Ground Beetles (Coleoptera Carabidae) as Bioindicators of the
Purmerbos
Research report Bachelorproject (thesis) 2017
Omar
O.I.
Elmecky
10772588
Male
omar.mecky@gmail.com
Front page image: Carabus Nemoralis, retrieved at: https://www.flickr.com/photos/atlapix/276140204 Online data repository: https://www.zenodo.org/record/824819#.WW4Q0lKp0uR Special thanks to supervisor Ben Brugge and coordinator Peter Roessingh.
Introduction
Bioindicators are species groups that can mirror the diversity of other species, give insights on the biotic and abiotic state of the ecosystem or give a representation of the influence of environmental change on ecosystems, communities and habitats (McGeoch 1998). The use of bioindicators has shown to be effective for reflecting the biotic and abiotic state of an environment in the past (Rainio and Niemelä, 2003). Though, research requirements for using ground beetles as bioindicator vary between studies. As Rainio and Niemelä (2003) stated, ground beetles have shown to be effective as bio indicator. However, are ground beetles still effective bio indicators when used for a forest in development, where the characteristics, including the fauna, are changing over time? This query leads to the overarching research question: “Can ground beetles be used as bio-indicator for young forests in development?” Because the Purmerbos is young and in development, this investigation tries to reveal the effectiveness of ground beetles in a developing environment, the Purmerbos was used as model for this research. The Purmerbos is a relatively young forest planted in 1987 close to the outskirt of Amsterdam (Nooren, 2010). It is a polder forest that growsAbstract
Bioindicators are species groups that can be an indicator for biotic or abiotic conditions of a certain ecosystem. The use of bioindicators has shown to be effective for reflecting the biotic and abiotic state of an environment in the past. However, whether ground beetles are still effective bio indicators when used for a forest in development, where the characteristics, including the fauna, are changing over time, has not been a point of elaboration. The aim of this research will be to identify the occurring ground beetle (Coleoptera Carabidae) species in the Purmerbos and to provide a answer on the research question: “Can ground beetles be used as bio-indicator for young forests in development?”. The Purmerbos is a young polder forest in development planted thirty years ago close to Amsterdam, and was used as model for this research. To investigate the ground beetle fauna in the Purmerbos, ground beetles will be caught in the forest and identified afterwards. Thereafter, results will be produced from statistical analyses of the data. The results provide no significant relationships due to a lack of data, but relationships between the ground beetles species and variables of the Purmerbos appear to be present. Results suggest that the age of a forest is not determinative for the effectiveness of the used bio indicator. Therefore, it can be stated that ground beetles can still be effective bio indicators when used for a forest in development.on a highly fertile clay soil. The overall tree species that make up the forest are ash, oak, popular, filbert and beech. The forest is fragmentized into separate vegetated patches and there is a high recreational pressure. Therefore, the development of the forest is likely to differentiate from that of a forest secluded from anthropological influences. Although the Purmerbos may look like a real forest in our eyes, it is expected not to be one in the terms of the species that occur there, because actual forest species would take much longer than thirty years to settle. And when the Purmerbos is missing most of the typically occurring forest species, it is ecologically speaking not a real forest. Since there is no existing literature addressing the ground beetle fauna of the Purmerbos, or one of the other newly constructed forests in the Netherlands it is interesting to investigate this forest. As it is expected that the Purmerbos is not actually a real forest considering the species that inhabit it, this research will investigate the distribution ground beetle species that cover the grounds of the Purmerbos to determine which species take in the place of the real forest species. As it is not clearly known and hard to predict what ground beetle species live in the Purmerbos, this research brings new insights on the distribution of ground beetles species that inhabit these types of young forests To investigate the effectiveness of ground beetles in the Purmerbos a set of statistical analyses was conducted. For example, the distribution of specialized (stenopic) species compared to that of unspecialized (eurytopic) species and the amount of caught species that are categorized as forest or eurytopic species, was a subject of investigation. It was expected that there is a bigger distribution of eurytopic species as these species spread more easily and are usually the first to colonize newly vegetated areas without a real particular preference, as opposed to stenotopic species which are specialistic and usually take longer to inhabit certain niches (Devineau, 2005). Also, the relationship between species that occur in shadowed areas and the percentage of canopy coverage is investigated, whereas such a relationship would contribute to the statement that ground beetles can be effective bioindicators. The same applies for the relationship between hydrophilic ground beetles and humid areas. Therefore these distributions and relationships will be investigated. Based on earlier research using ground beetles as bio indicators, and their effectiveness it is expected that at least a few relationships between the natural habitat preference of the beetles and matching variables of the Purmerbos will be found.
