Spatial distribution of ant nests and their influence on soil
properties in a semi-arid catchment in Spain
BSc Thesis
Image: Impression of fieldwork location (De Jong, 2017).
Marle de Jong (10786457)
Date: June 30st, Amsterdam
Future Planet Studies, Earth Sciences, University of Amsterdam
Supervisor: dhr. dr. L.H. Cammeraat
Abstract
Ants can be considered among the most important soil engineers in semi-arid areas. Within this research, the effects of the activity of seed harvesting ants (Messor bouvieri and Messor barbarus) on several physical and chemical soil properties were investigated in a semi-arid catchment in SE Spain. Water repellency, (macro)-nutrients concentrations, pH, electrical conductivity, organic carbon content and nitrogen content are measured by different methods. Besides this, spatial distribution and preferences of ant nests mounds related to lithology and soils are analysed, which is highly important, as nutrient and energy flows may be organised with the spatial distribution of organisms as ants. The density is calculated and the pattern is analysed by nearest neighbour analysis.
Nest mounds (0-10 cm) had a significantly lower pH than control soil (0-10 cm). Electrical conductivity, organic carbon, nitrogen and moisture content were significantly greater in soil from nest mounds, as well as concentrations of potassium, sodium, magnesium and calcium. No significant differences were found between the top layer (0-5 cm) and the uppermost layer (5-10 cm). Nest density of the area was extremely sparse with 3.54 nests ha-1, and highest densities occurred at the calcrete and calcisol units. The distribution of the ant nests can be considered as clustered, and does not differ between different units of lithology and soil.
In conclusion, it is confirmed that ants can be considered as ecosystem engineers and rejuvenators and their spatial distribution may have consequences for the strengthening of altering chemical and physical soil properties as well as the development of fertile islands in semi-arid ecosystems.
Keywords: Messor bouvieri, Messor barbarus, soil chemistry, Mediterranean ecosystems, ecosystem engineers,
Table of contents
Figures and tables
4
1. Introduction
5
2. Site and methods
7
2.1 Site location
7
2.2 Method: Spatial analysis
8
2.3 Method: Chemical and physical properties of soil
9
3. Results
11
3.1 Spatial analysis of ant nest locations
11
3.1.1 Pattern analysis
3.1.2 Nest density
3.2 Lab analysis
14
3.2.1 Summary of lab results
4. Discussion
18
5. Conclusion
20
6. References
21
7. Acknowledgements
24
Figures and tables
FiguresFigure 1: A catena over a limestone mountain ridge with abutting pediments. Numbers refer to slope sections and soil characteristics of these sections are given in Table 1. (Imeson and Cammeraat, 1999)
Figure 2: Fieldwork catchment area (ArcGis). Surface area of red square is 9.6 ha. (ArcGIS) Figure 3: Map of the pattern of ant nests. Red dots are ant nests locations (34 nests in total). Figure 4: Visualization of ant nest locations with convex hull (MatLab) Figure 5: Elbow method of k-means clustering convex hull (MatLab).
Figure 6: Clusters in ant nest locations (MatLab) Figure 7: Kernel density estimation of nest locations Figure 8: Visualization of compared combinations
Tables
Table 1: Hypotheses and consulted literature which confirm hypotheses.
Table 2: Soil characteristics along a catena on the Alqueria hill about 10 km north of the Puentes Reservoir. The numbers correspond to Figure 3 (Imeson and Cammeraat, 1999; Cammeraat, 2017).
Table 3: Resume of lab analysis Table 4: Patterns within lithology units Table 5: Patterns within soil units
Table 6: Nest density within lithology units Table 7: Nest density within soil units
Table 8: Properties soils of ants’ nest mounds and control sites in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing 0-5 cm
Table 9: Properties of soils of ants’ nest mounds and control sites in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing 5-10 cm
Table 10: Properties of soils of ants’ nest mounds in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing crust (0-5 cm) and subcrust layer (5-10 cm) Table 11: Properties soils of ants’ nest mounds and control sites in the catchment area (mean ± SE) and significance of
comparison between sites by a Wilcoxon signed rank test (p), comparing 0-10 cm Table 12: Summary of lab results with significant differences.
1. Introduction
Soil fauna plays an essential role in the ecosystem (Lavelle, 1996) as they provide an interaction with the water and air in the soil and surface, and they regulate the fragmentation and decomposition of organic matter (Hole, 1981). Ants are one of the most important soil engineers, as they denude vegetation (Crist & Wiens, 1996) and have an important impact on soils physical, chemical and biological properties by creating macro-voids, galleries and chambers and by changing the composition of carbon, nutrients and soil microbes within their nests (Frouz & Jilková, 2008). In recent years, the importance of ants in terms of biodiversity and their roles in ecosystems is more and more highlighted (Folgarait, 1998; Frouz & Jilková, 2008; Cammeraat & Risch, 2008). Especially in sub-humid and semi-arid areas, as in SE-Spain, ants are significantly important (Cammeraat & Risch, 2008). The activity of ants as burrowing and mounding can result in an increase in aggregates, which has a positive effect on the infiltration of water. Infiltration of water is highly important for biological activity in semi-arid landscapes. On the other hand, ants may decrease infiltration by producing compacted surfaces, which facilitate runoff (Cammeraat et al., 2002). The effect differs by species and more research is needed on the effect of different species on soil processes as structure, aggregate stability and other soil processes and properties (Six et al., 2004). For example, previous research has shown that ants can significantly influence soil properties as pH, concentrations of nutrients, structural stability and water repellency (Cammeraat et al., 2002; Cammeraat & Risch, 2008).
As stated above, it is already known that ants could affect different soil chemical and physical properties (Frouz & Jilková, 2008; Six et al., 2004), although more research is needed (Six et al., 2004). The aim of this research is therefore to measure the chemistry and the soil wettability of soil associated with nests of the seed harvesting ant species Messor bouvieri and the Messor barbarous in Southeast-Spain (Alméria Province).
However, the effect of their changing properties also depends on spatial distribution (Cammeraat & Risch, 2008). For example, research of Whitford & DiMarco (1995) showed that ant nests in slope bottom areas did not show significant differences in soil properties, whereas on the hillslope, they found larger and significant differences between nests and surrounding soil. They can influence pedological, hydrological and
geomorphological processes at larger scales as hillslopes or catchments (Whitford & DiMarco, 1995; Cammeraat & Risch, 2008). It is known that their pattern and density affects the distribution of resources and organise the energy and nutrient flow throughout ecosystems (Crist & Wiens, 1996), but this information is scarcely available (Frouz & Jilková, 2008).
Therefore, the second objective of this research is to analyse the spatial distribution and preferences of seed harvesting ant nests (Messor bouvieri and Messor barbarus) in a semi-arid catchment in Spain. It is essential to gain knowledge about the spatial distribution of organisms as ants to understand community interactions and ecosystem functioning. The focus in this study will be on the abiotic factors as soil, topography and lithology. The following research questions will be answered in this research:
1. Do ants significantly influence chemical and physical soil properties within the catchment and is this
different by depth (comparing 0-5 cm and 5-10 cm)?
Do ants have an influence on the pH, organic carbon, nitrogen and on the concentration of (macro)-nutrients potassium (K+), sodium (Na+), magnesium (Mg2+), calcium (Ca2+) and
phosphorus (P3+)?
Do ants have an influence on the moisture content, soil water repellency and conductivity?
2. What is the spatial distribution of seed harvesting ant nests in Southeast-Spain and why is this important
for the ecosystem at a catchment scale?
3. Are there preferences of the ant nest mounds related to lithology and soil type?
It is decided to measure these soil parameters because it is already known that ants could affect these properties significantly, based on the literature review of Cammeraat & Risch (2008). These parameters are also of great importance when assessing soil properties and quality. When surface cover is low, soil properties play a major role in the processes of infiltration and erosion (Scoging, 1989). Besides this, the proposed basic cations (K, Ca, Mg) can be easily measured, however they are important parameters in soil processes as cation exchange capacity (Frouz & Jilková, 2008).
Hypotheses
The different hypotheses are summarized in Table 1. Reasons why these results are expected are explained in the Discussion (p. 17).
