Author:
N. S. Dreijer
Primary supervisor:
Dr. L.H. Cammeraat
Secondary supervisor:
Dr. W.E. Morriën
Submitted:
April 2021, Amsterdam
Study program:
Future Planet Studies
Specialization:
Earth Sciences
BACHELOR THESIS
Short-term effects of non-inversion tillage and
straw-rich manure application on a marine clay
soil: A surface water and soil solution analysis
2
Abstract
Soil management can affect agricultural productivity and emissions to water and air, consequently affecting ecosystems and human health. With regard to low emission manure management, Dutch legislation demands farmers to incorporate farmyard manure in the soil by inversion tillage. The objectives of this study were to determine and explain the short-term effects of non-inversion tillage and straw-rich farmyard manure application on solute concentrations in surface water and soil solution of a marine clay soil in the Netherlands. Surface and soil water measurements from the first year of a multiyear (2019-2023) study were created, compiled and compared with general clay soil data from the Netherlands. Two field sites were established on cauliflower cultivated fields and are located approximately two kilometres apart. The samples were retrieved during late autumn and winter on four plots with non-inversion tillage (NT) and three plots with inversion tillage (T), both groups combined with straw-rich manure. Concentration changes of NO3- and NH4+ ions were analysed in surrounding surface water and at three different depths (Ap, A/C and C horizon) within the upper first meter of the soil. The mean NO3- concentration in soil solution fluctuated over time and depth and was different between the two field sites. NT treatment showed significantly lower NO3- (p = 3.9e-4) and on the other hand significantly higherNH4+ concentrations in soil solution (p = 0.0115). No significant difference in solute concentration was found for soil horizon Ap or A/C. Nevertheless, the C horizon showed a significantly lower concentration of NO3- (p = 0.00147) for NT. Similar to the total result, NH4+ concentration in the C horizon for NT was significantly higher (p = 0.006836) than the concentration in the C horizon of T. Mild weather conditions resulted in active nitrification in November. In combination with a relatively high input of organic material from crop residues, the highest NO3- concentration in either surface water and soil solution was observed in November 2020 after harvest. Surface water and soil solution maximum NH4+ -concentration was found in February 2021. All NO3- values were lower than the 2018 average nitrate concentration for surface waters on clay soils in the Netherlands and remained under 806 µmol NO3- / L of the EU Nitrate Directive limit.
Keywords: Water quality – arable farming – leaching – inversion tillage – manure – groundwater – soil solution – soil fertility – nitrate – ammonium – fertilizer policy – sustainable agriculture
3
Table of contents
1. Introduction ... 4
2. Materials and methods ... 5
2.1. Study site ... 6
2.2. Experimental design ... 7
2.3. Sampling and analysis ... 8
2.3.1. Soil solution, surface and rainwater ... 8
2.3.2. Lab analysis of samples ... 8
2.3.3. Data analysis ... 10 3. Results ... 11 3.1. Surface water ... 11 3.2. Soil solution ... 12 4. Discussion ... 17 5. Conclusion ... 19 Acknowledgements ... 20 References ... 21 Appendices ... 23
Appendix A – p-value tables ... 23
4
1.
Introduction
The world population has grown with more than seven billion people in the last 220 years to 7.8 billion people (United Nations, 2019). With a current average population growth of 81 million persons per year, it is estimated that the world population will reach ten billion people in 2057. Rapid growth of the global population goes hand in hand with an increase in food demand (United Nations, 2015). To cover food demand, farming systems that are known as conventional or industrial farming developed agricultural operations that ensured high productivity by primarily increasing nutrient availability to crops (Jain, 2010). Conventional farming became globally widespread after the second world war and is currently the main source of food for humans and animal stock. Half of all habitable land on Earth is used for agricultural purposes and with that, it is the largest engineered ecosystem by humankind. Widely used conventional agricultural methods are synthetic fertilizers, herbicides, pesticides, monocultures, intensive irrigation and tillage. Although these technologies increase yield quantities and are cost-effective in the short term, some disadvantages may cause irreversible problems in the long term (Liu et al., 2010). Primary effects caused by conventional methods are soil erosion and organic matter loss. These effects will eventually negatively affect soil fertility (Derpsch, 2003). Ultimately, loss of soil fertility can deprive the entire ecosystem of other important ecosystem services (Zhang et al., 2007). Services such as biodiversity, carbon sequestration, nutrient cycling and water provision and quality can lose their ability to function and will subsequently decrease agricultural productivity (Figure 1).
