Ho
w s
oi
l de
ve
lo
p
m
ent relate
s to su
cc
es
sion
- th
e
you
ng
, w
et
d
u
ne
slac
k
s,
sou
th
we
st
Te
xel
2015
Bache
lo
r The
sis
The southern coastline of Texel is one of the few remaining areas in The Netherlands which is marked by nutrient limitation. Therefore it is an unique site, featured by: low anthropogenic influence, high biodiversity and the occurrence of pioneer species, of which many occur on national and international Red Lists. Over the past four decades a steep decline in typical, species-rich dune slacks has been witnessed. Previous studies show that soil development parameters such as; OM accumulation, water table height, C:N ratio and soil pH, are of large influence on the
occurrence of pioneer species. Long-term monitoring of southwest Texel will contribute in establishing the exact interactions between these parameters and the occurrence, especially of the pioneer species Liparis
Loeselii. Therefore this study aimed to test and consolidate findings from
previous studies by having a more integrated approach. In this thesis the research addressed the multiple parameters related to organic matter content by quantifying developments in: depth of the mineral Ah horizon, carbon to nitrogen ratio and soil carbon and nitrogen contents. All parameters showed an increase over dune slack age and site sampling over the years 2010, 2014 and 2015. However this last relationship is less clear for the eldest dune slack H1, which may be explained by environmental and stoichiometric controls. This study should be seen in relation with two synchronously conducted studies, focussing on: acidity, electrical conductivity, bulk density and the presence of vegetation groups. Cohesion of these three theses has enhanced insights into the interactions of pioneer species, management techniques and soil
Universiteit van Amsterdam
Author: Dienke Stomph
Supervisor: Dr. A.W. Kooijman,
Peer reviewer: Dr. G. Oostermeijer
5,946 words excl. Figures and Tables
Submitted: 28
thof June 2015
Keywords:
Succession
Organic matter
Nitrogen
Carbon
C:N ratio
Mineral horizon
Pioneer species
Liparis Loeselii
Thesis – Soil development and succession, Texel
2015
Table of Contents
Introduction ...3
Dune dynamics ... 3
Dune slack development... 3
Early succession stage ... 4
Parameters for succession ... 4
The influence of management ... 4
Research focus ... 5
Methods...5
Composition of dune slacks ... 5
Fieldwork ... 6
Soil laboratory ... 6
Statistical analysis ... 7
Results ...8
Significant developments over time ... 8
Increased depth of the Ah horizon ... 8
Increase in carbon to nitrogen ratio ... 9
Increase in C-content ... 9
Increase in N-content ... 9
Figures 5-7 ...10
Tables 2 and 3 ...16
Predicting the occurrence of the Liparis Loeselii ...18
Discussion ... 19
Fieldwork set up ...19
Interrelations of carbon and nitrogen ...19
Best predictors ...19
Recommendations ...20
Conclusion ... 20
Evaluation and Acknowledgement ... 21
Reference list ... 22
List of Figures ... 24
Thesis – Soil development and succession, Texel
2015
Table with dune slack specifics ...26
Examining the correlation for both percentage and g/m
2for both C- and N-content ...27
Output of examining the best fit linear model with ‘lmfit’ ...28
Results correlation between two parameters of ages 21, 16, 12 and 3 yr ...30
Matlab Script for explorative analysis ...32
Matlab Script for visualizing correlation and regression ... 39
Matlab Script for regression and correlation analysis ...44
Matlab Script analyzing occurrence of the Liparis Loesilii 2010 and 2014 ...51
Matlab Script to compare 2010, 2014 and 2015 ... 52
Thesis – Soil development and succession, Texel
2015
The southern coast of the Wadden Island Texel is a spacious dynamic beach plain with young dunes. The beach plain is one of the European Natura 2000 areas, which aims to protect and preserve this habitat (Rijksoverheid, 2015). The young, wet dune slacks in this area are of special value for they have a high species richness of over 50 species per 4m2, and they are home to many endangered plant species, among which some Red List species (Meijer, Bilius & Vriens, 2015). However in the last 4 decades a decline of species richness has been reported, therefore monitoring of the impact of management activities is necessary. Liparis Loeselii (L.) Rich. is one of these species and functions as a benchmark for the ecosystem condition and dynamics in the area. As a follow-up study to van der Craats (2010), Hollaar (2014) and Jongejans (2014), this study
researches the soil development in the dune slacks, with a focus on: depth of the mineral Ah horizon, C:N ratio, C- and N-content.
Dune dynamics on the southern tip of Texel
The southern tip of the Island was formed by the movement of the two shoals De Hors and Onrust (Figure 1). They were connected to the coast in respectively mid-18th century and early-20th century (Meijer et al., 2015). In the coming decades the ‘Razende Bol’ (also known as Noorderhaaks) is expected to be linked to the south coast of Texel (Figure 2). Over the last years the Dutch government has realized ‘Dynamic Coastal Protection’ through off-shore sand supply in order to promote the movement of the sand shoals. This stimulates progradation and the
formation of small scale embryo dunes as a form of coastal protection (Meijer et al., 2015; van der Spek & Elias, 2006). This enables a more natural and dynamic management of the fore dunes. However in the past interventions to the natural dynamics were executed of which the most significant is the construction of drift dikes in the second half of the 20th century, which created the Kreeftepolder and Horsmeertjes. Another measure which influenced the dune dynamics was the stabilization of the foredunes by planting Ammophila Arenaria (L.) Link in mid-19th century, this enabled the development of primary dune slacks by
cutting areas off from the influence of the sea
Dune slack development
The part of Texel this research focuses on are the wet, young dune slacks, which are characterized by low nutrient availability, large variability in species composition, and a shallow water table during winter and spring. This water table evolves over the years under influence of precipitation and fresh groundwater flows (Figure 3) (Grootjans, 2002). Dune slacks are formed either as primary dune slack, by the formation of new dune ridges which seclude the area at least party from sea water influence. Or as secondary dune slack, which result
from blow-outs that remove the sand layer of the inner part of the dune area till it reaches the water table (van der Craats, 2011; Lammerts, Maas & Grootjans, 2001).
Figure 1: The aerial view on the southern tip of Texel.
Source: Staatsbosbeheer, 2008
Figure 2: Delta development Texel: shifting shoals Onrust & Razende
Thesis – Soil development and succession, Texel
2015
Figure 3: Typical transect with low embryo dunes near the shoreline followed by much taller
mature dunes. Source: David Rayner Beagle Graphics (GeoResources, 2001) Early succession stage
Texel’s young dune slacks feature “pronounced geomorphological and hydrological gradients” (Grootjans et al., 2002) and low nutrient availability, which favors the development of
habitats high in biodiversity of plants and animals. The wet dune slacks undergo a process of succession. The rate of succession influences which species have time to develop. L. Loeselii is a prominent species in classifying the succession stage, for it occurs solely in the early succession, or pioneer stage. This stage is known for its high species richness and therewith biodiversity. Many of these species are listed on the national and international Red Lists such as: L. Loeselii, Pedicularis palustris L., Schoenus nigricans L, Dactylorhiza incarnatee (L.) Soó, Parnassia palustris L.,
Eleocharis quinqueflora (Hartmann) O. Schwarz, Equisetum variegatum Weber and Gentialle amarelle (L.) Börner
(Shahrudin, 2014; Grootjans, Hartog, Fresco & Esselink, 1991).