Research methods
Data resources For one of the statistical analyses done in during this research data form other forests than the Purmerbos was needed as reference material. Thus, for this research data from three different forests was used; the Purmerbos, the Amsterdamse bos and the Eyserbos. The Purmerbos, which is with 30 years the youngest forest, is lying east of Purmerend, a small village close to Amsterdam, The Netherlands (Nooren, 2010). The Amsterdamse bos is 83 years old and is located south from the centre of Amsterdam (Damen, 1973). The Eyserbos is located at Eys, Limburg, The Netherlands (Vorst & Heijerman, 2015). The age of the Eyserbos is not traceable in literature, only indications are sometimes given e.g. that some of the trees are more than 200 years old. For thepurpose of this research it was just accepted that the Eyserbos is the oldest of the three. In this research only data from the Purmerbos was collected. This data will be used for all the analyses, while the other two forests are exclusively used in a Principal Component Analysis (PCA). Data from the Amsterdamse bos was retrieved from a peer (Kiki de Waart), which conducted her own research in this forest. The data from the Eyserbos originates form earlier research and was provided by the supervisor (Ben Brugge). Sampling the ground beetles To investigate the ground beetle fauna in the Purmerbos, ground beetles were sampled. To get an as complete as possible impression of the ground beetle fauna in the Purmerbos, eleven capture series of five traps per series were spread across the this forest according to Figure 2. The exact locations of catchment were chosen to be as differentiating from each other as possible, so the broadest spectrum of environmental conditions was included. For every location, pitfall traps were dug into the soil in a straight-line formation with a distance of five meters between every trap. To make sure that the catch was as reliable as possible for representing the environment of the location, the two traps at the end were kept at a distance of about ten metres from the forest paths, so the chance of disturbance by humans was minimized. Every trap was marked with a red-white ribbon close to it, and the coordinates of the location were noted. The traps were spread across the forest on the 8th of March and were emptied every three weeks for a period of about two months (three times), whereas the last harvest was done on the 9th of May. Figure 2: Maps of Purmerbos with the marked locations of the series Setting up the pitfall traps For making the pitfall traps the research methods section of (Turin & Nieuwkerken,
2000) was adopted. As described by Turin & Nieuwkerken (2000), 500 ml yogurt cans with an opening with diameter of about 10 to 15 cm were used. The traps were filled to about one third with a 4% formaldehyde solution as preserving liquid. Thereafter one drop of a detergent (dish soap) was added to lower the surface tension of the preserving liquid. The pitfalls were dug into the soil in such a way that the edges of the cans were levelled with the surface ground. Afterwards, the pitfall traps were camouflaged and protected by a cover plate pinned into the ground with two long nails, just a few centimetres above the opening of the trap (Figure 3). A schematic picture is shown in Figure 4. The trap with cover plate was covered by leaves, and marked with three sticks parallel to each other on top of the cover plate. After emptying the traps, everything was put in place again, and the traps are refilled with the preserving liquid if needed. Figure 3: Setup of the pitfall trap with cover plate
Figure 4: Schematic picture of pitfall traps (Turin & Nieuwkerken, 2000) Identifying the ground beetles After emptying a trap the whole content was stored in the freezer at about -18 °C. Before the ground beetles were identified, the stored content of the bag was first thawed under a hot stream of water and then flushed through a sieve till only the coarse material, including the ground beetles, were left. The preserved organisms along with the small soil particles were put under a binocular to sort all the ground beetles. Then, the ground beetles were identified by following the identifying table provided by Muilwijk et al. (2015) till species level. Data analysis Table 1 presents the traits that were used during the statistical analyses along with their code and reference. Table 1: Overview of used traits with their code and reference.
