.E., 2020. Topoclimates, refugia, and biotic responses to climate change. Frontiers in Ecology and the Environment 18, 288-297. Araujo, M.B., Luoto, M., 2007. The importance of biotic interactions for modelling species distributions under climate change. Global Ecology and Biogeography 16. Batllori, E., Parisien, M.-A., Parks, S.A., Moritz, M.A., Miller, C., 2017. Potential relocation of climatic environments suggests high rates of climate displacement within the North American protection network. Global Change Biology 23, 3219-3230. Brito-Morales, I., García Molinos, J., Schoeman, D.S., Burrows, M.T., Poloczanska, E.S., Brown, C.J., Ferrier, S., Harwood, T.D., Klein, C.J., McDonald-Madden, E., Moore, P.J., Pandolfi, J.M., Watson, J.E.M., Wenger, A.S., Richardson, A.J., 2018. Climate Velocity Can Inform Conservation in a Warming World. Trends in Ecology & Evolution 33, 441-457. Carter, T.R., Porter, J.H., Parry, M.L., 1991. Climatic warming and crop potential in Europe: Prospects and uncertainties. Global Environmental Change 1, 291-312. Dobrowski, S.Z., Abatzoglou, J., Swanson, A.K., Greenberg, J.A., Mynsberge, A.R., Holden, Z.A., Schwartz, M.K., 2013. The climate velocity of the contiguous United States during the 20th century. Global Change Biology 19, 241-251. Franklin, J., Davis, F.W., Ikegami, M., Syphard, A.D., Flint, L.E., Flint, A.L., Hannah, L., 2013. Modeling plant species distributions under future climates: how fine scale do climate projections need to be? Global Change Biology 19, 473-483. Hamann, A., Roberts, D.R., Barber, Q.E., Carroll, C., Nielsen, S.E., 2015. Velocity of climate change algorithms for guiding conservation and management. Global Change Biology 21, 997-1004. Heikkinen, R.K., Leikola, N., Aalto, J., Aapala, K., Kuusela, S., Luoto, M., Virkkala, R., 2020. Fine-grained climate velocities reveal vulnerability of protected areas to climate change. Scientific Reports 10:1678. Huntley, B., Green, R.E., Collingham, Y.C., Willis, S.G., 2007. A climatic atlas of European breeding birds. Durham University, The RSPB and Lynx Edicions, Barcelona. Huntley, B., Collingham, Y.C., Willis, S.G., Green, R.E., 2008. Potential Impacts of Climatic Change on European Breeding Birds. Plos One 3. Klok, E.J., Klein Tank, A.M.G., 2009. Updated and extended European dataset of daily climate observations. International Journal of Climatology 29, 1182-1191. Lenoir, J., Hattab, T., Pierre, G., 2017. Climatic microrefugia under anthropogenic climate change: implications for species redistribution. Ecography 40, 253-266. Luoto, M., Heikkinen, R.K., Pöyry, J., Saarinen, K., 2006. Determinants of biogeographical distribution of butterflies in boreal regions. Journal of Biogeography 33, 1764-1778. Moss, R.H., Edmonds, J.A., Hibbard, K.A., Manning, M.R., Rose, S.K., van Vuuren, D.P., Carter, T.R., Emori, S., Kainuma, M., Kram, T., Meehl, G.A., Mitchell, J.F.B., Nakicenovic, N., Riahi, K., Smith, S.J., Stouffer, R.J., Thomson, A.M., Weyant, J.P., Wilbanks, T.J., 2010. The next generation of scenarios for climate change research and assessment. Nature 463, 747-756. R Development Core Team, 2011. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing). Skov, F., Svenning, J.-C., 2004. Potential impact of climatic change on the distribution of forest herbs in Europe. Ecography 27, 366-380. Sykes, M.T., Prentice, I.C., Cramer, W., 1996. A bioclimatic model for the potential distributions of north European tree species under present and future climates. Journal of Biogeography 23, 203-233. Taylor, K.E., Stouffer, R.J., Meehl, G.A., 2012. An Overview of CMIP5 and the Experiment Design. Bulletin of the American meteorological Society 93, 485-498. Wood, S.N., 2011. Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society Series B 73, 3-36.
