The effects of mapping resolutions on
statistics that quantify heathland
vegetation succession
Student: Dimitri Tjon Sie Fat, BSc Examiner: dhr. dr. Albert Tietema
Summary
This MSc project aims to 1) determine the vegetation succession of heathland vegetation species and 2) test how data resolution influences rate of change statistics. High mapping resolutions minimize
calculation errors that can influence statistical outcomes on small plots. I will quantify these calculation errors using a reliable dataset and standardized statistical tests, such as one-way ANOVA and dependent t-tests. The dataset has been compiled between 2015 and 2018 on an experimental site near Oldebroek, the Netherlands. The dataset includes 72 maps created over the span of 4 years from 9 plots, each containing two 1 by 2-meter subplots mapped for vegetation coverage. The sites have different manipulation histories – increased temperature and drought – that together with control plots give insight into how predicted climate-change could influence these fragile ecosystems around the northern hemisphere.
Theoretical Framework
Climate change and its effects on heathlands
The Intergovernmental Panel on Climate Change (IPCC) (2014) has concluded that the average global temperature has increased by approximately 1°C in the last century. The majority of the model scenarios used by the IPCC indicate that the mean temperature on earth will increase with at least 2.0°C at the end of the 21st century compared to the end of the 20th century. The increase in temperature will vary regionally. In northern regions a higher rise in temperature, and an increase in the intensity, occurrence, and expanse of droughts and heatwaves (IPCC, 2014) is expected. The higher temperatures in the northern regions, including the Netherlands, are caused by the retreat of Arctic ice, which acts as a cold cover for the underlying Arctic sea (EEA, 2016). With the reducing Arctic ice cover, the temperature of the North and Baltic seas rose five to six times faster than the global average over the past 25 years, warming the air faster than the increase in global average temperature (EEA website, as seen on August 9th, 2019). An increase of mean global temperature leads warmer oceans, resulting in an increased evaporation rate. This results in a higher average worldwide precipitation, with fluctuations on regional scales (Van Boxel, 2008). The Netherlands borders the southern part of the North Sea, and is directly affected by its changes, being the southern border of the northern regions that have experienced a relatively high rise in temperature compared to the global mean temperature. In the following
paragraphs the combined effects of higher temperature and longer, drier periods, are proposed to have effects on the functioning, colonization, and regeneration of heathlands in northern regions, in
particular the Netherlands.
Water and temperature are main components in biological and chemical processes, which makes them key factors in how ecosystems function (Beier et al., 2004). The effects of climate changes on ecosystem functions are uncertain and complex (Beier et al., 2004). However, there is enough evidence that changes in temperature and precipitation will have an influence on the dynamics of ecosystems (Wessel et al., 2004).
European heathlands are important ecosystems that harbor a wide range of species (Newton et al., 2009), and offer important ecosystem services. Heathlands are sensitive to environmental changes such as changing rainfall rates and temperature (Beier et al., 2004). They are a common ecosystem in Europe, but only cover 4% of the European land mass (Maes et al., 2015). Most heathlands in the Netherlands
were formed by anthropogenic activities and still require active management (Wessel et al., 2004). Because heathlands are a nutrient poor shrubland type, they contain unique species that require specific management to be maintained and to keep their place within the high biodiversity of the ecosystem. Predicted climate-change will have an impact on heathlands in terms of vegetation nutrient cycles, carbon fluxes, and composition and phenology (Prieto et al., 2009; Wessel et al., 2004). In general, shrublands are important for the sequestration of carbon in their soils and biomass. Around 30% of the terrestrial carbon is sequestrated into heathlands in this manner (Adhikari & White, 2016). Carbon sequestration is an important tool in combating climate-change, since CO2 is taken out of the
atmosphere to be used in ecosystems. Efficient management can increase the carbon sequestration and maximize the heathlands’ potential to be efficient carbon sinks (Lal, 2004). On the other hand, the loss of shrubland, including heathlands, could release more CO2 into the atmosphere, and exacerbate the climate change problem (Lal, 2004).