Trait Code Reference
Eurytopy Ranging from 1 to 10 Turin et al. (1991) and Turin (2000)
Forest species W Lindroth (1949)
Eurytopic species EU Turin et al. (1991) and
Turin (2000) Hydrophillic species H Lindroth (1949) Shadow loving species F2 Turin et al. (1991) and Turin (2000) To assess if ground beetle fauna differentiates between different forests, which is a priority for a bio indicator because different forests have different
characteristics and therefore should contain different indicator species, the Purmerbos was compared to the Amsterdamse bos and the Eyserbos in terms of the ground beetle species found in each forest with a Principal Component Analysis (PCA). To investigate the distribution of specialized and unspecialized species, first simply the mean eurytopy of all the caught ground beetle species was calculated to see if the distribution of unspecialized (eurytopic) species compared to that of specialized (stenopic) species is as expected in a young forest like the Purmerbos. The eurytopy of a certain species was determined using the degree of eurytopy described by Turin et al. (1991) and Turin (2000), in Boeken et al. (2002) (Table 1). The degree of euytopy is described by a number ranging from 1 to 10, with 1 meaning very stenotopic and 10 meaning very eurytopic. Then, the relative amount of caught species that are categorized by Lindroth (1949) as forest (and usually specialist) species was calculated. For this the ecological group according to Lindroth (1949), described by Boeken et al. (2002), was used (Table 1). Also, the relative amount of caught species that are categorized by Turin et al. (1991) and Turin (2000) as “eurytopic species that occur in many of the 33 different area types”, was calculated. To reveal if there are more hydrophilic species occurring in the most humid sampling area, which is series 11, a histogram was made in which the number of hydrophilic species is shown per series. To determine if a certain species is thriving in humid areas, the ecological group according to Lindroth (1949), was used (Table 1). Furthermore, to determine if there is a relationship between species that occur in shadowed areas and the percentage of canopy coverage, a Two-sided Fisher's Exact Test (usual method using minimum likelihood) was done, followed by a logistic regression (similar to a linear regression, but applicable for binary data). To determine if a species is thriving in shadowed areas, the ecological group according to Turin et al. (1991) and Turin (2000), was used (Table 1). The percentages of canopy coverage that are being used are the result of the mean canopy coverage at which every species was caught.
For the all data analyses, excluding the analyses of the distribution of specialized and unspecialized species, the statistical computing program R (R Core Team, 2013) was used to carry out all the statistical tests.
Results
Variation between forests In Figure 5 the result of the Principal Component Analysis (PCA) is provided. Three forests were included in this analysis; the Purmer bos, the Amsterdamse bos and the Eyserbos. The values of PC1 on the x-axis and PC2 on the y-axis are determined by the biggest, and the second biggest part of the variation between the species quantities of every series from the three forests together, respectively.The spatial distribution of the arrows in the plot represents the variation in ground beetle composition between the different series. Thus, when two arrows are equal in length and direction, their species composition and quantities should be comparable. The red dots each represent a ground beetle species; the further a certain dot is present from the core, the more determinative it is for the variation between the series.