ABSTRACT Aim We explored the effects of prevalence, latitudinal range and spatial autocorrelation of species distribution patterns on the accuracy of bioclimate envelope models of butterflies. Location Finland, northern Europe. Methods The data of a national butterfly atlas survey (NAFI) carried out in 1991–2003 with a resolution of 10 × 10 km were used in the analyses. Generalized additive models (GAM) were constructed, for each of 98 species, to estimate the probability of occurrence as a function of climate variables. Model performance was measured using the area under the curve (AUC) of a receiver operating characteristic (ROC) plot. Observed differences in modelling accuracy among species were related to the species’ geographical attributes using multivariate GAM. Results Accuracies of the climate–butterfly models varied from low to very high (AUC values 0.59–0.99), with a mean of 0.79. The modelling performance was related negatively to the latitudinal range and prevalence, and positively to the spatial autocorrelation of the species distribution. These three factors accounted for 75.2% of the variation in the modelling accuracy. Species at the margin of their range or with low prevalence were better predicted than widespread species, and species with clumped distributions better than scattered dispersed species. Main conclusions The results from this study indicate that species’ geographical attributes highly influence the behaviour and uncertainty of species–climate models, which should be taken into account in biogeographical modelling studies and assessments of climate change impacts.
Global climate change is a major threat to biodiversity, posing increasing pressures on species to adapt in situ or shift their ranges. A protected area network is one of the main instruments to alleviate the negative impacts of climate change. Importantly, protected area networks might be expected to enhance the resilience of regional populations of species of conservation concern, resulting in slower species loss in landscapes with a significant amount of protected habitat compared to unprotected landscapes. Based on national bird atlases compiled in 1974-1989 and 2006-2010, this study examines the recent range shifts in 90 forest, mire, marshland, and Arctic mountain heath bird species of conservation concern in Finland, as well as the changes in their species richness in protected versus unprotected areas. The trends emerging from the atlas data comparisons were also related to the earlier study dealing with predictions of distributional changes for these species for the time slice of 2051-2080, developed using bioclimatic envelope models (BEMs). Our results suggest that the observed changes in bird distributions are in the same direction as the BEM-based predictions, resulting in a decrease in species richness of mire and Arctic mountain heath species and an increase in marshland species. The patterns of changes in species richness between the two time slices are in general parallel in protected and unprotected areas. However, importantly, protected areas maintained a higher level of species richness than unprotected areas. This finding provides support for the significance and resilience provision of protected area networks in preserving species of conservation concern under climate change.
This repository contains files that show optimal sites in Central Finland for the northern goshawk (Accipiter gentilis, hereafter goshawk), an indicator species of boreal forests with conservation values. The optimal sites were derived from the habitat suitability model outputs included in the following publication: Björklund Heidia, Parkkinen Anssib, Hakkari Tomic, Heikkinen Risto K.d, Virkkala Raimod, Lensu Anssib (2020): Predicting valuable forest habitats using an indicator species for biodiversity. Biological Conservation, https://doi.org/10.1016/j.biocon.2020.108682 . a Finnish Museum of Natural History Luomus, P.O. Box 17, FI-00014 University of Helsinki, Finland b University of Jyvaskyla, Department of Biological and Environmental Science, P.O. Box 35, FI-40014 University of Jyvaskyla, Finland c Centre for Economic Development, Transport and the Environment Central Finland, P.O. Box 250, FI-40101 Jyväskylä, Finland d Finnish Environment Institute, Biodiversity Centre, Latokartanonkaari 11, FI-00790 Helsinki, Finland The files are ArcGIS compatible shape files which indicate the spatial location of the 160 m × 160 m grid cells which include forest stands projected to be either highly suitable or suitable as a nesting site for the goshawk in Central Finland. The habitat suitability models and values were developed across the study area using Maxent software. The files show those 160-m grid cells from the study area which were included in one of the following two categories: (i) cells deemed as the most optimal (with high probability of suitable conditions) for goshawk nesting with suitability index values in Maxent outputs varying between 0.92–1.00 (‘best’ goshawk squares), and (ii) cells deemed as ‘good’ goshawk squares (with Maxent suitability index values of ≥ 0.69 and < 0.92). The coordinate system for the data files is: ETRS-TM35FIN (EPSG: 3067) (or YKJ Finland/Finnish Uniform Coordinate System (EPSG: 2393)). Summarization of