Vos et al. (2007) has identified the following ways in which climate change directly influences vegetation and ecosystems: 1) affecting physiological processes (e.g. photosynthesis, respiration, and
decomposition); 2) altering phenological processes (e.g. lifecycle processes); 3) change geographical spreading of species (e.g. the area expansion of species preferring higher temperatures); 4) local
adaptations (e.g. evolution). These four effects cause changes in how species interact. Species can profit from these changes and expand their habitat. For example, a more drought resistant species would outcompete species that require relatively more water in long periods of drought. It is also observed that species can be displaced, or become extinct by such negative effects (Vos, Nijhof, Veen, Opdam, & Verboom, 2007)
The influence of climate change on plant competition
Climate changes can influence plant production, causing changes in the competitive interactions for light, water, and nutrients between plants (Damgaard, Riis-Nielsen, & Schmidt, 2009). Species that are more negatively influenced by climate change are at a disadvantage when a higher growth rate leads to more available light, water, and nutrients (Damgaard et al., 2009).
Heathlands have relatively acidic and nutrient poor soils, with nutrient leaching being wide-spread (Vos et al., 2007). Nitrogen is the main factor that limits plant growth, with C. vulgaris best adapted to this low nitrogen environment by growing at a low rate and producing woody stems. In the case of Oldebroek, C. vulgaris is the dominant species in the area. It is a slow growing plant that favors low concentrations of nitrogen and it is relatively resistant to drought stress in comparison to most heathland vegetation (Wessel et al., 2004). A higher CO2 concentration and temperature can cause an increase in photosynthesis and extend the growing season (Beier et al., 2004), with a shift in
phenological patterns (Shaw et al., 2002). Higher temperatures will increase the availability of nutrients and mineralization over time (Kongstad et al., 2012). The changes in nutrient availability, temperature, CO2 concentration, and atmospheric deposition, will favor faster growing species. This can stimulate the growth of fast growing Poaceae species and biomass production, especially if nutrients are not limited, leading to changes in the ecosystem structure. Species that grow faster can quickly grow tall enough to access more sunlight, or have more and deeper roots to access more water and nutrients, while limiting the access to light, water, and nutrients of surrounding species. This interaction is important for the production by ecosystems (Brooker, 2006). Such a regime shift will change from an evergreen shrub ecosystem to a grassland ecosystem that stores less carbon and nutrients (Kongstad et al., 2012).
Vegetation under optimal growing conditions will benefit, but only if they are able to colonize areas with suitable habitats. Especially species such as grasses and sedges, that can disperse easily or have a wide suitable habitat range will be able to colonize more readily (Vos et al., 2007).
I will define colonization as the first plant species occupying an area that it did not occupy before. The Oxford English dictionary describes colonization in an ecological context as “the action by a plant or animal of establishing itself in the area”. This means that species can only colonize an area that they have not previously occupied. I will define regeneration as the regrowth of a species in an area that it has occupied before. The Oxford English dictionary also describes the verb ‘regenerate’ as “bring into renewed existence; generate again”. This means that species can only regenerate in an area that they are currently occupying. Colonization can occur in an area that has been impacted by an outside influence: for instance, mowing or removal of the (top-)soil. By temporary reducing the influence of species in the area, other species gain an opportunity to colonize the area, as a direct result of reduced competition. This can create a different succession of vegetation and can even lead to habitat shifts (Vos et al., 2007). Species are generally not able to colonize all the new suitable area after a shift of habitat, which follows a step-wise function (figure 1). This behavior is linked to distance and barriers between old and new habitats, and the ease in which species can expand their habitat (Vos et al., 2007). These combined factors can result in an overall lowered species diversity, with a shift from unique species towards more general species (Vos et al., 2007).
Figure 1:Colonization of vegetation after habitat shifts (Vos et al., 2006)
In figure 1, the original situation shows the green colored species as being dominant in the area. A habitat shift follows, in which a large part of the green colored species dies off, with red colored species colonizing the shifted habitat. In the third illustration a blue colored species makes its entrance into the area and eventually drives out the red colored species. This gives the green colored species an
opportunity to slightly expand its influence once more. However, the competition between species has caused the green species to grow in a more dispersed manner than was the case in the baseline
situation. This modeled competition between plant species closely resembles the previously mentioned competition between C. vulgaris and Poaceae species.