L. Loeselii is a small-sized orchid, which requires low-vegetation density, open vegetation, a pH above 6 and
relatively fresh calcium-rich groundwater flows, with a shallow groundwater table especially in winter (Grootjans, Stichting, Stuyfzand, Petersen & Shahrudin, 2014; Cederberg & Löfroth 2000; Stanová, Šeffer, & Janák 2008). When the conditions no longer favor the growth of the orchid the existing plants may survive for 3 to 4 more years (Stanová et al., 2008).
Four soil parameters promoting succession
Organic matter (henceforth: OM) is one of the important parameters in soil development. The rate at which OM accumulates depends on both biotic and abiotic factors. Amongst the abiotic factors are: temperature, precipitation, pH, salinity and inorganic nutrients. The biotic factors include: microbes, grazers (eg. rabbits) and plant species, which provide the litter. In general the net OM accumulation in dune slacks over time is positive, resulting in accumulation of humified OM. This enriched layer in the topsoil is the Ah-horzion. It has been established that in general the Ah increases in depth, however accumulation rate largely differs depending on litter quality, soil conditions and litter quantity (Kooijman & Besse, 2002). Carbon is an major component in the mineral Ah-horizon. It is fixed by the process of photosynthesis in the plants, and is subsequently deposited on the topsoil via dead plant material. Soil coverage and vegetation types therefore have a large impact on the C-content of this topsoil.
The initial content of nitrogen in the topsoil seems to be largely dependent on atmospheric deposition (Kooijman & Besse, 2002. However, after establishment of a vegetation cover, the biotic input of nitrogen becomes the most significant. Phosphorus (P) is another important macro nutrient. The availability of P is relatively high, this is caused by a high pH in the early successional stages, which prevents an immobilization of phosphorus in insoluble iron phosphates (Kooijman et al., 1998). Furthermore, continued succession with corresponding acidification of the soil, also leads to an increased OM content. When the OM content of the soil is high and the pH is low, P can be adsorbed, however this is in relatively loosely bound iron-OM complexes (Grootjans et al., 2002).Whilst the dune conditions favor the availability of the macro nutrient phosphorus, the macro nutrient nitrogen is a major limiting factor to the vegetation in early stages and therefore is expected to have the largest impact on the succession rate.
Since in early successional stages biotic input increases over time, both nitrogen and carbon concentrations have been reported to increase, moreover the rate of increase becomes larger over time (Berendse, Lammerts & Olff 1998). Secondly the ratio between C- and N-content is predicted to increase with time since the OM contains more carbon than nitrogen (van der Craats, 2010). When the soil has a very low (<10 ) C:N ratio, the efficiency in carbon consumption will be high (Manzoni, Trofymow, Jackson & Porporato, 2010). Although the C:N ratio is expected to increase during early successional stages, this increase is not expected to be continued, because when a high C:N ratio is reached the decomposition rate increases and microbes are expected to release more C as CO2. Overall, since all four factor have a positive correlation with the succession rate, they form a potential threat to the livelihood of the L.
Thesis – Soil development and succession, Texel
2015
The impact of managementNumerous studies have shown that over-stabilization of the site to a point where few new slacks are being formed by blowing sand and a reduction in the area of bare sand is a major threat to the occurrence of the L. Loeselii (van de Craats, 2011; Jones, 1988; Nienhuis & Gulati, 2002). Recently fixation of the dunes has been minimized and management techniques to promote pioneer species have been employed (Martínez, Maun & Psuty, 2004).
Mowing and grazing are traditional management techniques which prolong the pioneer stage by combatting
grass, shrub and tree encroachment. In the Horsmeertjes and the Kreeftepolder (Figure 1) these techniques are employed (Meijer et al., 2015), to favor the occurrence of non-competitive Red-List species. However, since grazing by larger cattle, such as horses and cows, severely disturbs sites, rabbits are the main grazers (Nienhuis & Gulati, 2002).
Sod cutting is one of the common practices in restoring or maintaining pioneer states in dune slack
(Shahrudin, 2014). Sod cutting encompasses the (partial) removal of the organic A-horizon to decrease the nutrient stock, however mechanical sod cutting is less precise and may also affect the mineral subsoil and plant populations (Nienhuis & Gulati, 2002). Currently, sod cutting is not executed in the studied dune slacks.
Research focus
Although several studies examined the dune slack development and its relation with the occurrence of early successional species such as L. Loeselii, uncertainties remain about site specific relevance of the acquired knowledge (Grootjans, 2004). Establishing knowledge of the occurrence and survival of pioneer species is essential to define an adequate management approach. Therefore long-term monitoring of actual dune slack development has large relevance in testing existing theories. As a follow-up study, this research attempts to create an integrated view on soil development in Texel’s young, wet dune slacks and its relation to the occurrence of the pioneer species L. Loeselii. The research questions are:
How do the soil parameters: depth of Ah-horizon, C:N ratio, carbon content, and nitrogen content; develop over time?
How do the soil parameters; Ah-horizon, C:N ratio, carbon content, and nitrogen content; interact with other soil parameters studied by Berghuis (2015) and van Middelaar (2015)?
How does soil development relate to the occurrence of L. Loeselii in Texel dune slacks?
The focus of this report is on the soil parameters: Ah-horizon, C:N ratio, C content and N content. However, synchronously with this study two other studies with the focus on pH, electrical conductivity, water level, soil bulk density, vegetation cover and vegetation characteristics were conducted. This enabled establishment of a more comprehensive understanding of Texel’s dune system.
Methods
This section consists of four parts describing: (1) the composition of dune slacks, (2)the fieldwork conducted to measure the soil development parameters, (3) the experiments in the soil laboratory and (4) the statistical analyses and comparison with the findings of the research by Van der Craat (2010) and Jongejans and Hollaar (2014).
Composition of dune slacks
Figure 4 depicts the spatial distribution of the 13 sampling locations and the four corresponding plots. At almost all locations mowing is practiced to limit overgrowth by Salix Repens L. (creeping willow). The series of dune slacks which are included in this study consist of: the 9 dune slacks (H1~9 in appendix II) which were studied in 2010 and 2014. Slack 10. which is added because in the summer of 2014 L. Loeselii was detected here for the first time. Slack 11, which is a very recent slack as it has been vegetated for at most 3 years. Finally, the other sampling locations H12 and H13 are, similar to H6, part of the ‘Horsvallei’, however during a recent flood in 2012 new sand was deposited at these locations therefore the soil has to redevelop hence they are now assigned to the youngest age group. Within the group of slacks, the locations H2 and H5 are classified as secondary slacks. This enables some comparison in development characteristics of primary and secondary slacks. Furthermore as shown in Appendix 1, the locations H4 and H6 have a high and a low sampling location. These were assigned in 2010 to the highest point of occurrence of L. Loeselii on the slope and the lowest point, and an adjacent higher and lower sampling plot
Thesis – Soil development and succession, Texel
2015
Figure 4: Map of the spatial distribution of all 2015 sampling locations. Source: OpenStreetMap
Field work
To enable comparison to previous datasets the same locations should be sampled as in 2011 and
2014 (Figure 4).
Therefore the previous
GPS locations were
uploaded onto a Yuma 2
GPS to support
orientation in the field. The thirteen research
locations and sixty
sampling sites (4-8 sites per location) require an identical sampling and observation procedure, therefore the treatment is described a single time.