Expected result ant nest mound
compared to control site Reference(s)
pH Lower pH in alkaline soil Folgarait (1998); Cammeraat et al., (2002); Wagner, Jones & Gordon, (2004); Organic carbon and
nitrogen Higher organic carbon and nitrogen level Lobry de Bruyn & Conacher (1990) Conductivity Higher conductivity Cammeraat et al (2002)
Concentrations of K, Na, Mg, Ca and P
Higher concentration of these nutrients Azcárate & Peco (2007); Cammeraat et al. (2002)
Water repellency More water repellent Cammeraat et al. (2002)
Moisture content Higher soil moisture Cammeraat & Risch (2008); Day & Ludeke (1993)
Table 1: Hypotheses and consulted literature which support hypotheses.
It is expected that this influence increases with depth, as the seeds are taken deep into the nests and ant activity is higher with increasing depth. Regarding the spatial distribution, it is expected that the nests are regularly distributed (uniform). This is based on the study of Levings & Traniello (1981), summarizing 160 studies involving 136 ant species. Finally, it is expected that ants do not prefer shallow soils (as leptosols) and they prefer marl, as this is finer grained and less hard material than limestone.
In conclusion, an analysis of the role of ants as ecosystem engineers, focused on the role of ants on different soil properties and on their spatial distribution can hopefully be made. At first, specifications of the research site and used methods will be explained and subsequently, the results of the research will be visualized and described. After this, a conclusion regarding the research questions will be given, as well as giving remarks and explanations for and on this study in the discussion.
2. Site and methods
2.1 Site locationThe research site is located in the province of Guadalentín Basin, Alméria Province (37°47’N, 1°50’W). The climate is semi-arid with an annual precipitation of 270 mm, and yearly potential evapotranspiration of 950 mm (Navarra Hervas, 1991). A high deficit occurs during summer, as rain falls mainly in spring and autumn. The average temperature ranges between 16 and 18 °Celsius (Navarra Hervas, 1991). For this research, information on the geology, lithology and soils within the study area is important.
The geology and geomorphology is strongly influenced by fold and fault structures (Cammeraat, 2002). The study area is part of the Betic Cordillera, the most western part of the Alpine Orogenic system. This zone was formed during the Cretaceous period through to the Tertiary period (the Miocene Epoch) (Cammeraat, 2017). The site is part of the Subbetic Zone (Geerlings, 1978). Marls, white and pink marly (sandy) limestones can be found within the area (Cammeraat, 2017). Besides this, the dominating lithology consists of more recent calcareous Holocene, Pleistocene and Late Tertiary deposits (Geerlings, 1978). The (green) marls in the lower areas belong to the Green Pelite Formation, whereas the white to green marls were deposited during the Lutitian Formation (mid-Eoceen). On top of the Green Pelite Formation, different types of limestones can be found. The area consists of blue, resistant limestones and less-resistant yellow limestone strata with marly characteristics (Geerlings, 1978). A calcrete zone can be found in the middle of the catchment (blue square, Figure 2, BSc Thesis Olaf de Haan). Remains of this calcrete can be found in other areas within the catchment. A calcrete is formed as a result of CaCO3-precipitation in
soil due to the evaporation of groundwater, and an impermeable calcrete crust is formed (Cammeraat, 2017). The average altitude ranges between 500 – 1200 m (Lambregts, no date). Partly due to the height differences, soils are generally very shallow or eroded (Imeson and Cammeraat, 1999), with low aggregate stability and low organic matter (Imeson et al., 1998). Mainly leptosols and regosols can be found, which are characterized by an AC-horizon. Calcisols can be found on the calcrete zone, and are recognized by the presence of a petrocalcic horizon (WRB, 2014). The soils in the fieldwork area can be considered as alkaline with a high CaCO3 content. Imeson and
Cammeraat (1999) have performed previous research within the same area and in Table 2, representative soil profiles for the limestone and marl areas can be found, corresponding to the catena in Figure 1.
Figure 1: A catena over a limestone mountain ridge with abutting pediments (Imeson and Cammeraat, 1999). Numbers refer to slope sections and soil characteristics of these sections are given in Table 2.
Table 2: Soil characteristics along a catena on the Alqueria hill about 10 km north of the Puentes Reservoir. The numbers correspond to Figure 1 (Imeson and Cammeraat, 1999; Cammeraat, 2017).
In areas with higher slopes (midslope and upsole), mainly leptosols can be found (Table 2). Regosols can be found in areas with a smaller slope (footslope and one in the midslope). It is expected that more calcisols can be found in the research area due to the formation of the calcrete zone.
Different ant species can be found in the area. Three main ant species have been collected in the field and are brought to Jinze Noordijk, research associate EIS Kenniscentrum Insecten, for determination. The three species are the Messor Bouvieri, Messor Barbarus and the Formica cf. rufibarbaris. The ant species M. Bouvieri and M.
Barbarus are commonly found in arid sites throughout the western Mediterranean and Northern Africa (Bernard,
1968). These ant species are seed harvesting ants (Cerda & Retana, 1994) as they collect seeds and mix them with moisture (Mierenboerderij, 2017). They play an important role in the ecosystem as they increase seed dispersal, provide nutrients for plants, provide soil aeration through their galleries and chambers and they mix deep and upper layers of soil (Uppstrom & Klompen, 2011). The Formica cf. rufibarbaris is a common type of ants throughout continental Europe, however, they are not considered as seed harvesting ants. This species occurred less frequently in the research area compared to the other two determined species.
2.2 Method: Spatial Analysis
In April 2017, a fieldwork period of five days has been conducted. Active ant nests are registered through the Collector App of ArcGIS (GPRS) and detailed maps of the soil types and lithology have been made. Figure 2 shows a detailed map of the research area (see Appendix 1). The upper area of the catchment (red square) has been vertically examined for ant nests, thereby reducing the chance of missing areas and possible nests.
A nearest neighbour analysis is conducted with the coordinates of the nests to analyse spatial distribution. The distances between each centroid and its nearest neighbour is measured, and observed mean distance is calculated. The expected nearest neighbour distance for a random point pattern is de = 1/(2√(n/A)) with n the number of points
and A the area of study region in km2. The ratio (R) is calculated by dividing the observed with the expected
neighbour distances. The result will produce a result between the 0 and 2.15, whereby a R close to zero means clustered, 1 means random, and around 2.15 means a dispersed pattern (Burt, Barber and Rigby, 2009). The code for this method can be found in Appendix 3E. Nest density is calculated by obtaining data from ArcGIS (square kilometres) combining with frequency tables.
2.3 Method: Chemical and physical properties of soil
Samples of approximately 50 grams were taken from nest mounds and a control site at approximately 2 m distance (11 locations, Appendix 2A). At every nest mound and control site, a soil sample of 0-5 cm and 5-10 cm has been collected. The control site must be at least 2 m from the nearest shrub, as it is known that shrubs can affect soil properties (Pugnaire et al., 1996). Samples are sieved in the field, over a sieve of 2 mm and half the weight of the samples are dried at a temperature of 60 degrees Celsius to remove soil moisture. The dried samples are milled in the milling machine (Fritsch Pulverisette 5).
Samples were analysed (IBED labs) and chemical analysis of the different parameters is summarized in Table 3.
Table 3: Resume of lab analysis. pH and
conductivity
pH & conductivity meter
suspension of unmilled soil in demineralized water (preferably 1:2.5 soil/water by volume). Shake for two hours (120 rotations/minute), rest overnight, and need to be put on the shaking plate again for 30 minutes, before measuring the pH and EC.
Organic carbon & nitrogen
CNS analysis (Thompson, 2008) & Van Wesemael method
Total carbon and nitrogen: a sample (50 mg) is burned in an excess of oxygen and various traps, collecting the combustion products: carbon, nitrogen and sulfur (Thompson, 2008). CNS analysis is conducted in duplo.
Inorganic carbon: Van Wesemael method is based on the release of CO2 from carbonates after addition of hydrochloric acid (4 M) to milled sample (2 grams). The Erlenmeyer is weighed before and after the reaction (26 hours on a shaking plate). As a reference, carbonate is used (NEN, 1993; Wesemael, 1955).
Organic carbon is calculated by subtracting the inorganic carbon from the total carbon. Moisture content Oven
Milled sample (1 grams) is put in the oven at 105 °C. After 24 hours, the samples need to cool down in an exicator and weighted again. Formula (3) shows the calculation of the moisture content.
Soil water repellency
Wetting depth penetrating method
1.5 grams of unmilled sample is spread out and droplets of distilled water are placed onto the surface (pipette 1 ml), time for complete infiltration is recorded.