Figure 5: Principal Component Analysis showing all the series from the Purmerbos (P1 till P11), the Amsterdamse bos (A1 till A9) and the Eyserbos (E1 till E5) as arrows and represents the variation in species composition between series. The red dots are representing the ground beetles species. Database During this research fieldwork was conducted in which ground beetles were caught using pitfall traps spread across the Purmerbos. After the fieldwork was conducted the database was build. In total, a set of 23 ground beetles species was caught. Table 2 provides the ground beetle species and the total caught amount. All the used data for
this research including the full database with all the caught species of the Purmerbos and Amsterdamse bos per series described with more variables is provided at the online database at: https://www.zenodo.org/record/824819#.WW4Q0lKp0uR Table 2: All caught ground beetle species, the total amount per species, the degree of eurytopy and the Ecological group by Lindroth (1949) and Turin et al. (1991) and Turin (2000). Species Total
amount Eurytopy Eco group LI Eco group TU Agonum emarginatum 1 7 H1 G4 Amara comunnis 1 10 N1 EU Asaphidion curtum 10 7 NA F2 Bembidion biguttatum 11 5 H1 G4 Bembidion femoratum 1 6 H2 H3 Bembidion harpaloides 1 4 H1 G4 Carabus granulates 2 7 H2 G2 Carabus nemoralis 65 8 N1 EU Clivina fossor 3 9 H2 EU Leistus fulvibarbis 8 4 HW D2 Limodromus assimilis 206 7 HW EU Loricera pilicornis 33 10 NH EU Nebria brevicollis 42 10 N1 EU Notiophlus rufipes 26 7 W1 D3 Ophonus rufibarbis 1 5 NX D1 Oxypselaphus obscurus 1 8 HW EU Paranchus albipes 5 4 H1 H2 Pterostichus anthracinus 2 4 H1 G4 Pterostichus melanarius 5 9 N1 EU Pterostichus oblongopunctatus 281 7 W1 D3 Pterostichus strenuus 37 9 NH EU Pterostichus vernalis 3 9 NH EU Trichocellus placidus 1 8 NW F2 The Limodromus assimilis and Pterostichus oblongopunctatus appear to be the most dominantly present across the series in the biggest numbers. Other species are only represented by a single individual. Distribution of specialized and unspecialized species The mean degree of eurytopy was calculated to be 7. Furthermore, 26% of the caught species were forest species according to Lindroth (1949). Furthermore, the Pterostichus oblongopunctatusis, a forest species, was caught the most. Finally, 43% of the ground beetles species are eurytopic according Turin et al. (1991) and Turin (2000).
Indicator for humid conditions A histogram with absolute amounts of hydrophillic ground beetles caught per series is provided below (Figure 6). Figure 6: Histogram with absolute amount of hydrophilic species per series. Series 11 is the most humid location. Table 3 below provides both the absolute and the relative amount of Lindroth “H” category ground beetles caught per series. Table 3: absolute and relative amount of hydrophilic species per series
Series
Absolute amount of
hydrophilic species Relative amount of hydrophilic species
P1 3 20%
As can be seen in Figure 6 and Table 3, series 11 does not have the most hydrophilic species, in absolute and relative numbers.
Indicator for canopy coverage A table with canopy coverage percentages, derived from the mean canopy coverage in which the different species were caught, and the amount of shadow loving, or Turin “F2” category ground beetles caught for every canopy coverage. The different canopy coverage percentages are each a result of the mean canopy coverage of a species. Thus, every data point in this table is representing one Turin “F2” category species. Only two Turin “F2” category ground beetles species were caught during the fieldwork, the Trichocellus placidus and the Asaphidion curtum, and they are caught at a mean canopy coverage of 90% and 95% respectively. Table 4: Amount of shadow loving species per canopy coverage in percentages Mean canopy coverage in % Amount of shadow loving species 15 0 20 0 50 0 52.5 0 54.5 0 55 0 59 0 61 0 63 0 64.5 0 65 0 70 0 72 0 73 0 90 1 95 1 100 0 To reveal if there is a significant relation between the distribution of the mean canopy coverage percentages and that of the shadow loving ground beetle species, a Two-sided Fisher's Exact test was done. The p-value produced by the Two-sided Fisher's Exact test, P3 4 44% P4 3 43% P5 2 33% P6 4 44% P7 3 38% P8 2 40% P9 0 0% P10 1 16% P11 2 22%
which 1, is not approaching significance (p-value < 0.05), this means that no assumption can be made that the mean canopy coverage percentages and shadow loving ground beetle species are causally linked. The alternative hypothesis that the test provides states that the true odds ratio is not equal to 1. The Odds Ratio (OR) is infinity and the 95% Confidence Interval (95% CI) of the OR is 0.0526 till infinity. Because the 95% CI contains OR=1, there is no reason to reject the null hypothesis that there is no relation between the distributions of shadow loving beetles and canopy coverage. In Figure 7, the shadow loving ground beetle species and the mean canopy coverage in percentages are presented in a plot. A y-axis value of 1 corresponds to a Turin “F2” category ground beetle species, and therefore to shadow loving species, and 0 to non-shadow loving ground beetle species. Thus, “shadow loving species” on the y-axis is a binary variable. The total of 23 caught ground beetles species are represented by the data points distributed in the plot. Figure 7: Binary plot of the Turin “F2” category ground beetle species and the other species and the mean canopy coverage in precentages A logistic regression of the plot in Figure 7 is conducted to, again, to test if there is a significant relation (regression) between the mean canopy coverage percentages and the shadow loving ground beetle species. The p-value with meaning regarding this research that resulted out of the logistic regression is 0.182, which is not significant (p < 0.05). This implicates that there is no indication in the used data that there is a relationship between the mean canopy coverage percentages and the shadow loving ground beetle species.