Besides phosphate (P2O5), ammonium (NH4+) and nitrate (NO3-) can be considered as main crop nutrients. Crops mostly prefer NO3- in aerobic(semi) arid environments andNH4+ in anaerobic wetter environments (Britto & Kronzucker, 2013). Nitrogen fixation, mineralization and nitrification are processes that increase crop available forms of nitrogen, whereas denitrification, volatilization, immobilization and leaching result in a decrease. Nitrate is a negatively charged ion and therefore mostly not attracted to the predominantly negatively charged clay particles (Cui et al., 2020). Due to the low attraction to soil particles, nitrate may leach out easily when water percolates through the soil. Nitrate leaching is most common in coarse-textured soils in comparison with finer-textured soils.
Soil aggregate formation is a major control factor in the process of soil structure formation which determines the amount of nitrate and other nutrient losses caused by leaching. Aggregates are arrangements of primary particles of sand, silt and clay around soil organic carbon (SOC) that are bound together (Adeyemo, 2010; Oades, 1984). Besides, fungi, bacteria, roots and mineral particles contribute to the cohesion and fragmentation of the aggregates. The pore space between aggregates enables water and air transmission which stimulates microbial processes and water uptake by plants. Tillage destroys aggregates by physical fragmentation and disturbing the microbial balance (Hansen & Djurhuus, 1997; McLaughlin & Mineau, 1995).
Agriculture is the greatest contributor of nitrate emissions to freshwaters in Europe (Grizzetti et al., 2007). The EU Nitrates Directive (EU, 1991) legislation aims to protect water quality by demanding every member state to comply with the prevention of water pollution by nitrate leaching from agricultural sources. The directive allows a maximum concentration of 50 mg NO3- / L or806 µmol NO3- / L in groundwater. To minimize nitrogen leaching, the Dutch government implemented several regulations on manure management. With regard to low emission manure management, current Dutch legislation demands farmers to incorporate farmyard manure into the soil by inversion tillage (T) immediately after broadcasting. Although nitrogen and phosphorus concentrations in the Netherlands on average decreased since 1990, surface water and groundwater in the upper layers of the soil still contain excessive amounts of these nutrients. Nitrate concentrations in surface and groundwater on clay and peat soil farmlands decreased almost everywhere in the Netherlands after the implementation of the Nitrates Directive in 1991. Nevertheless, measurements on sandy soils showed merely a decrease at fifty percent of the monitored locations (RIVM, 2020, n.d.). The cause of these differences is the low denitrification capacity for coarse-grained soils like sand. According to the nitrate monitoring report from the National Institute for Public Health and Environment (RIVM, 2020), an increase of nitrogen and phosphor has been measured in all soils of agricultural fields since 2015. The increase had subsequently induced a rise in nitrate concentrations of local surface waters and groundwater. Dry summers and an increase in nitrogen-rich fertilizer are accountable for the outcomes of the measurements (RIVM, 2020).
Climate conditions interact with the effect of tillage and manure management practices on nitrate leaching. According to the Royal Netherlands Meteorological Institute (KNMI, 2020), the average amount of rainfall in the Netherlands has increased from 692 mm to 873 mm a year between 1910 and 2019. The winter months have seen the largest increase with 43% on average within these 110 years. At the same time, longer and intense periods of drought occur during warmer seasons. The drought affects nitrogen uptake by crops and will lead to early leaf senescence (Leitner et al., 2020). Eventually, the leaves will increase soil nitrogen when decomposed. Moreover, microbes will increasingly resource gaseous and volatile substrates to survive as soluble resources are inaccessible in dry soils (Schimel, 2018). This will cause accumulation of nitrate at the soil surface and run-off when rainfall will appear.
Marine clay soils account for 23% of the soil distribution in the Netherlands. Few studies have researched the effect of inversion tillage on leaching in temperate climates with marine clay soils. Nitrogen transformation and leaching in agricultural soils are complex processes and depend on several factors. Nevertheless, some studies have shown that the timing and practice of manure application can impact nitrate leaching by affecting the amount and distribution of NO3- between rapid-flow and slow-flow drainage domains (Hansen & Djurhuus, 1997; Stenberg et al., 1999; Meisinger et al., 2015). In addition, timing and amount
5
of rainfall in relation to the time of tillage can influence total percolation and the percentage of percolation that is contributed by preferential flow, which are two main factors affecting nitrate leaching.The present paper contributes to the first year of an agricultural field study on the effects of non-inversion tillage (NT) on marine clay soils of the West-Friesland region in the Netherlands. The study will be conducted for four consecutive years as of March 2020. The overall aim of the research is to study biodiversity, yield, soil quality, crop quality and water quality under NT and T treatments and eventually change current Dutch legislation in favour of non-inversion tillage management. Adaptation of NT agricultural management could potentially contribute to minimalization or prevention of soil degradation to improve ecosystem services on a global scale.