The influence of water and temperature on the soil of heathlands
The nutrient availability in heathlands increases by vegetation succession. When dead vegetation material accumulates, microorganisms transform the litter into nutrients by decomposition. This succession is a dynamic factor of heathland ecosystems (Vos et al., 2007).
Drought is often accompanied by a lowering of the ground water table. This is the result of minimal moisture or reduced seepage and precipitation rates (Vos et al., 2007). Droughts directly lead to water being minimally available, and will have a major impact on the growing processes of plants (Larsen et al., 2011). Drought can cause changes in the water holding capacity of soil (Sowerby, Emmett, Tietema, & Beier, 2008). This decrease in soil moisture leads to more aerobe processes, resulting in a higher soil temperature. Mineralization rates of organic matter increase, resulting in eutrophication. Furthermore, dry soil can readily take up rain water. This can change the soil pH, depending on buffer capacity (Kopittke, Tietema, & Verstraten, 2012).
Precipitation changes lead to changing plant composition, in turn decreasing the biodiversity, which lowers ecosystem productivity (Kongstad et al., 2012). Droughts negatively influence water-dependent species. Seeds develop and germinate less with smaller amounts of water. Plants can still be affected negatively by a short drought (Vos et al., 2007).
Both temperature increase and drought cause a decrease in biodiversity, with a risk of a reduction in species diversity. Without responsible management, these important ecosystems will decline in both their functioning and the quality of their ecosystem services. In order to have effective management of heathlands, more research is required. Knowing how climate change affects local heathlands is crucial in providing a better understanding (Kongstad et al., 2012) and more accurate prediction on the complex climate change impacts on ecosystems (Shaw et al., 2002). Experiments and data of the effects of management on heathlands are key in increasing our understanding of these interactions (Kongstad et al., 2012). When the impacts of climate change are known, research on recovery management is made possible.
Managing heathlands in the Netherlands
The Dutch heathlands could grow by deforestation and reclaiming moorland. Agricultural exploitation maintained the heathlands, which were transformed into agricultural fields, pastures, and forests in the 1800’s (Diemont, University, & Berendse, 1996). Heathland vegetation has declined since then. At the start of the twentieth century, large heathland areas contained only C.vulgaris with a height of 10-15cm, which is ideal for grazing cattle. In the nineteenth century the heathland cover was 800,000 ha. This shrunk to 100,000 ha in 1940 (Diemont et al., 1996). The introduction of fertilizers, coal, and barbed wire in the twentieth century caused heathlands to be utilized less for economic reasons. This led to a period in which the vegetation could recover, even if the total heathland cover was further reduced to 42,000 ha in 1996 (Diemont et al., 1996).
Presently, the dominant heathland vegetation consists of Calluna vulgaris, Erica tetralix, Deschampsia
flexuosa and Molinia caerulea (De Jong, 2018). Active management is needed to preserve heathland
ecosystems. The restorative abilities of an ecosystem will determine its adaptability to climate change (György Kröel-Dulay et al., 2015). This can be influenced by efficient management. According to many fauna-experts this management has much room for improvement, leading to a decline of heathland species (Vos et al., 2007). The vegetation has to be exposed to several forms of disturbance in order to
maintain heathlands. Management practices consist of mowing, turfing, sheep grazing and burning (Webb, 1998); (Vos et al., 2007). The plots in the heathland of Oldebroek have been mowed for research.
Vegetation regenerates after a disturbance such as mowing. The first species to arise after clear cutting the vegetation are called pioneer species, and the species that grow in the following stage are
intermediate species. The ability of species to exploit the conditions of the community will determine the succession of species, which results in competition between them. Competition is defined by Tow & Lazenby (2000) “as an interaction between individuals brought about by a shared requirement for a resource in limited supply and leading to reduction in the survivorship, growth and/or reproduction of at least some of the competing individuals concerned.” (De Jong, 2018).