the key settings and elements of the study are provided below. A detailed treatment can now be found in the article published in Biological Conservation (Björklund et al.) for which the link is the following: https://doi.org/10.1016/j.biocon.2020.108682 . Summary of the study Intensive commercial use of boreal forests is an accelerating threat to forest biodiversity, highlighting the development of cost-effective tools to detect the locations valuable for conservation. We applied species distribution models (SDMs) in our study area, Central Finland, to locate the optimal nesting sites for the goshawk, an indicator bird species for biodiversity hotspots in mature boreal forests. The optimal sites (here, 160 x 160 m grid squares) for the goshawk were determined using the Maxent software. Optimal squares for the goshawk had forests with considerably high volumes of Norway spruce (Picea abies, hereafter spruce) covering only 3.4% of the boreal landscape, and they were located mostly outside protected areas. Many of the squares with optimal nesting forests appeared to be under threat due to recently intensified logging operations. Half of the squares were logged to some extent and 10% were already lost or notably deteriorated due to logging after 2015 for which our models were calibrated. Threats to biodiversity of mature boreal spruce forests are likely to accelerate with increasing logging pressures. Thus, there is an urgent need to secure the continuous supply of mature spruce forests in the landscape by developing a denser network of protected areas and applying measures that aid in sparing large entities of mature forest on privately-owned land. Our modelled optimal squares can be used for selection of potential areas with biodiversity values in conservation prioritization. The study species The goshawk is a raptor species which prefers mature forests for nesting in Europe. Old forests dominated by spruce are considered as important for the breeding success of the species particularly in northern latitudes. Thus, intensive forest management can impair the breeding possibilities of the goshawk, and changes in forest landscapes are likely to contribute to the decline of the species. For example, in Finland, the goshawk is classified as nearly threatened species. In our study, we used the goshawk as an indicator species to model the spatial locations of boreal forest with much potential for including biodiversity values. The indicator species status of the goshawk is based on earlier studies showing the close association of the goshawk with various taxa of mature spruce forest, as well as the reported declines of both the goshawk and associated species due to loggings. Developing Maxent models for the goshawk The location data on occupied nests of the goshawk gathered in spring and summer 2015 and 2016 in Central Finland – as a part of the Finnish Common Birds of Prey Monitoring – were related to a set of environmental predictor variables using a maximum entropy method, Maxent software, which is considered particularly useful for modelling presence-only data (such as our goshawk nest site data). In our case, the data on forest stand and tree characteristics were related using Maxent to the known nesting sites to predict suitable conditions for the species across the Central Finland. The forest data used in the modelling were extracted from the multi-source national forest inventory (MS-NFI) data sources governed by the Natural Resources Institute Finland. The MS-NFI data used in our modelling are based on field data of the 11th and 12th NFIs from 2009 to 2016 and satellite images from 2015 and 2016. Prior modelling, Pearson correlations were calculated between the continuous environmental variables at the nest sites. Of the highly (|r| ≥ 0.7) correlated variables, we chose those variables which are known to be important for the goshawk, which are useful for generalization in other areas, or whose impact was of specific interest. Our final selected set of predictor variables included one class variable, site fertility class, and nine continuous variables: growing stock volume of the spruce, pine, birches and other hardwood, canopy cover, canopy cover of broad-leaved trees, saw timber of other broad-leaved trees than birches, pulpwood volume of the birches, and the biomass of the stem residual of the spruce. The original MS-NFI data recorded at the resolution of 16 × 16 m were resampled to the resolution of 160 × 160 m for the Maxent models, to represent one potential nesting forest stand. The accuracy of Maxent models were assessed with cross-validation and associated averaged AUC-values. The relative importance of the variables was measured by variable contribution and model deterioration measures provided by Maxent. The cloglog-transformed output index values ranging from 0 to 1 described the relative suitability of the 160-m squares to goshawk nesting. Based on the index values, the squares were classified as ‘optimal’ (with index values of 0.69–1.00), ‘typical’ (0.46– <0.69) and ‘poor’ (<0.46). In addition, we divided optimal squares into ‘best’ goshawk squares (index values of 0.92–1.00 corresponding to a high probability of suitable conditions), and ‘good’ goshawk squares (index values ≥ 0.69 and < 0.92). Maxent model