Water quality is one of the main subjects within the four-year study and this paper describes and elaborates on the results of the first year. This study aimed to create, compile and compare surface and soil water analysis results and general water quality information of two field sites on marine clay soils during late autumn and winter. To achieve the aim of this study the following objectives have been formulated:
• to determine and explain short-term effects of NT and T straw-rich manure management on solute concentrations in surface water and soil solution of a marine clay soil
• to determine differences in NO3- in NH4+ concentrations between the two field sites of this study
• to assess NO3- and NH4+ concentration changes with depth for the A , A/C and C horizons of the soil during late autumn and winter
• to assess NO3- and NH4+ concentration changes in surface water bodies during late autumn and winter The emphasis of this study was mostly on the NO3- ion as this solute is currently addressed as the main polluting agent in soil solutions and water bodies as an effect of agricultural practices.
Tillage & carbon deprived manure management
Disruption soil structure
Surface runoff, soil erosion and nutrient leaching
Decrease nutrient availability
Decrease crop production
Figure 1 Simplified problem description of carbon deprived manure combined with tillage management. The emphasized subjects of the
6
2.
Materials and methods
2.1. Study site
The field study was conducted between November 2020 – March 2021 on marine clay soils at two agricultural fields cultivated by Brassica oleracea (Botrytis group) in Hem, the Netherlands (Figure 2) and is part of the first intermediate monitoring report on water quality. Both sites were levelled and smoothed in the past for agricultural purposes. Furthermore, the soil is drained from excess water at site A regularly. Further descriptions of the experimental sites are provided in Table 1. The soil type of the sites is considered a calcium carbonate-rich leek-/woudeerdgrond (Pons, 1974). Experimental site A contain a higher percentage of clay particles (< 2μm) than site B. The textural class and structural characteristics of the soil horizons are given in Table 2.
Table 1 Description of experimental sites. *Soil classification according to the Dutch National Key Registry of the Subsurface
Table 2 Soil characteristics of experimental sites. Figure 3 shows a visual representation of the soil profile according to the soil characteristics.
Site A Site B
Location Hem, West-Friesland, Noord-Holland, the Netherlands
Coordinates 52° 40’ 00” N, 5° 10’ 33” E 52° 38’ 52”N, 5° 11’49” E
Elevation [m-ASL] -0.8 -1.20
Slope < 0.25° 0.25° - 1°
Climate Temperate, humid ‘’
Rainfall [mm] Average for January: 67 August: 78 Annual: 788 Temperature [°C] Average for January: 4.4 August: 18.6 Annual: 11.2
Soil* pMn85A pMn55A/pMn85A
Parent material Young marine clay ‘’
Geomorphology Tidal plain and tidal creek ridge Tidal plain
Groundwater levels [m-ASL] max 0.8 - 1.4 min > 1.20
max < 0.4 min 0.8 - 1.2
Location Horizon Depth [cm] Description
Site A Ap 0-20 Dark grey, slightly humus-poor, calcium-rich, Westfrisian clayey marine clay, clay 17.5 - 25 %,
A/C 20-55 Transitional zone with oxidation mottles, silty sandy marine clay
‘’ C 55> Grey, humus-poor, calcium-rich, Westfrisian sandy marine clay, clay 17.5 - 25 %
Site B Ap 0-25 Dark grey, slightly humus-poor, calcium-rich, Westfrisian sandy marine clay, clay 12.5 - 17.5 %,
A/C 25-30 Transitional zone with oxidation mottles silty sandy marine clay
‘’ C 30 > Blue grey, humuspoor, calciumrich, Westfrisian sandy marine clay, clay 12.5
7
2.2. Experimental design
Two experimental sites were established in Hem, the Netherlands (Figure 2A). Physical, chemical and biological measurements were executed prior to the experiment to establish a starting point of the values. One non-inversion tillage (NT) plot of 0.4 x 0.6 km and one inversion tillage (T) plot of 0.15 x 0.7 km are established at Site A (Figure 2B). Additionally, five plots of approximately 0.04 x 0.3 km are created at site B with treatments: NT (n=3) and T (n=2) (Figure 2C). The two sites are two kilometres distant. Figure 2 shows pig (n=2), goat (n=4) and horse (n=1) manure alternately arranged between NT and T plots. All plots were shallowly tilled by milling between 3-10 cm, depth depending on soil moisture. Soil and crop management for this growing season showed minor differences between the two studied sites (Table 3).