Vegetation: a vegetation
relevée of one square meter, which is large enough to encompass all
characteristic species,
was marked and the
percentage of coverage for each of five categories was noted categories include: shrubs, herbs, mosses, bare sand and open water.
Soil description: an auger was used to make a soil profile. Horizons O, Ah, C1, C2 and C3 were described in terms of
depth, color, and signs of: mottling, oxidation and reduction. Furthermore the total soil depth and presence of roots in the profile was noted.
Water table: in the borehole the water table was measured after ten-fifteen minutes when the water table had
stabilized. Furthermore, throughout the observation period the water table was measured daily at two fixed locations. This was done to correct individual observations for changes in the water table during the fieldwork period.
Soil sampling: for each site three samples were taken which comes down to 180 samples in total. These samples were
taken with a 5cm diameter stainless steel pF_ring providing 100cm3 samples, which were stored in closed bags at a
temperature <100C.
Analysis in soil laboratory
To enable comparison with previous studies, chemical analyses of soil samples was almost identical to the methodology employed by Van der Craats (2010).
The fresh soil samples were weighed at two decimal precision (grams) and hereafter all bags were stored in the oven at 70°C for 48 hours. A 70°C temperature is suitable to prevent evaporation of any material except for soil moisture. Dry samples were again weighed at two decimal precision (grams). With these data, taking into account that the total
volume per sample is 200 cm3, the moisture content as well as the bulk density could be calculated with equation 2
(Appendix I).
Subsequently the samples were sieved, removing any material > 2 mm, such as roots. In preparation for the wet analysis, two times 10 grams of the sieved soil samples were solubilized in a 50 ml polyethylene bottle, for the first treatment with 25 ml of distilled water and for the second treatment with 25 ml of an 1M KCl solution. The extractions were shaken in an orbital shake machine for two hours and allowed to settle overnight. The next day, the samples were shaken again, this time for 30 minutes, after which electric conductivity (EC), pH(KCl) and pH(H2O)
Thesis – Soil development and succession, Texel
2015
activity after potassium (K+) has taken the place of the adsorbed H+ ions. Thus the latter represents the potential H+ activity. The samples with high pH have been tested for CaCO3 with a few drops of hydrochloric acid, to confirm none was present.
The CNS analyzer (Elementar Vario EL tube) requires extremely fine samples, so the samples were ground for 5 minutes at 400 rpm in an agate grinding bowl. These samples were heated again at 70°C for 24 hours, to make sure no moisture was left in the sample. Next, approximately 30-50 mg of ground material was placed in a tin cup intended for mineral sample analysis and then folded to make sure they were vacuum and nothing could spill. To be able to test the output all samples had a duplicate. The tin cups were weighed with two decimal (mg) precision, these weights were stored as an input for the instrument. To calibrate the instrument, three times eight tin cups with 7mg sulfanilic acid were analyzed before, in between and after the measurements. The tin cups were placed in the CNS analyzer where they were heated individually at 1150 °C in a combustion chamber. The output from the computer is a total percentage carbon and nitrogen and the C/N ratio. The average of the total percentage carbon and nitrogen is calculated for the samples with a duplicate. Percentages were then converted into mmol/kg according to equation 2 (Appendix I), with b=concentration of the measured parameter, and c=12 for the atomic mass of carbon and c=14 for the atomic mass of nitrogen.
The last step was to prepare the results for analysis with the help of 7 equations, shown in Appendix I.
Statistical analysis in Matlab®
Development over time of the soil parameters (Ah-horizon, C:N, C-content & N-content) and the dune slack age were
analyzed for correlation by calculating the Spearman’s Correlation Coefficient. Spearman's correlation applies to rank correlation and so provides a measure of a monotonic relationship between two continuous random variables. It is also useful with ordinal data and is robust to outliers. For significant correlations Matlab’s function ‘polyfit’ was used to fit a first order polynomial through all datasets, including those of previous years (2010, 2014) for the parameters: Ah-horizon, CN ratio, C- and N-content. Furthermore the parameters were tested for multi-linear regression over time with the function ‘regress’.
Visualization of this part includes a scatterplot of the raw data combined with the first order polynomial showing the development of all parameters over time for 2010, 2014 and 2015.
Comparing data with data of previous years by performing a non-parametric Friedman test. This analyzes whether
column effects are all the same or differ significantly, hereafter a multiple comparison test was performed to see which of the columns comparison intervals are disjoint.
Differences between dune slacks of the same age were tested on significant differences in means for each parameter,
since the dataset does not meet the assumption of equal variances, the non-parametric Kruskal-Wallis test was used. The method of testing for significant differences in dune slack development between slacks in one age group eliminates the impact time had on the development. Therefore conclusions can be drawn on how other factors influence soil development rates.
Secondly this is important for establishing the relation between soil parameters. The interactions will be analyzed in three ways: (1) based upon significant differences inside an age group (Kruskal-Wallis test and multiple comparison test), (2) examining the correlation coefficient between two parameters of an age group (Spearman correlation) and (3) by fitting and testing models (with ‘regress’ and ‘fitlm’) to predict response variables, taking 1 to 4 predictor parameters to generate the outcome of the response parameter.
Relations between soil parameters and the occurrence of L. Loeselii were analyzed by calculating the Spearman’s
Thesis – Soil development and succession, Texel
2015
Eq. 1 Ah-depth = 0.3997*C:N - 0.0354 * Bare - 0.0658
p-value = 7.75e-09 r2 adjusted = 0.462
Results
In this section first some overall trends are presented, followed by a more in-depth analysis of parameter predictions and interactions. To finalize specific information related to the occurrence of L. Loeselii is presented.
A significant increase of organic matter over time
Analysis of the development of the soil parameters over time (Ah-horizon, C:N, C, N), shows that all parameters significantly increase over dune slack age (Table 1). Both p-values for spearman correlation and regression are significant (P< 0.05). Figures 5a-d give a first impression of what the data look like for all 13 dune slacks and for all sampling years. Figures 6a-d show the fitted polynomial and observed values for all four parameters in 2015. Figures 7 a-d show this development for all three years.
Although most results show a distinct pattern, slack H1 shows conflicting results; firstly, the scatter within this location is wide and secondly, the average value is unexpectedly low. These results will be discussed in the next section.
Increased depth of the Ah horizon
Development over time; although the significance level of the regression line is high, the changes between the years
2010, 2014 and 2015 appear quite large as well (Figure 5a). However none of the datasets significantly differed (multiple comparison). This means that the variation between the datasets is not larger than within each dataset. Which is as expected since within each dataset the age gradient is approximately 40 years, whilst between datasets the maximal age variation is 5 years. Therefore this does not refute the observed regression over age within the 2015 dataset. This applies to all parameters.
Relation with other soil parameters
Within the four age groups the following results were obtained (Table 3a-d & 4): The depth of the Ah horizon was positively correlated with pH (3&21yr), shrubs (3&12yr), depth of water table (21yr), herbs(16yr), C-content(12yr), mosses (12yr), soil bulk density (3yr). Furthermore from Table 5 it can be understood that, for the younger sampling locations, the depth of the Ah horizon is often negatively correlated to the depth of the water table. This reflects that when the water table is more shallow the Ah horizon would be thicker.