Ion concentrations ICP-analysis
Diluted soil samples (1:2.5 soil/water by volume) are filtered by using 0.2µm membrane filters. Suspensions are prepared with 8 ml ELGA water, 1 ml of filtered soil sample, 0.5 ml yttrium as standard and 0.5 ml caesium to suppress ionization of natrium.
Notes:
Organic carbon:
Formula (1) and (2) show the calculation of the Wesemael method. Inorganic carbon calculated by the percentage C in CO2, which is approximately 27,3% (molar mass C is 12.0107 and CO2 44.0095
g/molecule).
Formula1 :%CO
2=
P× gram CaCO
3×44
Q ×100 × R
× 100 %
P = weight loss sample in gramsQ = weight loss CaCO3 in grams
R = dried sample weight in grams
Formula2 :%CaCO
3=%CO
2×
100
44
The moisture content needs to be considered for the correct calculation of R. The experiments need to be conducted preferably on the same day.
Moisture content
Formula3 :% moisture=
weight cup∧air dried sample−weight cup after drying
A TOC-analysis has been performed, which also gives the total, organic and inorganic carbon (see
Appendix 2F). Approximately 5 ml of diluted sample is (1:2.5 soil/water by volume) is followed by
determination in the Shimadzu machine. However, the results of the CNS and Van Wesemael will be used for the calculating of the organic matter as it is known that this gives a more precise result (Van Hall, personal correspondence, May 2017). The TOC outcomes will be analysed to see if this gives the same trends as the CNS and Van Wesemael.
Data is analysed for outliers and statistical analysis is performed by a Wilcoxon signed rank test. This test is a non-parametric statistical hypothesis test when comparing to paired samples and it analyses whether their population means differ. Data is checked for normal probability by producing normal probability plots and performing Kolmogorov-Smirnov tests in MatLab. Assumptions are that the data is paired and comes from the same population, each pair is chosen randomly and independently and the data is measured at least on an ordinal scale (Burt, Barber and Rigby, 2009).
3. Results
This section is split in two sections, namely the results of the spatial distribution and the chemical and physical analysis of nest mound. These results will be presented subsequently. Detailed results can be found in Appendix 2A-G (lab analysis) and 3A-H (spatial distribution).
3.1 Spatial analysis of ant nest locations
Within this section, the results of the pattern analysis and nest density will be presented subsequently. A map of the pattern can be seen in Figure 3.
Figure 3: Map of the pattern of ant nests. Red dots are ant nests locations (34 nests in total). 3.1.1 Pattern analysis
The pattern of the nests is analysed by calculating the nearest neighbour ratio (Appendix 3E, 3F for coordinates). The considered area is 9.6 ha. The ratio is close to zero (2.5609 x 10-5), meaning the pattern of the nests can be
classified as clustered. A visualization of the nest with their locations (x and y) can be seen in Figure 4.
Figure 4: Visualization of ant nest locations with Figure 5: Elbow method of k-means clustering (MatLab) convex hull (MatLab).
Based on this conclusion, the probable number of clusters within the research area can be calculated using the elbow method. The number of clusters is calculated by using the total sum square distances (tSSd) between each point location and the centre of its respective cluster (cluster’s centroid). One can determine when the inclusion of another cluster does not improve the tSSd. However, the correct choice of the k (clusters) is rather ambiguous. A k of 5 is decided as can be deduced from Figure 5 that the tSSd does not significantly improve from this amount of
cluster. The resulting clusters are visualized in Figure 6. Colours in the figure are randomly chosen. It should be kept in mind that these clusters are based solely on the location.
Figure 6: Clusters in ant nest locations based on their location (MatLab)
The largest cluster is the yellow cluster, which is in the northern part of the fieldwork area. This big cluster is situated on a mix of regosols and leptosols on a limestone bed (Appendix 3C/3D).
Secondly, the pattern is analysed for each unit of lithology and soil, thereby checking whether there is a different pattern for each unit. Appendix 3G shows which nest locations are covered within each unit. Separate simplified maps of the lithology and soil maps have been made (Appendix 3C and 3D), based on three main types for each parameter, namely calcrete, limestone and marl for the lithology, and regosol, calcisol and leptosol for the soil map. The nearest neighbor ratio is calculated and the results can be seen in Table 4 and Table 5.
Table 4: Patterns within lithology units
Unit Area A (km2) Nearest neighbor ratio Pattern
Limestone 0.0678 2.67 x 10-5 Clustered
Marl 0.0216 No pattern No pattern
Calcrete 0.0043 9.72 x 10-6 Clustered
For the marl unit, no pattern can be determined. This is because only two nest locations are covered by this unit, and patterns are not visible if n is 2. The patterns for limestone and calcrete can be determined as clustered, as the ratio is close to zero.
For the pattern at the different soil types, the overall nearest neighbour ratio is calculated (Table 5). Table 5 shows that the pattern is clustered for each unit as all ratios are close to zero.
Besides this, the ratio for the largest contiguous area with the most ant nest locations on this soil type is calculated as this might differ from the analysis of Table 5 (see Appendix 3H for further explanation and calculations). However, it becomes clear that the analysis of separate polygons also gives a clustered pattern.
Unit Area A (km2) Nearest neighbor ratio Pattern
Leptosol 0.0648 5.37 x 10-5 Clustered
Calcisol 0.0155 1.96 x 10-5 Clustered
Regosol 0.0152 8.211 x 10-5 Clustered
3.1.2 Nest density
The examined area (Appendix 1A) corresponds to approximately 0.096059 km2 (9.6 hectare), and 34 nests are found
within this area. This equals a nest density of 3.54 nests/ha.
A kernel density estimation of the points has been made (Figure 7). Kernel density estimation places a symmetric kernel on top of each point. The kernel describes the shape of the probability distribution around ant nest. For instance, a uniform distribution assumes that the point could have equally occurred at any position in the kernel distribution. The probability surface is then determined by summing the (potentially overlapping) kernels on top of each point in space.
Figure 7: Kernel density estimation of nest locations
The highest point density is in the middle of the fieldwork area, concluded from the highest kernel probability surface in the Figure 7. These high densities occur on the leptosol and calcisol, on a limestone and marl bed (Appendix 3C/3D).
The nest density is calculated for each unit of lithology and soil as well.
Table 6: Nest density within lithology units
Unit Area (ha) Nests Nest density
(nest/ha)
Limestone 6.78 26 3.83
Marl 2.16 2 0.93
Calcrete 0.43 6 14.0
Table 7: Nest density within soil units
Unit Area (ha) Nests Nest density
(nest/ha)
Leptosol 6.46 16 2.48
Calcisol 1.55 12 7.74
Regosol 1.52 6 3.94
The highest densities are on the calcrete zone (lithology) and on the calcisol (soil). The density on the calcrete is approximately 14 times higher than on the marl, and the density on the calcisol is approximately 3 times higher than on the leptosol. The least nests are located on the marl and on leptosols.
3.2 Lab analysis of nest mounds
Four combinations of sites and soil depths are compared to answer the research questions which are visualized in Figure 8.
- Nest site (0-5 cm) and control site (0-5 cm) - Nest site (5-10 cm) and control site (5-10 cm) - Nest site (0-5 cm) and nest site (5-10 cm) - Nest site (0-10 cm) and control site (0-10 cm)
In the following tables, the mean, standard deviation and the p-values of the different combinations and parameters are shown. It is decided to present the data with and without outliers, as small sample size and therefore the number of outliers might influence this research. When data is reported without outliers, it can be assumed that no outliers are present within this data set. The parameters which can be significantly influenced by outliers are mentioned, for the rest of the parameters it can be assumed that these outliers do not influence the significance of comparison between the combination (p remains significant or not), but need to be considered when analysing the data. This will be further discussed in the
Discussion (p.17).
A summary of the significant differences in and between the combinations can be found in Table 12.