Table 3 Overview of soil management for the studied sites 2020/2021
Site A Site B
March 2020 Manure (goat) application before plantation (± 25% of total fertilization, depending on fertilization values)
Inversion and preparation tillage (11 cm) mixing weeds in soil and loosen top soil and subsoil lifting Plantation spring nursery cauliflower (2x a week until the end of July)
Artificial fertilizer injection (±75% of total fertilization)
Harvest autumn cauliflower 2019
Manure (goat, pig & horse) application before plantation (±50% of total fertilization, depending on fertilization values)
Preparation tillage (3-5cm) mixing weeds in soil and loosen top soil
Plantation spring nursery cauliflower (every week until the end of July)
Artificial fertilizer addition (±50% of total fertilization)
Harvest autumn cauliflower 2019
April Plantation spring
Harvest autumn
Plantation spring Harvest autumn
May Plantation spring
Harvest autumn
Plantation spring Harvest autumn
June Plantation spring
Harvest spring cauliflower (every week until the end of August)
Plantation spring
Harvest spring cauliflower (every week until the end of August)
July Plantation spring
Harvest spring
Plantation spring Harvest spring August Plantation autumn nursery cauliflower (every week
until the end of September) Harvest spring
Plantation autumn nursery cauliflower (every week until the end of September)
Harvest spring
September Plantation autumn
Harvest spring Tillage
Sowing cover crop Japanese oats (Avena strigosa)
Plantation autumn Harvest spring Tillage
Sowing cover crop Japanese oats (Avena strigosa)
October Harvest autumn cauliflower 2020 Harvest autumn cauliflower 2020
November Harvest autumn cauliflower 2020 Harvest autumn cauliflower 2020
December Harvest autumn cauliflower 2020 Harvest autumn cauliflower 2020
January Harvest autumn cauliflower 2020 Harvest autumn cauliflower 2020
February Harvest autumn cauliflower 2020 Harvest autumn cauliflower 2020
March 2021 Harvest autumn cauliflower 2020 Harvest autumn cauliflower 2020
General comments
- Drainage canal
- Subsoil lifting (35-40 cm) to open up the subsoil to improve drainage and aeration - Round-up for weeds when soil moisture is too
high for shallow tillage
- Autumn cauliflower survived winter ’19-’20 therefore almost a year round harvest
- Harmonisator magnetic machine to loosen compact mud layer by locally altering the magnetic field
- Weed covering with mud (shallow tillage) between cauliflowers after rain
- Few to no irrigation (allow horizontal root growth)
- (Minimal) biological herbicide use - Compost used in the past
8
2.3. Sampling and analysis
2.3.1. Soil solution, surface and rainwater
The soil water solution was sampled in November 2020, February and March 2021 with an auger at three depths: top layer (Ap horizon), 25 cm (A/C transition) and 50 cm (C horizon). The samples were retrieved from A3, A8, B1 and B6 at site A and in the middle of each field (i.e. A3, B3, C3 and D3) at site B (Figure 2). Subsequently, the soil samples were stored in plastic zip-lock bags and immediately placed in a fridge until analysation. In addition, porous ceramic cups were installed at similar depths in the beginning of February and March, next to the rainwater sample locations. The cups were placed for validation of the soil samples. However, precipitation shortage resulted in a soil that was too dry for vacuum and together with limited time, validation was not possible.
Surface water was retrieved at three locations in surrounding ditches and from puddles that formed within tractor pathways within the plots (Figure 2). Rainwater samplers were placed at both Site A and Site B to collect precipitation (Figure 2). A bird scarecrow ring was attached to each sampler to avoid contamination by bird bodily fluids. The samplers were placed on the border of the plots at a height of 1.5 m above ground level. Rainwater was collected between 3 – 15 December (12 days), 7 February – 1 March ( 22 days) and 1 – 18 March (17 days). The amount of collected precipitation was compared to precipitation data from nearby weather stations (KNMI 2021) for validation. The stations were located in Hoogkarspel, Hoorn and Enkhuizen. The absence of rainwater data from the period between 15 December – 7 February (54 days), resulted in the interpolation of chloride concentrations of the sampled rainwater and daily precipitation data of the three weather stations.