How to predict the depth of the Ah horizon
In determining the best fit a multiple regression model was calculated. Based upon previous studies certain parameters were expected to be a predictor for the response parameter Ah horizon depth. These parameters include: age, C:N ratio, water level depth, percentage of bare soil and acidity. Tested the parameters showed that acidity has a significant, negative correlation with Ah-depth, whilst age has a significant, positive correlation with Ah-depth. However the ‘regress’ and ‘lmfit’ function gave the lowest p-value and highest r2 and adjusted r2 for a model with only ‘C:N ratio’ and ‘Bare’ as predictors. The formulated equation (1) is given in the textbox below. Another parameter which was also tested as a predictor value was ‘Site’. ‘Site’ is an 60 by 15 matrix with 4 ones per column representing the designated slack (Appendix X). Testing this model returned a p-value of 7.13e-13 and r2 adjusted of 0.77. This indicates that although Ah depth is significantly influenced by the studied parameters, the local circumstances per location are decisive.
Increase in carbon to nitrogen ratio
Development over time; C:N ratio is the parameter showing the strongest development over time. This is reflected by
Table 1: Correlation and regression coefficient for relations between dune slack age and the soil development parameters. Appendix III shows that differences in regression between age and percentage of N- and C-content and g/m2
of N- and C-content, are negligible.
Ah depth (cm) C:N ratio N-cont. (%) C-cont. (%)
Spearman correlation of dune slack age with
indicated soil parameters r = 0.733 p = 2.65e-11 r = 0.842 p = 3.36e-17 r = 0.7848 p = 1.17e-13 r = 0.8267 p = 4.06e-16 Regression coefficients and p-value where
Y indicates the soil parameter Y=a *Slack Age + b
a = 0.164 b = 1.673 p = 4.21e-07 a = 0.244 b = 8.372 p = 2.56e-13 a = 0.008 b = -0.028 p = 1.10e-06 a = 0.133 b = -0.738 p = 4.24e-07
Thesis – Soil development and succession, Texel
2015
Eq. 2 CN ratio = 0.0945*Age + 0.2123*Ah depth - 0.0766*Wat. tab. - 1.5802*pH + 21.6789
p-value = 1.2638e-17 r2 adjusted = 0.767
years 2010, 2014 and 2015 has continued steadily. Additionally, plotting the regression line for each of those years (Figure 7b) shows that the rate of C:N increase in 2015 is very similar to that in 2014.
Relation with other soil parameters
Similar to the comparison of Ah-depth with other parameters C:N can be compared. Within the age groups the following results were obtained (Table 2a-d & 3): positive correlations moss(16 yr), EC (3 yr) and with C- and N-contents at circa all ages. On the other hand C:N ratio has negative correlation with water table depth (3yr) and soil bulk density (16 yr).
How to predict the C:N ratio.
Based upon previous studies certain parameters were expected to be a potential predictor for the response parameter C:N ratio. These parameters included age, water level depth, percentage of bare soil and pH, all showed to be significant predictors. The lowest p-value and highest r2 and adjusted r2 were obtained in a model using predictors: age, depth of the Ah, water table depth and acidity (eq. 2). Note that the negative correlation with the depth of the water table encompasses that a high water table correlates with a large C:N ratio.
Increase in C-content
Development over time; C-content shows a steady increase over dune slack age. This is supported by the high
significance level of the correlation coefficient and the relatively low differences between the sampling years 2014 and 2015 (Figure 5c).
Relation with other soil parameters
For C-content, within age group comparisons gave the following results (Table 2a-d & 3): positive correlations with EC (16 yr), Ah (12 yr) and with C:N ratio and N-content at circa all ages. On the other hand C-content has a negative correlation with bulk density (16 yr).
How to predict the C-content
Based upon previous studies certain parameters were expected to be a potential predictor for the response variable C-content. These parameters included age, depth of Ah, bulk density, N-content, water level depth, soil coverage and pH. Bulk density, soil coverage, age and pH indeed resulted to be significant predictors (p-value 1,8e-7). However the multiple regression gave as an output the lowest p-value and highest r2 and adjusted r2 for a model using solely N-content as predictor (eq. 3). With an extremely low p-value and high r2. Using the same analysis for the 2010 and 2014 datasets provided the same results (p-value = 5.19e-43, r2 adjusted = 0.989)
Increase in N-content
Development over time; N-content shows a steady increase over dune slack age. The increase with age is supported by
the high significance level of the regression line which is high and the relatively low differences over the years 2014, 2015 (figure 5d).
Relation with other soil parameters
For N-content, within age group comparisons gave the following results (table 2a-d & 3): positive correlations with EC (16 yr), Ah (12 yr) and with C:N ratio and C-content at circa all ages. On the other hand N-content has a negative correlation with bulk density (16 yr).
How to predict the N-content
Multiple potential predictors were expected for N-content. Including: age, depth of Ah, bulk density, C-content, water level, soil coverage and pH. Soil coverage, bulk density, age and pH indeed resulted to be significant predictors (p-value 1,8e-7). However, as stated above, the lowest p-value and highest r2 and r2 adjusted were obtained in a model using solely C-content as an predictor (eq. 4).
Eq. 4 N-content = 0.0578* C-content + 0.0118
p-value = 5.5221e-59 r2 adjusted = 0.989
Eq. 3 C-content = 17.1065* N-content - 0.1864
Thesis – Soil development and succession, Texel
2015
Figure 5a:
Thesis – Soil development and succession, Texel
2015
Figure 5c:
Thesis – Soil development and succession, Texel
2015
Figure 6a:
Figure 6b:
Figure 6a:
Thesis – Soil development and succession, Texel
2015
Figure 6c:
Thesis – Soil development and succession, Texel
2015
Figure 7a:
Thesis – Soil development and succession, Texel
2015
Figure 7c:
Thesis – Soil development and succession, Texel
2015
Table 2b: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 16 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
Car. x 0.56 x - 0.98 0.81 x -0.66 x x x X
Nit. x 0.60 x 0.98 - 0.71 x -0.62 x x x x
C:N x X x 0.81 0.71 - x -0.75 0.58 x x X
Ah X X x X X - - - x x x 0.55
Table 2a: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 21 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
Car. x x x - 0.92 0.83 x x x nan x x
Nit. x x x 0.92 - 0.66 x x x nan x x
C:N x x x 0.83 0.66 - x x x nan x x
Ah 0.73 x 0.62 x x - - - x nan x x
Table 2c: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 12 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
Car. x x x - 0.83 0.90 0.61 x x x x x
Nit. x x x 0.83 - x x x x x x x
C:N x x x 0.91 x - x x x x x x
Ah X x x 0.61 x - - x 0.68 x 0.66 x
Table 2d: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 3 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
Car. x x x - 0.80 0.80 x x x x x x
Nit. x X x 0.80 - x x x x x x x
C:N x 0.66 -0.73 0.79 X - x x x x x x
Thesis – Soil development and succession, Texel
2015
Table 3: Comparing parameter values within different dune slack age groups. For all significant (p<0.05) differences between slacks the ranking is described. The last column is composed for the reason that H12 and H13 are originally of the same age as H6 and H7, but were flooded by which a large sand layer was deposited. By this column the impact of such flood on soil development can be studied. Note: for all significant differences the cell fill is green for all insignificant differences the cell fill is red.