Table 8
Properties soils of ants’ nest mounds and control sites in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing 0-5 cm
Property Site p
Control site 0-5cm Nest mound 0-5cm
pH 7.51 ± 0.25 7.20 ± 0.25 0.0059 Conductivity (µSv cm-1) 346 ± 156 510 ± 174 0.0156 Organic C (%) 1.54 ± 0.79 2.41 ± 1.64 0.1016 N (%) 0.22 ± 0.28 0.21 ± 0.13 0.4375 Without outliers 0.22 ± 0.28 Without outliers 0.14 ± 0.11 0.0625 K (mg kg-1) 25.79 ± 23.6 37.44 ± 20.7 0.0830 Without outliers 20.06 ± 14.79 Without outliers 37.4 ± 20.7 0.0137 Na (mg kg-1) 1.15 ± 0.37 1.56 ± 0.76 0.0098 Without outliers 1.06 ± 0.20 Without outliers 1.25 ± 0.34 0.0391 Mg (mg kg-1) 11.95 ± 5.46 18.63 ± 7.67 0.0137 Without outliers 10.57 ± 3.16 Without outliers 18.63 ± 7.67 0.0139 Ca (mg kg-1) 128.94 ± 64.48 178.12 ± 53.53 0.0537
Nest mounds had a significantly lower pH than control soil, and electrical conductivity and concentrations of Na and Mg were significantly greater in soil from nest mounds. When removing the outliers, K can become
significantly higher at the nest mound.
Table 9
Properties of soils of ants’ nest mounds and control sites in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing 5-10 cm
Property Site p Control site 5-10cm Nest mound 5-10cm
pH 7.48 ± 0.21 7.26 ± 0.18 0.0156 7.48 ± 0.21 7.21 ± 0.10 0.0313 Conductivity (mSv cm-1) 336 ± 88 474 ± 119 0.0154 Organic C (%) 1.57 ± 0.77 1.94 ± 0.85 0.1953 N (%) 0.243 ± 0.312 0.156 ± 0.056 0.9145 Without outliers: 0.1467 ± 0.081 Without outliers 0.156 ± 0.056 0.3750 K (mg kg-1) 25.29 ± 21.54 30.01 ± 16.40 0.250 Without outliers: 15.2331 ± 4.23 Without outliers: 30.01 ± 16.40 0.0469 Na (mg kg-1) 1.092 ± 0.423 1.085 ± 0.201 0.6406 Without outliers: 0.968 ± 0.165 Without outliers:1.019 ± 0.078 0.2188 Mg (mg kg-1) 12.80 ± 3.22 21.28 ± 8.85 0.0156 Ca (mg kg-1) 130.38 ± 36.78 174.96 ± 55.22 0.1094
Table 9 shows that in the uppermost layer, the parameters pH, conductivity and Mg are significantly different
between the nest and control site. When removing outliers, K can become significantly higher at the nest mound.
Table 10
Properties of soils of ants’ nest mounds in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing crust (0-5 cm) and sub crust layer (5-10 cm)
Property Site p
Nest mound 0-5 cm Nest mound 5-10 cm
pH 7.2 ± 0.25 7.3 ± 0.18 0.25 Conductivity (mSv cm-1) 510 ± 174 473 ± 119 0.99 Without outliers: 510 ± 174 Without outliers:440 ±77 0.68 Organic C (%) 2.4 ± 1.64 1.9 ± 0.83 0.3828 Without outliers: 2.0 ± 1.04 Without outliers: 1.9 ± 0.83 0.3828 N (%) 0.16 ± 0.13 0.21 ± 0.06 0.1719 K (mg kg-1) 37.4 ± 20.65 30.0 ± 16.40 0.5469 Na (mg kg-1) 1.56 ± 0.76 1.09 ± 0.21 0.3828 Without outliers: 1.25 ± 0.34 Without outliers: 1.02 ± 0.08 0.5781 Mg (mg kg-1) 18.6 ± 7.67 21.3 ± 8.85 0.6406 Ca (mg kg-1) 178.12 ± 53.54 175.00 ± 55.21 0.8438
When comparing the values of the upper and the uppermost layer of the nest locations, no significant values are detected. A remarkable trend is that concentrations of nutrients differ in depth, as K, Na and Ca are higher in the 0-5 cm layer compared to 0-5-10 cm, whereas Mg is lower in the 0-0-5 cm layer. However, differences are minor and not significant.
Table 11
Properties soils of ants’ nest mounds and control sites in the catchment area (mean ± SE) and significance of comparison between sites by a Wilcoxon signed rank test (p), comparing 0-10 cm
Property Site p Control 0-10 cm Nest mound 0-10 cm
pH 7.53 ± 0.23 7.26 ± 0.22 4.52 x 10-4 Organic C (%) 1.6 ± 0.76 2.2 ± 1.35 0.0364 Without outlier 1.6 ± 0.76 Without outlier 2.0 ± 0.93 0.0642 Conductivity (µSv cm-1) 341 ± 125 494 ± 151 9.94e-04 Without outliers 341 ± 125 Without outliers473 ± 121 0.0016 N (%) 0.14 ± 0.09 0.19 ± 0.11 0.0039 Without outliers 0.14 ± 0.09 Without outliers 0.17 ± 0.08 0.0078 K (mg kg-1) 25.61 ± 22 34.3 ± 19 0.0401 Without outliers 14.64 Without outliers 34.31 6.1035 x 10 -4 Na (mg kg-1) 1.12 ± 0.39 1.36 ± 0.63 0.0176 Without outliers 1.01 ± 0.19 Without outliers 1.17 ± 0.29 0.0131 Mg (mg kg-1) 12.36 ± 4.4 19.74 ± 8.1 8.3748 x 10 -4 Without outliers 11.69 ± 3.3 Without outliers 19.74 ± 8.1 8.6348 x 10 -4 Ca (mg kg-1) 129.62 ± 52 176.81 ± 53 0.010 Moisture content (%) 0.44 ± 0.19 1.46 ± 2.83 0.0141 Without outliers 0.44 ± 0.19 Without outliers0.53 ± 0.19 0.0442
Nest mounds had a significantly lower pH than bulk soil, whereas conductivity, moisture content, organic carbon and nitrogen, and the concentrations of K, Na, Mg and Ca were greater in soil from nest mounds. When outliers are removed at the organic carbon parameter, no significant differences are observed between nest and control site.
3.2.1 Summary of lab results
In Table 12 a schematic overview of the results can be found. Within this table, the first column shows the
comparisons which has been made, and the second column shows the parameters which are significantly influenced by ant activity. The – gives a negative influence from the nest site compared to the control site, whereas the + gives a positive effect of ant activity on this parameter. For example, when comparing the pH of the nest location (0-5 cm) and control location (5-10 cm), the pH was significantly lower at the nest mounds than at the control site. The last column shows the number of parameters influenced as part of the total. The combination of nest location (0-10 cm) and control site (0-10 cm) has one parameter more, which is the moisture content.
Table 12: Summary of lab results with significant differences.
Comparison Parameters influenced by ants (+/- gives the
influence of ants, either positive (+) or negative (-) x/x influenced as part of total parameters Nest 0-5 cm – Control 0-5 cm pH (-), EC (+), Na (+), Mg (+) 4/8 Nest 5-10 cm – Control 5-10 cm pH (-), EC (+), Mg (+) 3/8
Nest 0-5 cm – Nest 5-10 cm No significant differences 0/8
Nest site 0-10cm– Control site 0-10cm pH (-)
EC (+), organic carbon (+), N (+), K (+), Na (+), Mg (+), Ca (+), moisture content (+)
4. Discussion
4.1 Lab analysisAt the site, the nest mounds of the Messor bouvieri and Messor barbarus had higher concentrations of macronutrients and organic carbon than in the surrounding soil, when comparing 0-10 cm of both sites.
Among others, the higher percentage of nitrogen, the lower pH and the accumulation of nutrients are likely to be an effect of an increased amount of organic carbon at the nest site (+0.6% at the nest site, 0-10 cm, Table 12). The respiration and decomposition of organic matter is related to the lower pH at the nest sites, which is visible in this study. The pH is 7.26 at the nest site, and significantly higher at the control sites ( Table 12). For alkaline soils, it is known that the pH is reduced due to the decomposition of the organic matter, whereby organic acids are produced and thereby decreasing the pH. Previous ant studies support this (Cammeraat et al., 2002; Boulton et al., 2003), and for example Wagner, Jones and Gordon (2004) who investigated that the pH was 7.1 at control sites, and 6.3 at Pogonomyrmex barbatus nests. The previously mentioned research of Cammeraat et al. (2002) on the Messor
bouvieri confirmed the difference in pH, with 8.1 at control sites, and 7.4 at nest mounds. Overall, the pH can be
considered as low in the area, in line with previous research of Lambregts (no date) in the same study area. A lower pH affects the amount of nutrients and chemicals that are soluble in soil water, therefore the amount of nutrients available to plants.