2.3.2. Lab analysis of samples
Soil moisture content was determined by firstly subsampling five grams of soil per sample, drying the subsamples at 105 °C for 24 hours and leaving the dried soil in a desiccator for at least one hour. Subsequently, the ratio for extraction was determined from the moisture content per sample by calculation. Different dilution ratio’s were tested in order to find a ratio that would not deviate the sampled concentration from the original concentration. Eventually, a ratio with five grams of soil per sample were extracted with ultrapure water according to the calculated scale, 1:10. The diluted soil was shaken for a brief moment until all soil dissolved and centrifuged for 15 minutes on 10.000 g. Thereafter the centrifuged soil solutions, rainwater and surface water samples were filtered with a 0.45 μm membrane filter. All filtered samples were analysed for ion concentrations of NO3−, NH4+ and Cl− with the Skalar autoanalyzer (AA). Chloride was measured for the calculation of soil water element fluxes.
9
Figure 2 Field study map showing treatments and with A) Location of Hem, the Netherlands B) Experimental design of 2-plot field and sample locations
(Site A) C) Experimental site of 5-plot field and sample locations (Site B).
10
2.3.3. Data analysis
Soil solution samples were collected in November 2020, February and March 2021 (3 sampling events). The data distribution was preliminarily tested to assure validity for parametric tests according to the Shapiro-Wilk test. The test showed a p-value < 0.05 for all variables and therefore the data was considered as non-parametric. The dependent variables are measured at the continuous level, independent variables are related groups and the distribution of the differences between the related groups is symmetrical. According to these assumptions, the effect of non-inversion tillage on solute concentrations in the soil was analysed by comparison of the means with the non-parametric Wilcoxon signed-rank test (p < 0.05). In order to decrease the risk of a type I error (family-wise error) for multiple testing, the Holm correction was carried out. The following variables were tested:
Independent variables
- Tillage method (NT and T) - Location (Site A, Site B ) - Manure type (goat and pig ) - Soil horizon (Ap, A/C and C) Dependent variables
- NO3- concentrations - NH4 concentrations
The first two-sided statistical hypotheses for calculation of mean difference between T and NT can be defined as: H0: N concentration T = N concentration NT
H1: N concentration T ≠ N concentration NT
The second two-sided statistical hypotheses for calculation of mean difference between site A T and site B T can be defined as:
H0: N concentration site A T = N concentration site B T H1: N concentration site A T ≠ N concentration site B T
The third two-sided statistical hypotheses for calculation of mean difference between site A NT and site B NT can be defined as:
H0: N concentration site A NT = N concentration site B NT H1: N concentration site A NT ≠ N concentration site B NT
The first hypotheses was established for determination of the effect of NT on NO3- and NH4+ concentration in soil solution. The second and third hypotheses were formed in order to study possible significant mean differences between the two field sites of this study. Either of the three hypotheses were tested with the Ap, A/C or C and/or location as subset. In addition the complete dataset was tested. The calculations for the complete dataset without soil horizon or location subset are further referred to as total. Visualisation of graphs was carried out with a colour blind-friendly palette. A bluish colour was designated for graphs related to T treatment and reddish for graphs related to NT treatment.
11
3.
Results
3.1. Surface water
The NO3- concentration in surrounding surface water was strongly variable between sampling events (Figure A). Both sites followed approximately the same fluctuations over time with structural higher values for site B than site A. All values were less in comparison with the 2018 average nitrate concentration for surface waters on clay soils in the Netherlands. The highest value was observed in November 2020 at site B. The NH4+ concentrations were stable for the majority of the months at both sites (Figure 4). February showed a decrease for site A, while site B encountered a remarkable drastic increase. The 2018 average NH4+ concentration for surface waters, 27.7 µmol/L, was not exceeded by measurements at site A. Nevertheless, measurements of site B showed values strongly above the NL average concentration for every measured month.