3 yr 12 yr 16 yr 21 yr H6 H7 H12 H13
pH 11 significantly
higher than 13
8 significantly higher than 10
Not significant Not significant 12 significantly
higher than 6L
EC 11 significantly
higher than 13
not significant 7 significantly higher
than 5 & 6L 3 significantly higher than 4H. 7 is significantly higher than 13 Wat 13 is significantly higher than 12
Not significant 5 significantly higher
than 6L
4H significantly higher than 4L
13 is significantly higher than 6L & 7
Car. Not significant Not significant Not significant Not significant 6H significantly
higher than 13
Nit. 12 significantly
higher than 11
Not significant Not significant Not significant 6H significantly
higher than 13
C:N 12 significantly
higher than 13
Not significant Not significant 3 significantly
higher than 4H
6H significantly higher than 13
Ah 11 significantly
higher than 13
Not significant 6L significantly higher
than 7
Not significant 6L significantly
higher than 12 & 13 6H significantly higher than 12 Bulk 11 significantly higher than 12 8 significantly higher than 10 6L and 5 significantly higher than 6 4H significantly higher than 3 12, 13 significantly higher than 6H Moss 13 significantly higher than 11
Not significant Not significant Not significant 13 significantly
higher than 7
Bare Not significant Not significant Not significant Not significant 12 significantly
higher than 6H and 7
Shr. Not significant Not significant Not significant Not significant Not significant
Herb. 11 significantly
higher than 13
Not significant H6H, H6L and H7
significantly higher than 10
Not significant 7 significantly
higher than 12 and 13
Thesis – Soil development and succession, Texel
2015
Predicting the occurrence of Liparis Loeselii
Finally, the occurrence of L. Loeselii was tested on correlation with the soil parameters for both of the 2010 and 2014 datasets. The results are depicted in table 4 below. Noticeable is the fact that whereas C:N ratio was the single predictor for L. Loeselii occurrence in 2010, in 2014 all parameters show significant correlation. The correlation coefficients for predicting L. Loeselii occurrence ranking from most influential to least influential are: C:N, Ah depth, C-content and N-C-content. However in attempt to fit a model to the occurrence of L. Loesellii no significant p-values were obtained.
Table 4: Giving the significance of correlation and the correlation coefficient for a relation between L. Loesellii occurrence and the soil development parameters.
Ah depth C:N ratio N-content C-content
L. Loeselii 2010 occurrence r = 0.1276 p = 0.4090 r = -0.3106 p = 0.0401 r = 0.0078 p = 0.9599 r = 0.0069 p = 0.9645 L. Loeselii 2014 occurrence r = -0.3270 p = 0.0303 r = -0.3424 p = 0.0229 r = -0.3078 p = 0.0421 r = -0.3103 p = 0.0404
Thesis – Soil development and succession, Texel
2015
Discussion
This section provides further analysis and reflection of the results presented above, as well as suggested implications of the results and recommendation for further and future research.
Fieldwork set up
The distribution of sampling sites over slack ages is not homogenous with far fewer older (>25 yr) dune slacks. This complicates studying of (1) the development of the older locations, as well as (2) the trend in development rate over time. With regard to the results of H1, the unexpected OM measurements in this slack have a large impact on the overall studied regression which complicates drawing conclusions on the development. For this reason, supplementing the older age group is recommended.
As depicted in Figures 6 and 7 all four soil development parameters increase over time. When disregarding H1, this development appears well-nigh linear. All three studies (2010, 2014 and 2015) show a steady increase of the parameters, however cross-study comparison shows that follow-up measurements in the same slacks do not necessarily show a continuation pattern. This may have to do with spatial variation of the sampling sites over the years. Especially for non-observational linked results, such as C:N ratio, C- and N-content, this explains the variation. For results which are the product of human observation and interpretation, such as the Ah-depth, inconsistency in interpretation can be of influence as well. Therefore the cause of variation may be sought in both human subjectivity, and in alteration of the sampling location. A manner to resolve such conflicting findings in continuation patterns could be to sample more sites per slack which facilitates and strengthens the estimation of both spatial variation and average values of each parameter per slack.
Another aspect which this research aimed to investigate was the difference in development between secondary and primary slacks. For sampling location H5 this was an option as its age; 16 yr - equals that of H6 and H7. However for H2, the other secondary dune slack this is not the case, therefore it complicates comparison of soil development and parameter interactions. The results obtained by comparison of H5 in the 16 yr group with H6 and H7 has not provided a decisive answer to the question if dune slack ‘type’ is a major influence on soil development.
Interrelations of carbon and nitrogen
As depicted in Figure 6b low C:N ratios of between 6 and 11 were measured in the young dune slacks. These low initial values can be explained by the low litter input and therewith low levels of carbon, during this first part of the early succession stage. In this stage the soil has alkaline properties and vegetation growth is poor due to nitrogen limitation. Furthermore, the high pH values are known to have a positive effect on the decomposition rate of OM and therewith decrease the accumulation rate. An increase in the C:N ratio with age is observed which can be explained by the effect of microbial stoichiometry on mineralization rates. Which entails that the addition of carbon generally increases respiration and decreases nitrogen mineralization, whereas addition of nitrogen has the opposite effects (Manzoni et al., 2010; Buchkowski, Schmitz & Bradford, 2015). The high efficiency in carbon consumption by soil microbial life in the early succession stage relates to the low C:N ratio and low C-content. This well-nigh linear trend over the ages 12yr till 21yr, appears to decrease for the two oldest slacks. This may have to do with the fact that a C:N ratio in the range of 20-25 is ideal for maximum decomposition, since with this range a favorable soil environment is created to bring about equilibrium between mineralization and immobilization processes. In conclusion, C:N ratios reaching the value of 20 are expected to result in a decrease in OM accumulation and C:N ratio stabilization.
Best predictors
Although different parameters were found to predict soil C- and N-content, the prediction of the one as the product of the other was by far the most reliable (eq. 3 and eq. 4). This has to do with (1) the strong dependency of aboveground biomass on nutrient availability and (2) low concentrations of inorganic nitrogen which has to do with the coarse soil texture and shallow water table (Kachi & Hirose, 1983 ;Stanford & Epstein, 1974). Therefore a very strong correlation is expected between the presence of N and that of OM. Hence C and N accumulation go hand in hand.
Annual variation in hydrological regimes may also be an important predictor for especially N-accumulation, since the depth of the water table influences nitrogen mineralization, nitrification and nitrate leaching. Monitoring of annual variation is likely to provide insight into this soil-water-nitrogen interaction.
Thesis – Soil development and succession, Texel
2015
(Appendix X) and therefore does not correspond to an influential factor except for reflecting the coherence at one sampling location, this indicates that specific conditions at one location are the most influential parameter in predicting the formation of an Ah horizon. These conditions may include: slope, the occurrence of specific plant species, local hydrological fluctuations, local wind- and precipitation patterns. Bare sand and C:N ratio are other significant predictors for Ah horizon. This has to do with the fact that the C:N ratio increases simultaneously with the accumulation of OM in the topsoil, as reasoned above (Subheading: “Interrelations of carbon and nitrogen”). Furthermore, the percentage of soil which is covered results in a strong correlation, as the Ah horizon needs input of litter by plants.
Recommendations
In the dune slack composition two slacks were included with four high and four low sampling sites, these locations were compared and showed no dominant influence on any of the soil parameters. The fact that the sites do not significantly differ indicates that the altitude of the sampling site is not a dominant indicator for OM, yet it is advisable to test this in a follow-study for the conclusion is based upon the comparison of eight high with eight low
measurements.