Mowing has a short-term effect. Mowing does not only consist of the cut-off, but also the removal of the cuttings. It results in lower, denser vegetation. Roots and underground life stay intact, which makes it easier for plants to recover. Mowing C. vulgaris older than 15 years will reduce its coverage because it regenerates very slowly (Vos et al., 2007).
The availability of research on management of heathland vegetation is low, with most articles only relying on expertise and professional judgement (Vos et al., 2007). This makes general statements especially difficult to formulate, with each location having its own characteristics.
Image resolution and subsampling
Maps give a representation of an area and emphasize relationships between elements in this space. This makes mapping a useful tool to give a visual overview of the area. These maps can be made annually under different circumstances, such as higher temperatures and droughts. Collectively, the maps then give an overview of the growth or decline of the vegetation, under different growing conditions, and throughout time. Vegetation data, such as area and density of growth, can be collected and visualized in digital maps. This data makes the succession of vegetation throughout time quantifiable, and can be used to statistically test if higher temperatures and droughts impact the growth area of different species. Since the data and maps are digitized, computers are used to rapidly process the data and perform the statistical tests, saving valuable time compared to processing the data manually. Computers do visualize maps and images on a screen or monitor instead of on paper, and the method in which this happens can influence both how the image is perceived, and the data that is linked to the maps. A digital image or map consists of a specified number of pixels. The same principle applies to a raster map, which is made up of cells, all with their individual values, represented by colors. When rows or columns of cells are discarded, the same image will be formed by a smaller number, but larger cells. This results in edges deforming because of the lower resolution, and becoming less smooth than in the higher resolution image (Herrera-Navarro, Jiménez-Hernández, & Terol-Villalobos, 2012) (Figure 2).
Figure 2: Circles at different resolutions (Herrera-Navarro et al., 2012)
By lowering the resolution (i.e. reducing the number of pixels to visualize the same object within a map), unexpected deformities within the map can have a significant effect on the feature properties (Herrera-Navarro et al., 2012). Subsequently, these altered properties affect the mapped features
(Palamakumbure, Flentje, & Stirling, 2015), for example the calculated growth area of species, dispersal patterns or unwanted omission of entire units in the map: whenever the pixel size exceeds the smallest feature size. To overcome this, vegetation maps are traditionally created with the highest resolution as possible. However, high-resolution mapping requires a lot of effort in the field and leads to large file sizes: slowing down the post-processing stage (Palamakumbure et al., 2015). This might cause problems when projects are upscaled and will have to deal with more data. Currently, the optimal mapping resolution for high resolution vegetation applications is not well known.
Spatial and temporal statistics
An experimental site near Oldebroek in the Netherlands is used to collect vegetation data of the local heathlands under normal, raised temperature, and drought conditions. The dataset has been compiled between 2015 and 2018, and includes 72 maps created over the span of 4 years from 9 plots, each containing two 1 by 2-meter subplots mapped for vegetation coverage. The different growth coverages in the drought, temperature, and control plots will be compared and tested for statistical differences. This will determine the influence of climate change on the vegetation succession in the Dutch heathland. Furthermore, high mapping resolutions minimize calculation errors that can influence statistical
outcomes on small plots. I will quantify these calculation errors and advise on optimal resolutions for similar projects on different scales. This will result in more insight into how climate change affects heathlands, and improved methods for collecting future data on similar projects.
Research Aim
The ultimate aims are to create optimal workflows for research on the effects of climate change on vegetation coverage and high-resolution mapping of vegetation coverage. This will be achieved by standardizing heathland vegetation data that has been collected between 2015 and 2018 at an
experimental site near Oldebroek, the Netherlands. Optimal mapping resolutions for different scales are assessed by comparing the statistical results for the vegetation succession at different resolutions as well. By standardizing methods, similar research in different areas can be compared against each other and even combined for research in larger areas. This will lead to future research having uniform data that can be analyzed efficiently in a standardized manner.
Research Questions
1. How does sub-sampling data with different resolutions influence the statistics that quantify heathland vegetation succession?