Figure 4 Concentration of NO3- and NH4+ in surface water at site A and site B for November 2020, February and March 2021. Average
concentrations of NO3- and NH4+ in the Netherlands are displayed at 419.3 and 27.7 µmol/L respectively with the red dashed line. NH4+ concentrations [µmol/L]
12
3.2. Soil solution
The mean and median of NO3- concentration were higher for T treatment than for NT treatment. The highest measured NO3- concentration, 804 µmol/L, was observed in the topsoil horizon (Ap) at site A in November 2020 (Table 4). November additionally showed the highest values for the A/C and C horizon with 594 and 432 µmol/L respectively. All three maximum values originate from the same measurement point at site A (location B6). The lowest observed concentration was 32.4 µmol/L in the C horizon at site B, location C3, in March 2021 (Table 4). The Ap and A/C horizon minimum was also found in March 2021 at 43.3 (location B6) and 39.8 µmol/L (location B3) respectively. All values remained under the 806 µmol NO3- / L of the EU Nitrates Directive. In, addition no measurement showed values higher than the 2018 average NO3- concentration for soil solution on clay soils in the Netherlands.
In comparison with the NO3- concentration, the mean and median for the NH4+ concentration were higher for NT treatment. The NH4+ concentration maximum, 40.5 µmol/L, was observed in the A/C horizon at site B (location B6) in February 2021. The maximum in the Ap and C horizon was 20.9 and 25.4 µmol/L respectively. Either were observed in March 2021 at site B (location D3 and C3). The lowest NH4+ concentration, 3.2 µmol/L, was measured in the C horizon at site A (location B6) in November 2020. The minimum NH4+ concentration values in the Ap and A/C horizon were 4.8 and 6.1 µmol/L respectively. The 2018 average NH4+ concentration for soil solution on clay soils in the Netherlands did not exceed by any value.
Table 4 Descriptive statistics on T and NT treatment and the effect on NO3- and NH4+ concentration in soil solution
The paired comparison of means with the Wilcoxon signed-rank test (p < 0.05) showed that the mean of both NO3- andNH4+ concentration in soil solution is significantly affected by T or NT treatment when total values were tested (Table 5, Figure 5). The difference in mean between T and NT treatment was also significant for NO3- (p = 0.00147) and NH4+ (p = 0.006836) concentration in the soil solution of the C horizon. No significant difference was found for soil horizon Ap and A/C (Table A1, Table A2). An overall decrease in solute concentration over time was observed for both T and NT treatments. The mean NO3 -concentration in soil solution fluctuated over time and depth (Figure 5) and was different between sites A and B (Figure 6). While NO3- decreased over time, NH4+ seemed to slightly increase with time for every horizon and both T and NT (Figure 7). However, only the C horizon showed significantly higher NH4+ concentrations for NT treatment.
Table 5 p-values of NO3- and NH4+ concentration in soil solution per location and both locations together
NO3
-N Median Mean Std. Dev. Minimum Maximum
T 36 231.5 266.8 184.3 34.3 804.0 NT 36 161.5 172.7 99.4 32.4 386.0 T 36 8.5 10.5 7.1 3.2 40.5 NH4+ NT 36 10.7 12.1 5.1 5.2 25.4 Location subset El eme n t
site A site B Total
Treatment T- NT T-NT T-NT
NO3- 0.06654 0.001579 0.0003859
NH4 0.01203 0.2462 0.01152
Wilcoxon signed rank test p-values for NO3- and NH4+ at p < 0.05. The test was calculated separately for each site and for the total
13
Figure 5 Total concentration of NO3- in soil solution and for three depths (Ap, A/C and C) divided in sample event months November 2020,
February and March 2021. Average concentration of NO3- in the Netherlands are displayed at 629 µmol/L. P-values were included in
boxplot figure if mean differences were significant.
NO3- concentrations [µmol/L] Total NO3- concentrations [µmol/L] Ap horizon NO3- concentrations [µmol/L] A/C horizon NO3- concentrations [µmol/L] C horizon Wilcoxon, p = 3.9e-04 Wilcoxon, p = 0.00147
14
Figure 7 Concentration of NH4+ in soil solution in total and for three depths (Ap, A/C and C) divided in sample event months November 2020,
February and March 2021. Average concentration of NH4+ in the Netherlands is displayed at 36 µmol/L. P-values were included in boxplot figure
if mean differences were significant.