Outliers need to be considered. The organic carbon is not significantly different between nest and control site when the outliers are removed (see Table 11). However, the value of the outlier (6.35% at nest location, see
Appendix 2B-iii) seems a reasonable value. Previous research from Cammeraat et al. (2002) on the Messor bouvieri
showed that organic carbon can be two to eight times higher at nest mounds than in the surrounding soil.
Nutrient concentrations
Harvester ants collect organic matter from surrounding soils and concentrate this at the nest, resulting in a higher accumulation of nutrients at nests (Wagner, Jones & Gordon, 2004), which is also visible in this research (Table
12). The burrowing activities result in a mixing of (top)soil, thereby influencing the nutrient dynamics in the area.
It is generally known that higher concentrations of P can be found within ant nests (Kilpeläinen et al., 2007) also concluded in their research. The ant nests serve as a C, N and P pool and ants increase the spatial heterogeneity in C and nutrient distribution at the ecosystem level (Kilpeläinen et al., 2007).
The concentrations of P in the research area were too low to measure, which is remarkable as in previous research of Cammeraat et al. (2002) in the same region, almost significant differences (p = 0.006) were found between the concentration of phosphorus between ant nests and a control site, with an accumulation of this element at nest locations. Cammeraat & Risch (2008) state that in Ca-rich soils as in the research area, total P is limiting for vegetation. The lower pH at nest sites makes it difficult for plants to obtain phosphorus from substrates (Louw & Seely, 1982). If ants influence this total P, further research may be of importance to analyse the phosphorus cycle within this area.
The concentrations of Na, K, Ca and Mg (mg kg-1) show variable results, however, the overall conclusion
is that all four nutrients had a significant higher concentration (0-10 cm) at nest mounds, which is in line with previous research from Cammeraat et al. (2002, K and Mg concentrations). However, the concentration of both the nest and control sites are relatively low in this research compared to outcomes in the scientific paper, which study was conducted in November 1997 and 1998 (Table 13).
Table 13: Comparison outcomes between this research and Cammeraat et al. (2002)
Cammeraat et al. (2002) This study
K (µg g-1) Control: 92 Nest site: 227 Control: 26 Nest site: 34 Mg (µg g-1) Control: 68 Nest site: 156 Control: 12 Nest site: 20
An explanation for this difference can be the different research periods. The harvester ants collect the seeds from the Stipa tenacissima and the harvest/flowering season was started in the research period, but the stipa was not fully in bloom and ant nests were still in the beginning of development. It could be possible that less seeds are collected in this research, influencing the organic carbon. The percentage organic carbon is lower in this research compared
to Cammeraat et al. (2002), which calculated a percentage of 2.7% at nest mounds, and 2.2% is determined in this research. This lower percentage in organic carbon could be an explanation for the lower nutrient concentrations.
Conductivity and water repellency
Conductivity increases with increased soil moisture content (Marshall & Holmes, 1988). Moisture content is higher at nest mounds, and the increased conductivity is in line with previous research (Cammeraat et al., 2002; Lambregts, no date). Two outliers in moisture content are removed (9.6531 and 9.2897% at nest site, see Appendix
2E), as these values were considered as high relative to other values (between 0-2% moisture content). The higher
moisture content is mostly related to the increased infiltration and evaporation rates, due to the pores, burrows and channels in mounds (Cammeraat & Risch, 2008). As previously mentioned, infiltration is a highly important process in arid areas and the influence of ants is therefore important for the ecosystem. Differences in soil water repellency could not be measured, as each soil sample was determined as non-water repellent (all below the although arbitrary wetting depth penetrating threshold of 5 s, Dekker et al., 2009), and minor differences were found. This is contrary to what has been found in literature. It is known that water repellency of soils is positively related to organic matter and it is expected that the microbial activity could result in soil which is more water repellent at ant nests than in control areas (Harper & Gilkes, 1994; Hallet & Young, 1999). The organic matter can act as hydrophobic particles, thereby influencing the hydraulic behaviour of soil (Franco et al., 1995). The non-water repellency of the samples can be partly explained by the lithology in the area, as the lime in the area absorbs water and could thereby influence the soil wettability within the region.
Differences in parameters with increasing depth
Not all significant differences as described above are observed when comparing the nest site with the control site at 0-5 cm and 5-10 cm (combination 1 and 2 in Figure 8). Only pH, conductivity and Mg were significantly different at both combinations. This suggests that the affected soil properties are less influenced by ants when depth is taken into account and different depths of nests sites and control sites are compared. It should however be noted that the sample size is smaller in combination 1 and 2, compared to combination 4.
Only Na was significantly higher at the upper layer (0-5 cm, Table 8) compared to the paired control site, whereas this difference did not occur at the uppermost layer (5-10 cm, Table 9). Remarkably, the values of Na were higher at the control site than at the nest mound (5-10 cm), however this difference is minor and non-significant (0.007 mg kg-1 difference).
Potassium (K) can become significant at both combinations when the outliers are removed from the data set. The values of the outliers seem reasonable values. At 0-5 cm the value of the outlier (Control 0-5 cm, 83,12 mg kg-1, see Appendix 2C) seems a reasonable value, as this value falls between the ranges of previous studies
(Cammeraat et al., 2002) and a high(er) value is also noticeable at the nest site (80.91 mg kg-1). At 5-10 cm, the
values of the two outliers (64,49 and 66,58 mg kg-1, see Appendix 2C) have a value close to each other, and seem a
reasonable value when compared to the nest location, which has a value of 61,32 mg kg -1. Potassium is a highly
important nutrient in dry areas, as it is known that this nutrient reduces stress from drought by stimulating the water uptake at the roots by plants (Armstrong et al., 1998). Values are higher (although not significant at 0-5 and 5-10cm) at nest sites, thereby indicating a positive influence of ants on the concentration of this important nutrient in ecosystems.
Besides this, when comparing the depths of the same nests (0-5 cm and 5-10 cm of nest mound, combination 3), no significant differences were found, suggesting that the influences do not in- or decrease with depth. It should be said that trends are visible, as the values of the parameters pH, conductivity, organic carbon, K, Na and Ca decrease with depth, whereas N and Mg increase with depth. This is contrary to the hypotheses, which was that the influence increases with depth, as the seeds are taken deep into the nests and ant activity is higher with increasing depth. The results confirm the role of ants in influencing soil properties, and therefore accepting the hypothesis. The effect is however highly dependent on the ecosystem and ant specie (Cammeraat & Risch, 2008; Frouz & Jilkova, 2008). The influences do not significantly increase with depth therefore rejecting the hypothesis and trends are visible which support the rejection of this hypothesis.
4.2 Spatial analysis
Spatial patterns of changed soil physical, chemical and hydrological properties as described above, are important at the nest scale, but they also have important consequences for hydrological, geomorphological and hydrological processes at catchment scales (Cammeraat & Risch, 2008).
Within this research area, the nest density is approximately 3.54 nests ha-1, which is an extremely sparse
pattern compared to previous studies. Possible ant nest could have been missed in the area as nests were sometimes hard to notice due to the missing characteristic ant highways, despite examination in vertical lanes. Nest density of the Lasius niger in the Czech Republic showed a dense pattern with 165 nests on 2500 m2 (Holec et al., 2006).
Another study at the Messor barbarus (NE Spain) reported an ‘normal average nest density in the area’ of 468 nests ha-1 (Baraibar et al., 2011 via Blanco-Moreno et al., 2014). An explanation for this lower density could be the
different research periods, as both researches took place in late Summer, namely July and August (Blanco-Moreno et al., 2014) and June (Holec et al., 2006). It is possible that less nests are developed earlier in the harvest season of Stipa, which was the case in this research. A decrease in seed availability has led to a reduced ant nest density in the previous study of Blanco-Moreno et al. (2014). A low ant nest density can increase the diversity and density of particularly herbivores and other decomposers and in this way influencing the ecosystem in the environment (Sanders & Van Veen, 2011).