Since all slacks included in this study are under uniform management regimes, the impact of different management regimes on soil development and succession rate cannot be assessed. However, a way to assess the impact of the current management regime is to regard it in wider context; as a form of Dynamic Coastal Management. This can be done by assessing the impact of the new management form on coastal landscape formation by sand drift and wash over with salt water. The aim of this management form is to prevent the ageing of vegetation and to increase the transitional areas in terms of salt, water, altitude and soil gradients (van der Spek & Elias 2013). The impact can be measured in two ways.
(1) New observations on L. Loeselii occurrence of 2015 will be available soon. It is recommended to assess the impact of flooding on the occurrence of the L. Loeselii, since the two neighboring locations H12 and H13 where recently studied and last year’s data on the occurrence of the L. Loeselii suggest a positive effect of flooding, in the year 2012, on the number of L. Loeselii.
(2) Another option is to include more abiotic factors in a follow-up study as to create a more integrated view on the influences of the eolian processes in the area. Recommended areas of study are variation in hydrological regimes, flood movements and sand movements.
Conclusion
This section describes the patterns, principles and relationships of soil development parameters and succession which have been explored and/or confirmed by this study.
Soil development over an age gradient
In line with previous studies the soil development parameters: Ah-horizon, C:N ratio, C-content and N-content showed an increase over the age gradient. Together with bulk density these parameters can be regarded as representatives for the soil organic matter. The two major drivers of accumulation of OM are the trends of decreasing pH and increasing soil coverage. Based upon the 2015 findings the rate of OM accumulation decreases when the slack ages.
L. Loesellii had a stronger correlation with the soil parameters in 2014 than in 2010. These results showed
that sustained increase of the Ah horizon, C:N ratio, C- and N contents in the soil form a threat to the occurrence and persistence of L. Loesellii, as was expected for a pioneer species. As discussed above, favorable conditions for a prolonged pioneer stage can be achieved by increasing the dynamic eolian processes. These eolian processes which are promoted by dynamic coastal management, are anticipated to influence many soil development parameters. Firstly, by depositing sand in the dune slacks, which increases the pH (Berghuis, 2015) and decreases both soil coverage (van Middelaar, 2015) and the OM content. Secondly, brackish water is more likely to flood established dune slacks, disturbing both the topsoil and furthermore increasing the amount of salts (Berghuis, 2015). Therefore an increase in eolian processes could very well be a window of opportunity for pioneer species since, tall grasses and other later successional species are less competitive under these circumstances.
Thesis – Soil development and succession, Texel
2015
Evaluation & Acknowledgement
The bachelor project enabled me to exercise all skills acquired during my three-year bachelor program Future Planet Studies. With the research I practiced my investigative, collaboration, presentation, planning and writing skills. I appreciated the relatively long-term character of the bachelor project which especially taught me to anticipate on my findings and to practice long-term planning.
As a final addition I wish to thank A. Kooijman for her guidance throughout the thesis and especially during the fieldwork on Texel. Furthermore, the guidance given by J. Schoorl and L. Hoitinga during the soil analysis in the laboratory was much appreciated. Assistance provided by E. van Loon concerning the statistical analysis and by T. de Boer concerning the GPS device and ArcGIS, has been very helpful, for which I am grateful. At last, I want to show my gratitude to H. Berghuis and J. van Middelaar, with whom I worked side-by-side with throughout the research.
Thesis – Soil development and succession, Texel
2015
Reference list
Berendse, F., Lammerts, E. J., & Olff, H. (1998). Soil organic matter accumulation and its implications for nitrogen mineralization and plant species composition during succession in coastal dune slacks. Plant Ecology, 137(1), 71-78.
Berghuis, H. S., (2015). Soil development and interactions in dune slacks of the dynamic coastal area De Hors, Texel: A field study concerning the acidity, electrical conductivity and water table. Bachelor thesis.
Buchkowski, R. W., Schmitz, O. J., & Bradford, M. A. (2015). Microbial stoichiometry overrides biomass as a regulator of soil carbon and nitrogen cycling. Ecology, 96(4), 1139-1149.
van der Craats, A., (2010). Window of opportunity of Liparis loeselii in relation to dune slack development on Texel.
Master thesis.
Elias, E. P., & van der Spek, A. J. (2006). Long-term morphodynamic evolution of Texel Inlet and its ebb-tidal delta (The Netherlands). Marine Geology, 225(1), 5-21.
Ernst, W. H. O., Slings, Q. L., & Nelissen, H. J. M. (1996). Pedogenesis in coastal wet dune slacks after sod-cutting in relation to revegetation. Plant and soil, 180(2), 219-230
Jones, P. S. (1998). Aspects of the population biology of Liparis loeselii (L.) Rich. var. ovata Ridd. ex Godfery (Orchidaceae) in the dune slacks of South Wales, UK. Botanical journal of the Linnean Society, 126(1‐2), 123-139.
Jongejans, L. (2014). Exploring the change in succession in dune slacks in the Hors, Southwest Texel between 2010 and 2014, focussing on the acidity and electrical conductivity of the soil and the water table level: a field study.
Bachelor thesis.
Grootjans, A., Stichting, E. R. A., Stuyfzand, P., Petersen, J., & Shahrudin, R. (2014). Ontwikkeling van zoet-zoutgradiënten met en zonder dynamisch kustbeheer.
Grootjans, A. P., Adema, E. B., Bekker, R. M., & Lammerts, E. J. (2004). Why coastal dune slacks sustain a high biodiversity. In Coastal Dunes (pp. 85-101). Springer Berlin Heidelberg.
Grootjans, A. P., Geelen, H. W. T., Jansen, A. J. M., & Lammerts, E. J. (2002). Restoration of coastal dune slacks in the Netherlands. In Ecological Restoration of Aquatic and Semi-Aquatic Ecosystems in the Netherlands (NW Europe) (pp. 181-203). Springer Netherlands.
Grootjans, A. P., Everts, H., Bruin, K., & Fresco, L. (2001). Restoration of Wet Dune Slacks on the Dutch Wadden Sea Islands: Recolonization After Large‐Scale Sod Cutting. Restoration Ecology, 9(2), 137-146.
Grootjans, A. P., Hartog, P. S., Fresco, L. F. M., & Esselink, H. (1991). Succession and fluctuation in a wet dune slack in relation to hydrological changes. Journal of Vegetation Science, 2(4), 545-554.
Hollaar, T. P. (2014). Exploring the influence of succession in wet coastal dune slacks on pioneer species Liparis Loeselie, with focus on soil organic matter and nutrient concent along an age gradient. Bachelor thesis.
Kachi, N., & Hirose, T. (1983). Limiting nutrients for plant growth in coastal sand dune soils. The Journal of Ecology, 937-944.
Thesis – Soil development and succession, Texel
2015
Kooijman, A. M., & Besse, M. (2002). The higher availability of N and P in lime‐poor than in lime‐rich coastal dunes in the Netherlands. Journal of Ecology,90(2), 394-403.
Kooijman, A. M., Dopheide, J. C. R., Sevink, J., Takken, I., & Verstraten, J. M. (1998). Nutrient limitations and their implications on the effects of atmospheric deposition in coastal dunes; lime‐poor and lime‐rich sites in the Netherlands. Journal of Ecology, 86(3), 511-526.
Lammerts, E. J., Maas, C., & Grootjans, A. P. (2001). Groundwater variables and vegetation in dune slacks. Ecological
Engineering, 17(1), 33-47.