NH4+ concentrations [µmol/L] Ap horizon NH4+ concentrations [µmol/L] Total NH4+ concentrations [µmol/L] A/C horizon NH4+ concentrations [µmol/L] C horizon Wilcoxon, p = 0.0115 Wilcoxon, p = 0.006836
15
When compared, Site B showed lower mean concentrations than site A for almost every depth and month except for the Ap horizon in March 2021 (Figure 6). The NT treatment at site A showed no significant difference between mean NO3-and NH4+ concentration in comparison with NT treatment at site B. However, site A showed significantly lower NO3- concentrations than site B for T treatment. Furthermore, significantly higher NH4+ concentrations were observed at site A for T treatment than site B.Table 6 p-values NO3- and NH4+ concentration based on comparison of site A against site B
November 2020, showed a similar pattern of NO3- distribution within the soil column for both sites. However, opposite distributions were found in February and March 2021 between site A T and site B T. Site A showed an increase with depth while site B decreased between the Ap and A/C horizon to slightly increase into the C horizon. This opposing pattern was even more clear in March 2021.
Although mean NO3- concentrations of the NT treatment were almost halved between November 2020 and February 2021, distribution within the soil column remained fairly the same. The concentration increased between the Ap and A/C horizon to subsequently decrease between the A/C and C horizon. March 2021 showed that the mean concentration in site A continuously increased with depth while in contrast site B decreased.
Location subset El eme n t Site A-B Treatment T- T NT-NT NO3- 0.00769 0.1415 NH4 0.00193 0.7019
Wilcoxon signed rank test p-values for NO3- and NH4+ at p < 0.05. The test was calculated by comparing the values of site A against
16
Figure 6 Mean concentrations of NO3-in soil solution over soil depths (Ap, A/C and C) for non-inversion (NT) and inversion tillage (T) treatments. Changes with dept are shown for site A, site B and the total (Avg)
mean values. Chronologically ordered from left to right: November 2020, February and March 2021
Mean NO3- concentration [µmol/L]
November 2020 Mean NO3 concentration [µmol/L] February 2021 Mean NO3 concentration [µmol/L] March 2021
So
il ho
rizon
So
il ho
rizon
17
4.
Discussion
This study aimed to create, compile and compare surface water and soil solution analysis results and general water quality information in the Netherlands of two field sites on marine clay soils during late autumn and winter. The short-term effect of non-inversion tillage on solute concentration in surface water and soil solution was measured in November 2020, February 2021 and March 2021. NO3- and NH4+ concentration changes in surface water bodies during late autumn and winter were studied. In addition, NO3- and NH4+ concentration changes with depth for the Ap , A/C and C horizon of the soil during late autumn and winter were assessed. Lastly, differences in concentrations between the two field sites of this study were analysed.
The NO3- concentration in surrounding surface water was strongly variable between the sampled months (Figure 3). Both sites followed approximately the same fluctuations of NO3- and NH4+ concentration over time with structural higher values for site B than site A. All values were less in comparison with the 2018 average nitrate concentration for surface waters on clay soils in the Netherlands. The NH4+ concentrations were stable for the vast majority of the months at both sites (Figure 4) except for a strong increase in February at site B. This increase could have been caused by dissimilatory nitrate reduction to ammonium by fermentative bacteria (Philips et al., 2002). The decrease in NO3- supports this hypothesis as within this transformation, NO3- is dissimilated in to NH4+ (Figure 7). Nevertheless, although nitrification activity may lower in winter some studies suggest elevated NH4+ concentrations in winter are due to high nitrification rates (Cavaliere & Baulch, 2019). Poorly drained ditches with an ice layer can eventually shift into anaerobic conditions and dissimilation of NO3-. Redox conditions along flow paths of water within the soil can influence soil and groundwater nitrate concentrations (Tesoriero et al., 2015). Microbes depend on oxidation and will first oxidize organic carbon or reduced minerals for energy generation. In case of oxygen depletion, nitrate is used as oxidising agent. Warmer winters result in an increase of microbial activity during this period and eventually more mineralisation. February consisted of 2.5 continuous weeks of temperatures below 0 °C and therefore anaerobic circumstances could have been plausible. To conclude, the 2018 average NH4+ concentration for surface waters (27.7 µmol/L) was not exceeded by measurements at site A, and strongly above the NL average concentration for every measured month at site B.
.
The highest NO3- concentration for both surface water and soil solution were observed in November 2020. The crops were just harvested and harvest residues were left on the soil. Dry summer conditions in the month before and mild weather in autumn possibly induced high nitrogen mineralisation rates (RIVM, 2020). The NO3- concentration in soil solution was significantly higher for T. However NH4+ was significantly higher for NT, especially in the C horizon. Cauliflower consists of a short root system between 30-45 cm. Accumulation of solutes in the C horizon could result in percolation to groundwater as the C horizon is almost outside the area of capillary rise.