Furthermore, there was a difference in nest density between the different lithology and soil units. There is a strong relationship between lithology and the occurrence of soils (Gray, Bishop & Wilford, 2016). In that way, it is a logical consequence that the highest nest densities occur at both the calcrete and the calcisol. It is expected that the calcrete layer could negatively influence the nest development in depth, as the nests of the Messor barbarus are generally deep and the calcrete layer in the research area is found at an average of 10.7 cm depth (see Bachelor
Thesis of Olaf de Haan, and generally ≤ 100 cm from surface, WRB, 2014). However, the highest nest densities
occur on these units. It is hard to determine why ants seem to prefer these units, as from previous studies, it becomes clear that spatial trends in ant nest locations do not solely depend on soil type and altitude, but should be sought more in dynamic factors as seed availability, microclimate, soil surface conditions and intraspecific competition (Blanco-Moreno et al., 2014). It could be the case that a higher vegetation density was on the calcisols and calcrete and therefore more available seeds, however, that goes beyond this study. The BSc Thesis of Bart ter
Mull focusses more on the vegetation. It is therefore difficult to give an explanation which is solely based on the
lithology and soil.
The pattern of the nests is analysed as clustered. Most studies at spatial patterns of ant nests showed a regularly distributed pattern or were randomly distributed with tending towards regularity (Levings & Traniello, 1981, summarizing 160 studies involving 136 ant species). The pattern of Messor barbarus in NE Spain also showed this pattern (Blanco-Moreno et al., 2014). Holec & Frouz (2006) stated that the spatial distribution differs with nest size and the size of the sample plots. A larger sampling plot results in more aggregated distribution than smaller plots (Holec & Frouz, 2006). The plot in the research area is of relatively large compared to other plots as in the study of Holec et al. (2006), (2500 m2, random pattern) and Blanco-Moreno et al., 2014 (random, 3750 m2).
A clustered pattern reinforces the development of fertile islands in arid lands. In arid lands, the effect of changing soil properties as described in Chapter 4.1 can be larger. This can be due to the fact that fertility is intrinsically lower in arid lands (Farji-Brener & Werenkraut, 2017). Arid landscapes are often characterized by patchy vegetation. Ants take seeds from the existing densely covered patches to nutrient-poor areas, thereby influencing the soil fertility at these locations (Pugnaire et al., 1996). Within this assumption, spatial distribution plays an important role as it for example can be said that clustering reinforces this development, whereas a dispersed pattern would result in the development of small(er) fertile islands. However, it should be said that low nest densities were calculated in this study, and therefore, the process of the formation of these fertile islands in this study might not be as great as in other areas with higher nest densities.
Conclusion
Within this study, it is tried to give an insight in the role of ants (Messor barbarus and Messor bouvieri) as ecosystem engineers. After analysing soil properties and the spatial distribution, this study confirms that can be seen as ecosystem engineers in the semi-arid area.
Concluded from the lab analysis, it becomes clear that ants influence chemical and physical soil properties within the catchment. When comparing the 0-10 cm, all nine measured parameters (pH, EC, organic carbon, nutrient concentrations and moisture content) are significantly influenced by ant activity, thereby confirming the role of ants as soil engineers. This has for example consequences for nutrient availability for plants and fertility of the soil within the catchment. However, this difference is not clearly visible when comparing different depths, because no differences are noticed when comparing 0-5 cm and 5-10 from the nest site. Concluded from the spatial analysis, nest activity was relatively low in the area. This can result in the increase the diversity and density of herbivores and other decomposers as earthworms and in this way influencing the ecosystem in the environment. Nest density was the highest on the calcisol and calcrete zone, and the lowest on the leptosol and the marl. Spatial preferences and reasons are difficult to explain, as it becomes clear that not only lithology and soil units are relevant. An integrated framework with other influences as seed availability, microclimate, intraspecific competition and topography could be made, which is a suggestion for further research. The pattern of the nests can be considered as clustered in the whole study area and on each lithology and soil parameter as well. The clustering can reinforce the process of the development of fertile islands in the area, although, due to the low nest density it is expected that the clustering does not affect this process significantly in the catchment area.
Regarding all outcomes, it should be considered that nests were not fully developed in the research period, and that this could have influenced outcomes within this research. The same study can be conducted later in the year and within the flowering season, to analyse if this might have influenced the results.
References
Armstrong, D.I. et al. (1998). Potassium for agriculture. Better Crops, 3 (82), 34-36.
Azcárate, M. F., & Peco, B. (2007). Harvester ants (Messor barbarus) as disturbance agents in Mediterranean grasslands. Journal of Vegetation Science, 18(1), 103-110.
Boulton, A. M., Jaffee, B. A., & Scow, K. M. (2003). Effects of a common harvester ant (Messor andrei) on richness and abundance of soil biota. Applied Soil Ecology, 23(3), 257-265.
Blanco-Moreno, J. M., Westerman, P. R., Atanackovic, V., & Torra, J. (2014). The spatial distribution of nests of the harvester ant Messor barbarus in dryland cereals. Insectes sociaux, 61(2), 145-152.
Cammeraat, L.H. (2017). Fieldwork manual, Southeast Spain. IIS, University of Amsterdam.
Cammeraat, L.H. (2002). A review of two strongly contrasting geomorphological systems within the context of scale. Agric., Ecosyst. Environ., 104, 317-332.
Cammeraat, E. L. H., & Risch, A. C. (2008). The impact of ants on mineral soil properties and processes at different spatial scales. Journal of Applied Entomology, 132(4), 285-294.
Cammeraat, L. H., Willott, S. J., Compton, S. G., & Incoll, L. D. (2002). The effects of ants' nests on the physical, chemical and hydrological properties of a rangeland soil in semi-arid Spain. Geoderma, 105(1), 1-20.
Cerdá, X., & Retana, J. (1994). Food exploitation patterns of two sympatric seed‐harvesting ants Messor bouvieri (Bond.) and Messor capitatus (Latr.)(Hym., Formicidae) from Spain. Journal of Applied Entomology, 117(1‐5), 268-277.
Crist, T. O., & Wiens, J. A. (1996). The distribution of ant colonies in a semiarid landscape: implications for community and ecosystem processes. Oikos, 301-311.
Dekker, L. W., Ritsema, C. J., Oostindie, K., Moore, D., & Wesseling, J. G. (2009). Methods for determining soil water repellency on field‐moist samples. Water resources research, 45(4).
Folgarait, P. J. (1998). Ant biodiversity and its relationship to ecosystem functioning: a review. Biodiversity & Conservation, 7(9), 1221-1244.
Farji‐Brener, A. G., & Werenkraut, V. (2017). The effects of ant nests on soil fertility and plant performance: a meta‐ analysis. Journal of Animal Ecology, 86(4), 866-877.
Franco, C.M.M., Tate, M.E., Oades, J.M., 1995. Studies on non-wetting sands: I. The role of intrinsic particulate organic matter in the development of water-repellency in non-wetting sands. Aust. J. Soil Res. 33, 253–263.
Frouz, J., & Jilková, V. (2008). The effect of ants on soil properties and processes (Hymenoptera: Formicidae). Myrmecological News, 11(11), 191-199.
Frouz, J., Holec, M., & Kalčík, J. (2003). The effect of Lasius niger (Hymenoptera, Formicidae) ant nest on selected soil chemical properties. Pedobiologia, 47(3), 205-212.
Geerlings, I.P.A. (1978). Geologie van het gebied ten noordoosten van Fuensanta La Parroquia, Murcia, Zuid-Oost Spanje. Department of Phys. Geogr. And Soil Sc., University of Amsterdam: 1-72.
Gray, J. M., Bishop, T. F., & Wilford, J. R. (2016). Lithology and soil relationships for soil modelling and mapping. Catena, 147, 429-440.
Hallett, P.D., & Young, I.M. (1999). Changes to water repellence of soil aggregates caused by substrate-induced microbial activity. Eur. J. Soil Sci. 50, 35–40.
Harper, R.J., Gilkes, R.J., (1994). Soil attributes related to water repellency and the utility of soil survey for predicting its occurrence. Aust. J. Soil Res. 32, 1109–1124
Hole, F. D. (1981). Effects of animals on soil. Geoderma, 25(1-2), 75-112.
Holec, M., & Frouz, J. (2006). The effect of two ant species Lasius niger and Lasius flavus on soil properties in two contrasting habitats. European Journal of Soil Biology, 42, S213-S217.
Holec, M., Frouz, J., & Pokorný, R. (2006). The influence of different vegetation patches on the spatial distribution of nests and the epigeic activity of ants (Lasius niger) on a spoil dump after brown coal mining (Czech Republic). European journal of soil
biology, 42(3), 158-165.
Imeson, A.C. & Cammeraat, L.H. (1999). Scaling up from field measurements to large areas using the Desertification Response Unit and Indicator Approaches. Rangeland Desertification. O. Arnalds and S. Archer. Dordrecht, Kluwer Academic Publishers: 99-114.