Manzoni, S., Trofymow, J. A., Jackson, R. B., & Porporato, A. (2010). Stoichiometric controls on carbon, nitrogen, and phosphorus dynamics in decomposing litter. Ecological Monographs, 80(1), 89-106.
Martínez, M. L., Maun, M. A., & Psuty, N. P. (2004). The fragility and conservation of the world's coastal dunes: geomorphological, ecological and socioeconomic perspectives. In Coastal Dunes (pp. 355-369). Springer Berlin Heidelberg.
Meijer, J.,Bilius, M., and Vriens, G., (2015). Document PAS-analyse: Herstel-strategieën voor Texel.
Van Middelaar, J. C., (2015). Dune slack succession - Succession in dynamic, wet dune slacks - southern Texel. Bachelor
thesis.
Nienhuis, P. H., & Gulati, R. D. (Eds.). (2002). Ecological restoration of aquatic and semi-aquatic ecosystems in the
Netherlands (NW Europe) (Vol. 166). Springer Science & Business Media.
Rijksoverheid. (2015). Natuur & Biodiversiteit. Collected from Rijksoverheid: http://www.rijksoverheid.nl/onderwerpen/natuur-en-biodiversiteit/natura-2000.
Shahrudin, R. (2014). Do we really need management to preserve pioneer stages in wet dune slacks? (Doctoral dissertation, PhD Dissertation University of Groningen, Groningen).
Stanford, G., & Epstein, E. (1974). Nitrogen mineralization-water relations in soils. Soil Science Society of America
Journal, 38(1) , 103-107.
Stanová, S.V., Šeffer, J., & Janák, M. (2008). Management of Natura 2000 habitats. 7230 Alkaline fens. Tehnical
Report, 20(24), 20.
van der Spek, A., & Elias, E. (2013). The effects of nourishments on autonomous coastal behaviour.Coastal Dynamics, 221, (pp.1-10).
Westhoff, V., and Van Oosten, M.F., 1991, De plantengroei van de Waddeneilanden: Utrecht, Stichting Uitgeverij Koninklijke Nederlandse Natuurhistorische Vereniging, 417 p.
Thesis – Soil development and succession, Texel
2015
Reference list figures
Picture on front page: ‘Enkele bloeiende exemplaren van de groenknolorchis in een loopduin (secundaire uitblazingsvallei).’ Available from: http://duinenenmensen.nl/komen-en-gaan-van-een-kleine-pionier-groenknolorchis/. Viewed on 20 April 2015.
Figure 1: The aerial view on the southern tip of Texel. Source: Staatsbosbeheer, 2008: https://staatsbosbeheertexel.wordpress.com/tag/kreeftepolder/. Viewed on 20 April 2015.
Figure 2: Available from: http://www.ecomare.nl/typo3temp/GB/b61569e759.png. Viewed on 20 April 2015.
Figure 3: typical transect with low embryo dunes” Copyright ©2001 David Rayner Beagle Graphics (GeoResources). Available from: http://www.georesources.co.uk/csd1.htm. Viewed on 20 April 2015.
Figure 4: J. W. van Aalst. adjusted by D. Stomph. Map of the spatial distribution of all 2015 sampling locations.
Available from: OpenStreetMap GIS open street map, with GPS sampling locations of the thirteen sampling locations in 2015. Created on: 15 June 2015.
Thesis – Soil development and succession, Texel
2015
Appendix I: Equations
Eq. 1: Moisture content (%)
Eq. 2: Bulk density (gr/cm3)
Eq. 3: Total alkalinity [mmol OH-/L]
* a = “consumption of HCl at the last (first or second) inflection point”.
Eq. 4: Converts percentages of the concentration (CNS-results) to mmol/kg * b = concentration of the measured parameter
* c = 12 for the atomic mass of carbon and c = 14 for the atomic mass of nitrogen
Eq. 5: Converts concentration in water extracts from μmol/L to mmol/kg * d = moisture loss at 105 °C
* f = 105 °C oven dry sample
Eq. 6:Converts concentration in water extracts from mg/L to mmol/kg
Eq. 7: Converts all concentration in mmol/kg to concentration in mmol/m2. Giving the concentration in the relevèe
Thesis – Soil development and succession, Texel
2015
Appendix II: Dune slack specifics
Table 1: Thirteen sampling locations. The sampling plots on the lower part of the slope are indicated as H4L and H6L, the higher part of the slope as H4H and H6H. “-L” indicates the presence of L. Loeselii in 2010. The quantity of sampling plots per sampling location is shown in the column “n”. Furthermore, the table lists the year at which a vegetation cover established and when the first
L. loeselii was observed in the dune slack. Adjusted table from: Van der Craats, 2011.
Location Code n Age of the vegetation cover L. loeselii arrived
Dune slacks N-Nw of western
Horsmeertje H1 4 1970 1978
Grauwe Gans vallei H2-L 4 1986* 1992
Kreeftepolder East H3-L 4 1994 1995
Kreeftepolder middle – wet H4L-L 4 1994 1995
Kreeftepolder middle – dry H4H-L 4 1994 1995
Secondary dune slacks north of
Horsvallei H5-L 4 1999* 1998
Horsvallei – wet H6L-L 4 1999 1998
Horstvallei - dry H6H-L 4 1999 1998
Saltier part Horsvallei H7-L 4 1999 1998***
Western new dune slack H8-L 4 2003 2008
Future slack in Horsvallei H9 4 2009 X
Future slack in Horsvallei H10 4 2003 2014
Future slack H11 4 2012 X
Horsvallei H12 4 2012** X
Horsvallei H13 4 2012** X
Total 60
* Secondary blowing out processes
** Originally from same age as H6, however after new sand deposits (>5cm) new soil and vegetation developments
Thesis – Soil development and succession, Texel
2015
Appendix III: Examining the correlation for both percentage and g/m2 for both C- and N-content
Table 1b: Correlation and regression for N and C in percentage and g/m2 of the top 5cm.