2. What is the role of climate change in the regeneration rates of heathland vegetation species?
Methods
The Dutch experimental site Oldebroek (52°24’N 5°55’E) is located in the province of Gelderland at the artillery range belonging to the Royal Netherlands Army. The average temperature is 10.1 °C, and the average annual precipitation is 1072 mm. The heathland area is dominated by Calluna vulgaris,
Deschampsia flexuosa and Molinia caerulea with scattered Betula pendula, Pinus Sylvestris, and Juniperus communis (Schaap, 2015). The research plots are situated within the 30 year old part of the
heathland (Kopittke et al., 2012).
Experimental plots
The experimental site consists of nine plots. For six plots, the precipitation and temperature conditions are altered (figure 3). Three plots are kept warmer at night and three plots have the precipitation blocked during the summer. The remaining three plots are control plots (figure 4).
Figure 3: The Oldebroek site layout shows the nine experimental plots are located in the old community of the heathland. (source: Kopittke et al., 2012)
Figure 4. Climate manipulation techniques. (Source: http://increase.ku.dk/experimental_approach/)
The temperature plots are kept warmer by being covered by reflective curtains at night. This reflects and keeps in infrared radiation (IR), and minimizes heat loss (Beier et al., 2004). This results in a higher temperature in the soil of 0.5-1.0 °C during the day and 1.0 °C at night. Light sensors enable the curtains to operate automatically, which ensures that the three temperature plots are covered simultaneously.
The drought plots are covered by a transparent curtain for 2-3 months during the summer, to reduce the precipitation and simulate droughts (Beier et al., 2004). Rain sensors trigger the curtains to cover the plot during rainfall. The curtains are opened during dry periods so that the plots are only covered when it rains (Llorens et al., 2004). This coverage reduces precipitation by about 20 percent. In order to avoid drought simulations in the temperature plot, a sensor keeps the curtains open during nighttime rainfall (Llorens et al., 2004).
The three control plots have similar scaffolding as the other plots for accurate comparisons.
Management
Heathlands are mowed regularly to avoid grass overgrowth. In 2009, an area of 2m2 was cut in all of the plots. In 2013, another 2m2 area of the plots was (Figure 4). The mowed areas are studied for
interactions between climate change and recolonization.
The data has been collected by bachelor students (Schaap, Knibbeler, Blok, de Jong) from 2015 to 2018. They have digitized maps of the growth coverage of different species in the plots, and have calculated the percentages of the growth areas of the different species under the different conditions. They have also calculated whether drought and temperature have a statistically significant influence on the succession of the vegetation in the plots.
Figure 4: A plot, cut in 2009 and 2013.
Subsampling maps to adjust mapping resolutions
The resample tool in Arcmap 10.6.1 will be used to adjust the mapping resolutions of the maps. The nearest neighbor technique will be used to interpolate the data to provide an as accurate output raster file as possible. It is used primarily for discrete data such as land-use classification. It does not change the value of the cells. The maximum error for this method is one-half the cell size (ERDAS website, as seen August 1st 2019).
Spatial and temporal statistical methods
I will use SPSS to execute the one-way analysis of variance (ANOVA) test and the paired (samples) t-test to verify if the groups have differences that are statistically significant.
A one-way analysis of variance (ANOVA) is used when there is a categorical independent variable (with two or more categories) and a normally distributed interval dependent variable. The test is used to calculate for differences in the means of the dependent variable broken down by the levels of the independent variable. The ANOVA test will be used to compare different plots with each other, and will be useful for testing the effects of temperature or drought with the control group (UCLA Statistical Consulting Group, as seen August 1st 2019).
A paired (samples) t-test is used when there are two related observations (i.e., two observations per subject). It is used to calculate if the means on two normally distributed interval variables differ from one another. The paired (samples) t-test will be used to compare the differences of one plot between two years. It will also be used to compare the effects of different mapping resolutions on the same plot with each other (UCLA Statistical Consulting Group, as seen August 1st 2019)..