Kilpeläinen, J., Finér, L., Niemelä, P., Domisch, T., Neuvonen, S., Ohashi, M., ... & Sundström, L. (2007). Carbon, nitrogen and phosphorus dynamics of ant mounds (Formica rufa group) in managed boreal forests of different successional stages. Applied
Soil Ecology, 36(2), 156-163.
Lavelle, P. (1996). Diversity of soil fauna and ecosystem function. Biol Int, 33, 3-16.
Lambregts, C. (no date). MSc Thesis: The impact of harvesting ants on soil surface properties of a degraded Mediterranean Ecosystem. University of Amsterdam.
Levings, S. C., & Traniello, J. F. (1981). Territoriality, nest dispersion, and community structure in ants. Psyche, 88(3-4), 265-319.
Lobry de Bruyn, L.A. & Conacher, A.J., (1990). The role of termites and ants in soil modification: a review. Aust. J. Soil Res. 28, 55–93
Louw, G.N., & Seely, M.K., (1982). Ecology of Desert Organisms. Longman, London.
Navarra Hervas, F. (1991). El Sistema Hidrografico del Gualalentin, Consejeria de Politica territorial, Murcia.
Whitford, W. G., & DiMarco, R. (1995). Variability in soils and vegetation associated with harvester ant (Pogonomyrmex rugosus) nests on a Chihuahuan Desert watershed. Biology and Fertility of Soils, 20(3), 169-173.
NEN, 1993. Bepaling van het gehalte aan carbonaten in grond, Gravimetrische methode. Uitgave Nederlands Normalisatie-Instituut.
MacMahon, J. A., Mull, J. F., & Crist, T. O. (2000). Harvester ants (Pogonomyrmex spp.): their community and ecosystem influences. Annual Review of Ecology and Systematics, 31(1), 265-291.
Marshall, T.J. & Holmes, J.W., 1988. Soil Physics, 2nd edn. Cambridge Univ. Press, Cambridge.
Mierenboerderij (2017). Messor barbarus. Retrieved on 11 June 2017 from https://www.mierenboerderij.nl/messor-barbarus-kolonie-koningin-en-1-4-werksters.html
Pugnaire, F.I., Haase, P., Puigdefabregas, J., Cueto, M., Clark, S.C., Incoll, L.D., (1996). Facilitation and succession under the canopy of a leguminous shrub, Retama sphaerocarpa, in a semiarid environment in Southeast Spain. Oikos 76, 455–464. Sanders, D., & van Veen, F. J. (2011). Ecosystem engineering and predation: the multi‐trophic impact of two ant species. Journal of Animal Ecology, 80(3), 569-576.
Scoging, H., (1989). Runoff generation and sediment mobilisation by water. In: Thomas, D.S.G. Ž . Ed. , Arid Zone Geomorphology. Belhaven Press, London, pp. 87–116.
Six et al. (2004). A history of research on the link between (micro) aggregates, soil biota, and soil organic matter dynamics. Soil Tillage Res., 7-31.
Uppstrom, K. A., & Klompen, H. (2011). Mites (Acari) associated with the desert seed harvester ant, Messor pergandei (Mayr). Psyche: A Journal of Entomology, 2011.
Wagner, D., Jones, J. B., & Gordon, D. M. (2004). Development of harvester ant colonies alters soil chemistry. Soil Biology and Biochemistry, 36(5), 797-804.
Warburg, I., & Steinberger, Y. (1997). On the spatial distribution of nests of the ants Messor arenarius and Messor ebeninus. Journal of arid environments, 36(4), 671-676.
Weijermans, R. (1991). Geology and tectonics of the Betic Zone, SE Spain. Earth-Sci. Rev. 31, 153-236.
Wesemael, J.Ch. van (1955). De bepaling van het calciumcarbonaatgehalte van gronden, Chem. Weekbl., 51, 35-36. Whitford, W. G., & DiMarco, R. (1995). Variability in soils and vegetation associated with harvester ant (Pogonomyrmex rugosus) nests on a Chihuahuan Desert watershed. Biology and Fertility of Soils, 20(3), 169-173.
World Reference Base for Soil Resources (2014). International soil classification system. World Soil Resources Report 106. FAO, Rome, 181 pp.
Acknowledgements
First and foremost, I want to thank my research supervisor Mr. Erik Cammeraat. I would like to thank you very much for your support, ideas and useful comments on and in this research. Your passion and knowledge about the study area and the subject has helped me a lot in this research.
I would also like to show gratitude to Ms Lieke Mulder, who has helped me with (statistically) analysing my data and who has given me some valuable remarks on my research. This research was not established without Mr Rutger van Hall, who has shown me around in the laboratories of the IBED, analysed and prepared my end results and who has always picked up the phone, even when I called several times a day.
Besides this, I want to thank Mr Jinze Noordijk (Naturalis) for the determination of my ant species and Mr Rudolf van Hengel (Mierenwerkgroep Nederland), who has send me useful literature on ants and ant species.
8. Appendices
Table of content Appendix
1. Detailed map of fieldwork area and ant nest locations 2. Results chemical and physical soil analysis
A. Map of sample locations B. Carbon analysis
i. CNS analysis ii. Inorganic carbon iii. Organic carbon C. ICP analysis
D. pH and conductivity E. Moisture content F. TOC analysis G. N analysis
3. Spatial distribution analysis A. Detailed lithology map B. Detailed soil map C. Simplified map lithology D. Simplified map soil
E. MatLab code to calculate nearest neighbour ratio F. List of ant nest locations including coordinates
G. List of ant nests locations separated in lithology and soil unit H. Nearest neighbour ratio of largest contiguous area
2. Chemical and physical soil analysis A. Map of sample locations
B. Results carbon analysis a. CNS Analysis
Location What Depth Sample 1,
C (%) Sample 2, C (%) C (%), mean of two samples 1 Nest 0-5cm 10,61 10,62 10,615 1 Nest 5-10cm 10,96 10,94 10,95 1 Control 0-5cm 10,02 10,01 10,015 1 Control 5-10cm 10,63 10,64 10,635 2 Nest 0-5cm 10,84 10,85 10,845 2 Nest 5-10cm 10,59 10,59 10,59 2 Control 0-5cm 9,69 9,35 9,52 2 Control 5-10cm 10,13 10,15 10,14 3 Nest 0-5cm 12,01 12,01 12,01 3 Nest 5-10cm 11,96 11,97 11,965 3 Control 0-5cm 11,14 11,07 11,105 3 Control 5-10cm 11,63 11,66 11,645 4 Nest 0-5cm 12,16 12,05 12,105 4 Control 0-5cm 9,73 9,74 9,735 4 Control 5-10cm 9,38 9,38 9,38 5 Nest 0-5cm 10,07 10,08 10,075 5 Control 0-5cm 10,55 10,54 10,545 5 Control 5-10cm 10,68 10,73 10,705 6 Nest 0-5cm 10,81 10,84 10,825 6 Nest 5-10cm 10,08 10,09 10,085 6 Control 0-5cm 9,06 9,08 9,07 6 Control 5-10cm 8,52 8,56 8,54 7 Nest 0-5cm 9,7 9,7 9,7 7 Control 0-5cm 10,55 10,55 10,55 8 Nest 0-5cm 8,51 8,52 8,515 8 Nest 5-10cm 8,56 8,56 8,56 8 Control 0-5cm 9,02 9,02 9,02 8 Control 5-10cm 8,75 8,76 8,755 9 Nest 0-5cm 11,52 11,51 11,515 9 Nest 5-10cm 11,34 11,39 11,365 9 Control 0-5cm 10,91 10,91 10,91 9 Control 5-10cm 10,8 10,82 10,81 10 Nest 0-5cm 12,21 12,22 12,215 10 Nest 5-10cm 12,22 12,34 12,28 10 Control 0-5cm 11,05 11,04 11,045 10 Control 5-10cm 11,35 11,33 11,34 11 Nest 0-5cm 11,92 11,93 11,925 11 Nest 5-10cm 11,6 11,63 11,615 11 Control 0-5cm 11,11 10,98 11,045 11 Control 5-10cm 11,35 11,35 11,35