outputs Spruce volume was the most important variable in defining habitat suitability for goshawk nesting, but hardwood cover, other hardwood logs and site fertility class contributed also to some extent to habitat suitability. In Maxent outputs, the set of 160-m squares deemed as optimal for goshawk nesting included 6 895 (cover 0.9% of the study area) best goshawk squares and 19 421 (cover 2.5%) good goshawk squares. The projected best and good goshawk squares were mostly located in unprotected areas: 95.0% of the best and 96.0% of the good goshawk squares occurred completely outside protected areas. For further details concerning the data and the model outputs, see the referred article Björklund et al. (2020). State of the optimal goshawk squares In total, 11% of best and over 9% of good goshawk squares were severely altered due to recent harvesting, typically clear-cutting, of the forests during the time period between 2015 and 2019. Altogether, some level of logging occurred in 3 062 (44%) of best goshawk and 9 846 (51%) of good goshawk squares during the recent years. However, many of the squares still included enough unlogged area for the goshawk in 2019. In our article, we conclude that while most of the optimal squares for the goshawk were still preserved in 2019, they are under risk as they are mainly situated outside protected area network. This stresses the importance of conserving biodiversity with complementary measures in privately-owned managed forests. In conclusion, a denser network with more PAs for forest-dwelling species should be secured in areas with intensive forestry, e.g. in southern Finland where PAs currently cover a smaller proportion of land compared to northern Finland.
Abstract Aim To analyse the effect of the inclusion of soil and land‐cover data on the performance of bioclimatic envelope models for the regional‐scale prediction of butterfly (Rhopalocera) and grasshopper (Orthoptera) distributions. Location Temperate Europe (Belgium). Methods Distributional data were extracted from butterfly and grasshopper atlases at a resolution of 5 km for the period 1991–2006 in Belgium. For each group separately, the well‐surveyed squares ( n = 366 for butterflies and n = 322 for grasshoppers) were identified using an environmental stratification design and were randomly divided into calibration (70%) and evaluation (30%) datasets. Generalized additive models were applied to the calibration dataset to estimate occurrence probabilities for 63 butterfly and 33 grasshopper species, as a function of: (1) climate, (2) climate and land‐cover, (3) climate and soil, and (4) climate, land‐cover and soil variables. Models were evaluated as: (1) the amount of explained deviance in the calibration dataset, (2) Akaike’s information criterion, and (3) the number of omission and commission errors in the evaluation dataset. Results Information on broad land‐cover classes or predominant soil types led to similar improvements in the performance relative to the climate‐only models for both taxonomic groups. In addition, the joint inclusion of land‐cover and soil variables in the models provided predictions that fitted more closely to the species distributions than the predictions obtained from bioclimatic models incorporating only land‐cover or only soil variables. The combined models exhibited higher discrimination ability between the presence and absence of species in the evaluation dataset. Main conclusions These results draw attention to the importance of soil data for species distribution models at regional scales of analysis. The combined inclusion of land‐cover and soil data in the models makes it possible to identify areas with suitable climatic conditions but unsuitable combinations of vegetation and soil types. While contingent on the species, the results indicate the need to consider soil information in regional‐scale species–climate impact models, particularly when predicting future range shifts of species under climate change.
The EU aims at reaching carbon neutrality by 2050 and Finland by 2035. We integrated results of three spatially distributed model systems (FRES, PREBAS, Zonation) to evaluate the potential to reach this goal at both national and regional scale in Finland, by simultaneously considering protection targets of the EU biodiversity (BD) strategy. Modelling of both anthropogenic emissions and forestry measures were carried out, and forested areas important for BD protection were identified based on spatial prioritization. We used scenarios until 2050 based on mitigation measures of the national climate and energy strategy, forestry policies and predicted climate change, and evaluated how implementation of these scenarios would affect greenhouse gas fluxes, carbon storages, and the possibility to reach the carbon neutrality target. Potential new forested areas for BD protection according to the EU 10% protection target provided a significant carbon storage (426-452 TgC) and sequestration potential (- 12 to - 17.5 TgCO2eq a-1) by 2050, indicating complementarity of emission mitigation and conservation measures. The results of the study can be utilized for integrating climate and BD policies, accounting of ecosystem services for climate regulation, and delimitation of areas for conservation.