Workflow
1. Prepare the data of previous years for statistical analysis
2. Use the highest resolution to compare the revegetation with the previous years with statistical tests
3. Repeat step 2 for lower resolutions
4. Statistically verify if different resolutions have an impact on the results 5. Analyze and comment on the results
6. Propose an optimalized, uniform workflow for the research project
Time Schedule
The time schedule shows a significant break of 4 months after the training period. This is to account for my internship period, which is less flexible than writing the thesis itself. The four-month internship is very difficult to combine with a Master thesis, and makes it possible to finish my MSc. study in March.
Budget & Funding
Item
Price per item (€)
Amount
Price (€)
Training Scotland Tickets (return) 260 1 260 Train 25 2 50 Car 50 1 50 lodging 0 1 0 Consumption 100 1 100 Visa 285 1 285 Field work Oldebroek Transport 0 1 0 Equipment 0 1 0 Software Licenses 0 1 0 Labor
Setup, analyzing, and writing costs
15 624
(24 ects in hours)
9360
Total
10.105
All of the expenses will be funded by the student. Student loans are not possible as of now for this research, but possible grant applications will be considered to alleviate the personal cost.
Insurance & Safety
I have a worldwide health insurance (AON) that will cover emergency health expenses during my training period in Scotland. According to the website of the Dutch Ministry of Foreign Affairs, there are no safety risks in Scotland or the rest of the UK. This green color code gives me leeway to freely receive training in Scotland.
Equipment
The writing and data processing will be done on my personal laptop. I have installed a licensed version of ARCGIS 10.6.1 on my laptop, which I received from the GIS studio of the University of Amsterdam. Microsoft Office is also installed with a student license.
The necessary field work equipment to identify species in the field will be purchased by the student. The car that will be used during field work will be arranged by the supervisor for the duration of the field work.
References
Adhikari, A., & White, J. D. (2016). Climate change impacts on regenerating shrubland productivity.
Ecological Modelling; Ecological Modelling, 337, 211-220. doi:10.1016/j.ecolmodel.2016.07.003
Beier, C., Emmett, B., Gundersen, P., Tietema, A., Peñuelas, J., Estiarte, M., . . . Williams, D. (2004). Novel approaches to study climate change effects on terrestrial ecosystems in the field: Drought and passive nighttime warming. Ecosystems, 7(6), 583-597. doi:10.1007/s10021-004-0178-8
Brooker, R. W. (2006). Plant–plant interactions and environmental change. New Phytologist, 171(2), 271-284. doi:10.1111/j.1469-8137.2006.01752.x
Damgaard, C., Riis-Nielsen, T., & Schmidt, I. (2009). Estimating plant competition coefficients and predicting community dynamics from non-destructive pin-point data: A case study with
calluna vulgaris and deschampsia flexuosa. Plant Ecology; an International Journal, 201(2), 687-697. doi:10.1007/s11258-008-9521-z
De Jong, T. C., 2018. The effects of climate change on the regeneration of disturbed heathland
in Oldebroek, The Netherlands’Diemont, W. H., University, A., & Berendse, F. (1996). Survival of dutch heathlands.
EEA, 2016: Climate change, impacts and vulnerability in Europe 2016. An indicator-based report. EEA, Copenhagen, Denmark, 424 pp.
György Kröel-Dulay, Ransijn, J., Inger, K. S., Beier, C., Angelis, P. D., Dato, G. D., . . . Penuelas, J. (2015). Increased sensitivity to climate change in disturbed ecosystems. Nature Communications, 6(1) doi:10.1038/ncomms7682
Herrera-Navarro, A., Jiménez-Hernández, H., & Terol-Villalobos, I. (2012). A probabilistic measure of
circularity doi:10.1007/978-3-642-34732-0_6
IPCC, 2014: Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the
Fifth Assessment Report of the Intergovernmental Panel on Climate Change [Core Writing Team,
R.K. Pachauri and L.A. Meyer (eds.)]. IPCC, Geneva, Switzerland, 151 pp.