Discussion
The main objective of this research was to reveal if ground beetles can be reliable bio indicators for the Purmerbos. To investigate this, a set of diverse statistical analyses was run to detect relationships between the natural habitat preference of the beetles and matching variables of the Purmerbos. First, to investigate the variation in species composition between different forests, a PCA was done using the catchments of the Purmerbos, the Amsterdamse bos and the Eyserbos. The PCA illustrates that series originating from the same forest are clustering, meaning that they are similar in species composition. Only the species composition of one series of the Eyserbos (E1) tends to be more related to the series of the Purmerbos than other series of the Eyserbos. Thus, this PCA shows that the different forests have a different ground beetle species composition, suggesting that based on the ground beetle fauna certain characteristics of the forest could be determined. This supports the claim that ground beetles can be a bio indicator for the Purmerbos. Second, the distribution of ground beetle species in the Purmerbos was investigated to reveal the species composition of a young forest, where specialists should theoretically not be abundantly present. In total 23 different species were caught (Table 2). The mean degree of eurytopy was 7, whereas a degree of eurytopy of 7 was described as “quite eurytopic” (Turin et al., 1991 & Turin, 2000). Furthermore, a big share of 46% of the ground beetles were categorized as eurytopic species. A smaller 26% of the beetles were categorized as forest species, which are usually specialists. Though, in some cases there is conflict between the ecological groups according to Lindroth (1949) and Turin et al. (1991) and Turin (2000). For example, the Limodromus assimilis is categorized by Lindroth (1949) as “HW”; a hydrophilic forest species, while it is categorized by Turin et al. (1991) and Turin (2000) as “EU”; eurytopic species. These descriptions suggest a controversy and therefore it is important to note that there appears to be some disagreement between these authors. However, the amount of eurytopy is as expected from a forest of this age, meaning that the distribution of ground beetles species can be a indicator for age. Nevertheless, 26% forests species is still higher than expected from a 30 year old forest. Additionally, the Pterostichus oblongopunctatus was the most abundantly present species and is a forest species. How these specialists have managed to populate this area would be a focus of further research. Third, to see if the caught ground beetles could be an indication for humid conditions in the Purmerbos, a histogram was made showing the amount of hydrophilic species caught per series (Figure 6). Based on the knowledge that series 11 was definitely the most humid location, the expectations were that in this series the amount of hydrophilic species would be the highest. However, this is not the case. In absolute numbers, but also relatively (Table 3) series 11 was one of the series with the least hydrophilic species. Therefore, this result is not contributing to the evidence that ground beetles are a reliable bio indicator for the Purmerbos. Fourth, to reveal if the caught ground beetles could be an indication for canopy coverage in the Purmerbos, first a table was made (Table 4) that shows at which (mean) canopy coverage the shadow loving species occurred. Just from looking at this table it seems that there is a relationship between “F2” categorized species (shadow loving species) and the canopy coverage of the Purmerbos, as the two shadow loving species are