i. Inorganic carbon (Van Wesemael methode)
Referring to the colours in the Location column: Red: round 1 Blue: round 2 Beige: round 3
Locatio
n What Depth Weight of
sand (gram) Left over (gram) Weigh t of sample (air dried) (gram) Moisture content (%) Weight sample (oven dried) (gram) Erlenmeye r with sample (gram) Erlenmeye r after 24 hours at shaking plate (gram) Weight loss Erlenmeye r after shaking plate (gram) CO2 (%) Inorganic carbon (%):calculated by the percentage C in CO2: (12,0107/44,01 ) Calcium Carbonate (CaCO3): calculated by %CO2 x 100/44 1 Nest 0-5cm 2,1986 0 2,1986 0,86124401 9 2,17966468 9 156,9311 156,2785 -0,6526 31,4227263 4 8,575526908 71,4152871 3 1 Nest 5-10cm 2,1964 0,0006 2,1958 0,84219481 1 2,17730708 6 153,8672 153,2152 -0,652 31,4278297 6 8,576919675 71,4268858 2 1 Control 0-5cm 2,0996 0,0009 2,0987 0,746129453 2,083040981 155,6074 154,9825 -0,6249 31,48467341 8,592432786 71,55607593 1 Control 5-10cm 1,9013 0,0006 1,9007 0,272952854 1,895511985 152,9443 152,4005 -0,5438 30,10919544 8,217053254 68,42998964 2 Nest 0-5cm 2,1581 0,0018 2,1563 0,528494671 2,144904069 155,1949 154,592 -0,6029 29,5001275 8,050833477 67,04574431 2 Nest 5-10cm 2,0084 0,0026 2,0058 0,323465987 1,999311919 154,9201 154,3555 -0,5646 29,63785718 8,08842107 67,35876633 2 Control 0-5cm 2,1771 0,004 2,1731 0,252131999 2,16762092 145,3108 144,6682 -0,6426 31,11314191 8,491038707 70,71168615 2 Control 5-10cm 2,0039 0,0023 2,0016 0,341530055 1,994763934 155,7812 155,2001 -0,5811 30,57354867 8,343779164 69,48533789 3 Nest 0-5cm 1,8927 0,0005 1,8922 0,401685118 1,884599314 155,7106 155,0814 -0,6292 35,55983387 9,704578429 80,81780425 3 Nest 5-10cm 2,0526 0,0002 2,0524 0,48951049 2,04235328 7 155,5017 154,7728 -0,7289 38,0125601 10,37394809 86,3921820 4 3 Control 0-5cm 2,2145 0,0015 2,213 0,34249833 5 2,20542051 2 148,5968 147,8965 -0,7003 33,3256937 6 9,094862759 75,7402130 9 3 Control 5-10cm 2,1128 0,0012 2,1116 0,50533408 2 2,10092936 6 156,457 155,6987 -0,7583 37,8805335 10,33791692 86,0921215 8 4 Nest 0-5cm 1,726 0,0001 1,7259 0,760558339 1,712773524 155,0721 154,7332 -0,3389 21,07471286 5,751466796 47,89707467 4 Control 0-5cm 1,9601 0,0006 1,9595 0,31131433 1,95339979 154,6648 154,195 -0,4698 25,6160352 6,990831967 58,2182619
6 8 9 4 Control 5-10cm 2,0138 0,0002 2,0136 0,440859206 2,004722859 154,0284 153,5845 -0,4439 23,58418325 6,436322422 53,60041648 5 Nest 0-5cm 1,9966 0,002 1,9946 0,159221583 1,991424166 156,0423 155,3821 -0,6602 35,31033071 9,636486913 80,25075161 5 Control 0-5cm 1,8682 0,0019 1,8663 0,281786355 1,861041021 150,5333 149,8929 -0,6404 36,65096513 10,00235735 83,29764802 5 Control 5-10cm 2,1131 0,0017 2,1114 0,30463220 1 2,10496799 6 151,1845 150,4721 -0,7124 36,046944 9,837514889 81,9248727 2 6 Nest 0-5cm 1,9834 0,0011 1,9823 0,58462718 1,970710935 158,0864 157,565 -0,5214 28,17981506 7,690509082 64,04503422 6 Nest 5-10cm 1,9534 0,0013 1,9521 0,3894081 1,944498364 156,6746 156,1471 -0,5275 28,89381714 7,885366269 65,66776622 6 Control 0-5cm 2,0224 0,003 2,0194 0,275286387 2,013840867 156,7129 156,1202 -0,5927 31,34727658 8,554936032 71,2438104 6 Control 5-10cm 2,0157 0,002 2,0137 0,35441149 2,006563216 153,7141 153,1575 -0,5566 29,54475505 8,063012712 67,14717056 7 Nest 0-5cm 2,1162 0,001 2,1152 9,653121903 1,911017166 149,8422 149,3606 -0,4816 26,84181806 7,325358423 61,00413196 7 Control 0-5cm 2,0027 0,0004 2,0023 0,994506535 1,982386996 156,8829 156,3562 -0,5267 28,29859778 7,722925889 64,31499496 8 Nest 0-5cm 1,8021 0,0137 1,7884 0,327943868 1,782535052 158,4168 157,9152 -0,5016 29,9715696 8,179494001 68,11720364 8 Nest 5-10cm 1,798 0,0011 1,7969 0,544294294 1,787119576 153,814 153,3096 -0,5044 30,06155932 8,204052953 68,32172572 8 Control 0-5cm 1,9626 0,0017 1,9609 0,346636687 1,954102801 155,7701 155,2218 -0,5483 29,88552412 8,156011465 67,92164573 8 Control 5-10cm 1,9578 0,002 1,9558 0,47331054 4 1,94654299 2 154,2599 153,7221 -0,5378 29,4270573 6 8,030892021 66,8796758 2 9 Nest 0-5cm 2,0263 0,0162 2,0101 0,633977216 1,997356424 149,4752 148,8644 -0,6108 31,95194016 8,719953822 72,61804583 9 Nest 5-10cm 2,3708 0,0229 2,3479 0,659448819 2,332416801 146,6504 145,9532 -0,6972 31,2323738 8,523578097 70,98266774 9 Control 0-5cm 1,7408 0,0125 1,7283 0,810341501 1,714294868 155,1277 154,5785 -0,5492 34,12204424 9,312193519 77,55010055 9 Control 5-10cm 2,0404 0,0163 2,0241 0,475603312 2,014473313 158,6419 158 -0,6419 33,93875411 9,262172098 77,13353207 10 Nest 0-5cm 2,077 0,0103 2,0667 0,488138241 2,056611647 161,4685 160,7995 -0,669 33,98815603 9,275654298 77,24580917
10 Nest 5-10cm 2,1203 0,0245 2,0958 0,440233515 2,086573586 147,1641 146,4485 -0,7156 35,83359683 9,77929065 81,43999281 10 Control 0-5cm 2,2639 0,0166 2,2473 0,372404804 2,238930947 156,8651 156,1159 -0,7492 34,96317355 9,541744799 79,46175807 10 Control 5-10cm 2,1653 0,0157 2,1496 0,35977536 2,14186626 9 155,8145 155,1169 -0,6976 34,0304660 7 9,287201063 77,3419683 4 11 Nest 0-5cm 2,262 0,0243 2,2377 0,50152247 9 2,22647743 1 157,68 156,8296 -0,8504 39,9078867 2 10,8911987 90,6997425 4 11 Nest 5-10cm 2,2202 0,0056 2,2146 9,289671547 2,008870934 152,4749 151,7385 -0,7364 38,30147696 10,45279594 87,04881127 11 Control 0-5cm 2,3112 0,0177 2,2935 0,555500446 2,280759597 154,0688 153,3305 -0,7383 33,82261125 9,230475732 76,86957102 11 Control 5-10cm 2,0079 0,0071 2,0008 0,397200681 1,992852809 155,941 155,2559 -0,6851 35,91968513 9,802784871 81,63564802
References and blanco
Round 1 Weight of
sample (grams)
Left over
(grams) CaCO(grams)3 Weight of Erlenmeyer with sample (grams) Weight after shaking plate (grams) Weight loss (grams) R1 Reference 1: Calcium carbonate 0,2606 0,001 0,2596 151,717 151,6073 0,1097 B1 Blanco: only HCl - - 150,9893 150,981 0,0083 Round 2 R2 Reference 2: Calcium carbonate 0,265 0,0001 0,2649 145,7355 145,623 0,1125 R3 Reference 3: Calcium carbonate 0,2479 0,0001 0,2478 153,6085 153,5092 0,0993