Kongstad, J., Schmidt, I., Riis-Nielsen, T., Arndal, M., Mikkelsen, T., & Beier, C. (2012). High resilience in heathland plants to changes in temperature, drought, and CO 2 in combination: Results from the CLIMAITE experiment. Ecosystems, 15(2), 269-283. doi:10.1007/s10021-011-9508-9
Kopittke, G. R., Tietema, A., & Verstraten, J. M. (2012). Soil acidification occurs under ambient conditions but is retarded by repeated drought: Results of a field-scale climate manipulation experiment.
Science of the Total Environment, 439, 332-342. doi:10.1016/j.scitotenv.2012.09.044
Lal, R. (2004). Carbon sequestration in dryland ecosystems. Environmental Management, 33(4), 528-544. doi:10.1007/s00267-003-9110-9
Larsen, K. S., Beier, C., Albert, K. R., Ambus, P., Carter, M. S., Ibrom, A., . . . Maraldo, K. (2011). Reduced N cycling in response to elevated CO 2 , warming, and drought in a danish heathland: Synthesizing results of the CLIMAITE project after two years of treatments. Global Change Biology, 17(5), 1884-1899. doi:10.1111/j.1365-2486.2010.02351.x
Llorens, L., Peñuelas, J., Beier, C., Emmett, B., Estiarte, M., & Tietema, A. (2004). Effects of an experimental increase of temperature and drought on the photosynthetic performance of two ericaceous shrub species along a North–South european gradient. Ecosystems, 7(6), 613-624. doi:10.1007/s10021-004-0180-1
Maes, J., Fabrega, N., Zulian, G., Barbosa, A. L., Vizcaino, P., Ivits, E., . . . Lavalle, C. (2015). Mapping and
assessment of ecosystems and their services: Trends in ecosystems and ecosystem services in the european union between 2000 and 2010 doi:10.2788/341839
Newton, A. C., Stewart, G. B., Myers, G., Diaz, A., Lake, S., Bullock, J. M., & Pullin, A. S. (2009). Impacts of grazing on lowland heathland in north-west europe. Biological Conservation, 142(5), 935-947. doi:10.1016/j.biocon.2008.10.018
Palamakumbure, D., Flentje, P., & Stirling, D. (2015). Consideration of optimal pixel resolution in deriving landslide susceptibility zoning within the sydney basin, new south wales, australia. Computers and
Geosciences, 82, 13-22. doi:10.1016/j.cageo.2015.05.002
Prieto, P., Peñuelas, J., Niinemets, Ü, Ogaya, R., Schmidt, I. K., Beier, C., . . . Estiarte, M. (2009). Changes in the onset of spring growth in shrubland species in response to experimental warming along a north–south gradient in europe. Global Ecology and Biogeography, 18(4), 473-484.
Schaap, J. 2015. The influence of climate change on heathland vegetation colonization and regeneration
in Oldebroek, the Netherlands.
Shaw, M. R., Zavaleta, E. S., Chiariello, N. R., Cleland, E. E., Mooney, H. A., & Field, C. B. (2002). Grassland responses to global environmental changes suppressed by elevated CO 2. Science, 298(5600), 1987-1990. doi:10.1126/science.1075312
Sowerby, A., Emmett, B. A., Tietema, A., & Beier, C. (2008). Contrasting effects of repeated summer drought on soil carbon efflux in hydric and mesic heathland soils. Global Change Biology, 14(10), 2388-2404. doi:10.1111/j.1365-2486.2008.01643.x
Vos, C. C., Nijhof, B. S. J., Veen, v. d.,M., Opdam, P., & Verboom, J. (2007). Risicoanalyse kwetsbaarheid
natuur voor klimaatverandering
Webb, N. R. (1998). The traditional management of european heathlands. Journal of Applied Ecology,
35(6), 987-990. doi:10.1111/j.1365-2664.1998.tb00020.x
Wessel, W. W., Tietema, A., Beier, C., Emmett, B. A., Peñuelas, J., & Riis–Nielsen, T. (2004). A qualitative ecosystem assessment for different shrublands in western europe under impact of climate change.
Ecosystems, 7(6), 662-671. doi:10.1007/s10021-004-0219-3