occurring at a mean canopy coverage of 90% and 95%. To test if this relationship is significant; when there is a p-value (chance) of lesser than 0.05 that there is no relationship between the current distributions of the two variables, the Two-sided Fisher's Exact test was conducted on Table 4. This test showed a p-value of 1, meaning that no assumption can be made that there is a relationship between shadow loving species and canopy coverage in the Purmerbos. Then, a binary plot was made that shows, both the shadow loving and the non-shadow loving species distributed over the mean canopy coverage percentages. A logistic regression was done to, again, check if the relationship between shadow loving species and canopy coverage is significant. The p-value of this logistic regression was 0.182. This implies that based on the data, there is no indication for a relationship between shadow loving species and canopy coverage in the Purmerbos. The reason that both the Two-sided Fisher's Exact test and the logistic regression did not give significant p-values, is most likely due to the fact that not enough data was provided to confirm significance. To conclude, the eurytopy of the ground beetle fauna of the Purmerbos is somewhat as expected from a young forest. Also, in this research no evidence was found that ground beetles can be a good indicator for humid conditions, the relationship between shadow loving species and canopy coverage seems to be present but could not be significantly confirmed and convincing results claim that different forests are characterized by different ground beetle species composition. The main research question was: “Can ground beetles be used as bio-indicator for young forests in development?” When all the results and their meaning are taken into account, it could not be stated that ground beetles are, in the case of the Purmerbos, a reliable bio indicator. However, the sampling period of this research was about two months, whereas the actual minimum length for the sampling period to reliable is about a year. Therefore, due to the short sampling period in this research and the relatively low numbers of ground beetles that have been caught in this period, significant relationships are hard to reveal. Based on the fact that relationships between the ground beetles species and variables of the Purmerbos appear to be present, with exception of the relationship between hydrophilic species and humid conditions, ground beetles could indeed be used as bio indicator for the Purmerbos, provided that more data is collected during a longer sampling period of at least one year. Thus, results suggest that the age of a forest is not determinative for the effectiveness of the used bio indicator. Therefore, it can be stated that ground beetles can still be effective bio indicators when used for a forest in development.
References: Boeken, M., K. Desender, B. Drost, T. van Gijzen,. B. Koese J. Muilwijk, H. Turin & R.J. Vermeulen. ( 2002). De Loopkevers van Nederland & Vlaanderen (Coleoptera: Carabidae). Stichting Jeugdbondsuitgeverij, Utrecht: 212 pagina’s. Damen, J. P. N. (1973). Een onderzoek naar de relatie bezoek-herkomst, de bezoek frequentie en enkele inrichtingsaspecten van het Amsterdamse Bos (No. 716). Instituut voor Cultuurtechniek en Waterhuishouding. Devineau, J. L. (2005). Generalist versus specialist: a contrasted sociology of woody and herbaceous species in a fallow-land rotation system in the West African savanna (Bondoukuy, Western Burkina Faso). Phytocoenologia, 35(1), 53-78. McGeoch M. 1998. The selection, testing and application of terrestrial insects as bioindicators. Biological Reviews 73: 181–201. Muilwijk, J., R. Felix, W. Dekoninck & O. Bleich. (2015). De loopkevers van Nederland en België (Carabidae). Nooren, M. J. (2010). Staatsbosbeheer. De Levende Natuur, 111(2), 118-118. R Core Team. (2013). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. Rainio, J., & Niemelä, J. (2003). Ground beetles (Coleoptera: Carabidae) as bioindicators. Biodiversity and conservation, 12(3), 487-506. Turin, H., & Nieuwkerken, E. J. (2000). De Nederlandse loopkevers: verspreiding en oecologie (Coleoptera: Carabidae) (Vol. 3). Nationaal Natuurhistorisch Museum. Vorst, O., & Heijerman, T. (2015). Enige aantekeningen over Quedius dilatatus in Nederland (Coleoptera: Staphylinidae). Entomologische berichten, 2-10.