N-cont. (%) N-cont.(g/m2) C-cont. (%) C-cont.(g/m2) Spearman correlation of dune slack age
Thesis – Soil development and succession, Texel
2015
Appendix IV: output of examining the best fit linear model with ‘lmfit’
Y = Ah depth X1 = constant X2 = age
Linear regression model: y ~ 1 + x1 + x2
Estimated Coefficients:
Estimate SE tStat pValue ________ ________ ______ __________
(Intercept) 0 0 NaN NaN x1 1.6734 0.56154 2.9801 0.0042298 x2 0.16425 0.028809 5.7014 4.4118e-07
Number of observations: 60, Error degrees of freedom: 58 Root Mean Squared Error: 2.28
R-squared: 0.359, Adjusted R-Squared 0.348
F-statistic vs. constant model: 32.5, p-value = 4.21e-07
Y = C:N
X1 = constant X2 = Age X3 = Water X4 = Bare
Linear regression model: y ~ 1 + x1 + x2 + x3 + x4
Estimated Coefficients:
Estimate SE tStat pValue _________ ________ _______ __________
(Intercept) 0 0 NaN NaN x1 10.749 0.63889 16.825 2.3933e-23 x2 0.17385 0.026296 6.6112 1.6325e-08 x3 -0.078996 0.022688 -3.4818 0.00098367 x4 -0.031534 0.010623 -2.9683 0.0044275
Number of observations: 60, Error degrees of freedom: 56 Root Mean Squared Error: 1.73
R-squared: 0.728, Adjusted R-Squared 0.713
Thesis – Soil development and succession, Texel
2015
Y = N contentX1 = constant X2 = C content
Linear regression model: y ~ 1 + x1 + x2 + x3
Estimated Coefficients:
Estimate SE tStat pValue _________ _________ _______ __________
(Intercept) 0 0 NaN NaN x1 -0.27132 0.057758 -4.6976 1.7501e-05 x2 0.0069393 0.0036096 1.9225 0.059638 x3 16.794 0.27925 60.141 1.4389e-52
Number of observations: 60, Error degrees of freedom: 57 Root Mean Squared Error: 0.232
R-squared: 0.99, Adjusted R-Squared 0.99
F-statistic vs. constant model: 2.84e+03, p-value = 8.54e-58
Y = C content X1 = constant X2 = N content
Linear regression model: y ~ 1 + x1 + x2 Estimated Coefficients:
Estimate SE tStat pValue ________ __________ ______ __________ (Intercept) 0 0 NaN NaN x1 0.011809 0.0021267 5.5527 7.6745e-07 x2 0.057838 0.00078569 73.614 3.3141e-58
Number of observations: 60, Error degrees of freedom: 58 Root Mean Squared Error: 0.0138
R-squared: 0.989, Adjusted R-Squared 0.989
Thesis – Soil development and succession, Texel
2015
Appendix V: Results correlation between two parameters of ages 21, 16, 12 and 3 yr
Table 3b: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 16 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
pH - x x x x x x x x x x x EC x - x 0.56 0.60 x x -0.57 x x x -0.70 Wat x x - X X x x X x x x X Car. x 0.56 x - 0.98 0.81 x -0.66 x x x X Nit. x 0.60 x 0.98 - 0.71 x -0.62 x x x x C:N x X x 0.81 0.71 - x -0.75 0.58 x x X Ah X X x X X - - - x x x 0.55 Bulk X -0.57 x -0.66 -0.62 -0.75 x - x x x 0.49 Moss X x x x 0.55 x X - x x x Bare x X x x x X x X x - x x Shr. x X x x x x x X x x - x
Table 3a: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 21 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb
pH - x 0.63 x x x 0.73 x x nan x x EC x - x x x x x -0.92 0.75 nan x x Wat 0.63 x - x x x 0.62 x x nan x x Car. x x x - 0.92 0.83 x x x nan x x Nit. x x x 0.92 - 0.66 x x x nan x x C:N x x x 0.83 0.66 - x x x nan x x Ah 0.73 x 0.62 x x - - - x nan x x Bulk x -0.92 X X X X X - X nan X X Moss x 0.75 X X X X X X - nan X -0.96
Bare nan nan nan nan nan nan nan nan nan - nan nan
Shr. x x x x x x x x x nan - x
Thesis – Soil development and succession, Texel
2015
Table 3c: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 12 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
pH - x x x x x x 0.76 x x -0.75 x EC X - x x x x x x x x x x Wat x x - x x x x x 0.75 x x x Car. x x x - 0.83 0.90 0.61 x x x x x Nit. x x x 0.83 - x x x x x x x C:N x x x 0.91 x - x x x x x x Ah X x x 0.61 x - - x 0.68 x 0.66 x Bulk 0.76 x x x x x x x x x x x Moss x x 0.75 x x x x x x x x x Bare x x x x x x x x x x x -0.88 Shr. -0.75 x x x x x x x x x x -0.75 Herb. x x x x x x x x x -0.88 -0.75 x
Table 3d: Rho for all significant (p<0.05) correlations between parameters in dune slacks aged 3 yr. Note for all positive correlation the cell fill is green for all negative correlations the cell fill is red.
pH EC Wat. Car. Nit. C:N Ah Bulk Moss Bare Shr. Herb.
pH - 0.66 X x x x 0.66 0.63 -0.86 x -0.64 0.70 EC 0.66 - -0.59 x x 0.66 x x -0.88 0.58 -0.62 0.78 Wat X -0.59 - x x -0.73 x x 0.60 x x x Car. x x x - 0.80 0.80 x x x x x x Nit. x X x 0.80 - x x x x x x x C:N x 0.66 -0.73 0.79 X - x x x x x x Ah 0.66 X X X X - - 0.69 x x -0.63 x Bulk 0.6 X X x x x 0.69 - -0.64 x Moss -0.86 -0.88 0.60 x x x x -0.64 - x 0.60 -0.85 Bare x 0.58 x x x x x x x - x x Shr. -0.64 -0.62 x x x x -0.63 x 0.60 x - x
Thesis – Soil development and succession, Texel
2015
Appendix VI: Matlab Script for explorative analysis
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Dienke Stomph
% Bachelor thesis concerning the succession in dune slacks at De Hors, Texel %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %start close all clear clc %load data texeldata = 'H:\teksel\Texel_complet.xls' texel = xlsread(texeldata);
%general data [years]
Age10=texel(1:44,7); %9 slacks
Age14=texel(46:93,7); %10 slacks
Age15=texel(98:157,7); %13 slacks
%general data [slacks]
Slack10 = texel(1:44,2); Slack14 = texel(46:93,2); Slack15 = texel(98:157,2);
%water [cm]
Water10 = texel(1:44,30); %water below soil surface
RHeight10= texel(1:44,29); %relative height with respect to the deepest point
Water14 = texel(46:93,30); RHeight14 = texel(46:93,30); Water15 = texel(98:157,30); RHeight15= texel(98:157,29);
WaterDeepest15 = texel(98:157,49);
%water table (with respect to soil surface)separated for slacks
Water10_1=Water10(1:4);Water14_1=Water14(1:4);Water15_1=Water15(1:4); Water10_2=Water10(5:8);Water14_2=Water14(5:8);Water15_2=Water15(5:8); Water10_3=Water10(9:12);Water14_3=Water14(9:12);Water15_3=Water15(9:12); Water10_4L=Water10(13:16);Water14_4L=Water14(13:16);Water15_4L=Water15(13:16); Water10_4H=Water10(17:20);Water14_4H=Water14(17:20);Water15_4H=Water15(17:20); Water10_5=Water10(21:24);Water14_5=Water14(21:24);Water15_5=Water15(17:20); Water10_6L=Water10(25:28);Water14_6L=Water14(25:28);Water15_6L=Water15(25:28); Water10_6H=Water10(29:32);Water14_6H=Water14(29:32);Water15_6H=Water15(29:32); Water10_7=Water10(33:36);Water14_7=Water14(33:36);Water15_7=Water15(33:36); Water10_8=Water10(37:40);Water14_8=Water14(37:40);Water15_8=Water15(37:40); Water10_9=Water10(41:44);Water14_9=Water14(41:44);Water15_9=Water15(41:44); Water15_10=Water15(45:48); Water14_11=Water14(45:48);Water15_11=Water15(49:52); Water15_12=Water15(53:56); Water15_13=Water15(57:60); %PH pH10 = texel(1:44,16); pH14 = texel(46:93,16); pH15 = texel(98:157,16);
%pH separated for slacks
pH10_1=pH10(1:4);pH14_1=pH14(1:4);pH15_1=pH15(1:4); pH10_2=pH10(5:8);pH14_2=pH14(5:8);pH15_2=pH15(5:8); pH10_3=pH10(9:12);pH14_3=pH14(9:12);pH15_3=pH15(9:12);