Clay soils often consist of multiple impermeable layers that cause precipitation surpluses to run-off to surface water in the area. This may cause the high nitrate concentrations of negatively charged ions in surface water of both sites. Impermeable layers of clay soil compaction can be variable within the plots. Heavy machinery involved in agricultural practice could have formed local impermeable layers resulting in a perched water table. After heavy rainfall, the local perched aquifer can rise to levels as high as the root zone. This is likely to match with site B as the grey and bluish soil together with orange oxidation mottles indicate poor water drainage or waterlogging. During oxygen depletion, denitrification will form nitrogen gas from NO3- (Figure 7). This could be an explanation for lower mean NO3- concentrations at site B.
The plots were located in active agricultural field sites owned by two different farmers and were two kilometres distant. The farmers executed certain soil and crop treatments, due to continuation of cauliflower production. Hence, both field sites experienced slightly diverging treatments. Site A was subsoil lifted (40 cm) at both T and NT plots. Although subsoil lifting is not considered as inversion tillage, this practice could possible compact soil below the rootzone (>40 cm). Accumulation of nitrate is
NH4+ NO2
-Organic
N
NO3-Ammonification
Ammonification
Nitrification
Dissimilation
NH4+ NO3-D
eni
tr
ifi
ca
tio
n
N2 (g)Figure 7 Simplified diagram of the nitrogen cycle
Aerobic
18
therefore possible above such compacted soil. This could possibly explain the similar dynamics for T and NT of solute concentration increase with depth at site A. Site B on the contrary, followed opposite distribution of concentration with depth between T and NT plots. Another important matter is the incorporation of straw within the used farmyard manure. Straw establishes a higher C:N ratio in the soil and immobilizes nitrates during winter when solely catch crops or no crops are growing on the soil (Mary et al., 2008). Timing, practice and the fertilising value of manure application can impact nitrate leaching by affecting the amount and distribution of NO3- between rapid-flow and slow-flow drainage domains (Meisinger et al., 2015). The farmyard manure/artificial fertilizer ratio is different at both sites. It is unknown what amount ands what type of straw was used. Variability caused by fertilisation differences could be explained with a control group. Nevertheless, no control group with only artificial fertilization and tillage treatment was established within the timeframe of this study. The control group should consist of equal basic requirements such as similar soil type and cultivation. A field site of this sort was not found at times when this study was conducted.Solute dynamics in either soil water and surface water are influenced by multiple factors. A direct causal relationship between tillage management on surface and soil water solute concentration is therefore merely possible with a broad approach and incorporation of the most important factors. Disputable factors are variations in meteorological, chemical and ecohydrological conditions. Firstly, timing and amount of rainfall in relation to the time of tillage can influence total percolation and the percentage of percolation that is contributed by preferential flow. These two are the main factors affecting nitrate leaching. Soil inversion tillage is only executed immediately after fertilization and before crop plantation. This study only shows winter period measurements between yield and plantation (i.e. November until end of March). Nitrate flux fluctuates between seasons and during an agricultural year. During winter, soil management is put to a halt as there is no crop plantation, harvest or fertilization. The focus of this study was merely on one late autumn and winter season. Late autumn and winter are most representative for NO3- concentration measurements in surface water and soil solution in the Netherlands. One of the reasons is eutrophication in summer which leads to the reduction of NO3- by algae consumption. Due to eutrophication in sensitive waters during summer, NO3- concentrations measured within these warmer periods can be biased. In addition, polders in the vicinity are actively used during summer and can therefore influence solute concentrations.
Mass balance calculations can estimate the amount of NO3- that percolates from the rootzone to the groundwater. Initially, the element flux was supposed to be calculated according to the CMB method by Wood (1999). Soil water drainage can be calculated assuming the chloride is directly originating from precipitation. According to Wood (1999), other mandatory conditions are conservation of chloride in the soil, no alteration of chloride-mass flux over time and no recycling or concentration of chloride within the soil solution. Nevertheless, compliance with these mandatory assumptions was not possible, due to missing rainwater values between the 15th of December and the 7th of February. The measured samples from other dates showed strongly different chloride concentrations and therefore interpolation would not be a valid option. In addition, high chloride values in soil solution and surface water were observed. Concentrations in urban areas and around highways are often above average during winter months possibly due to application of road salt. Nevertheless, seawater intrusion from the coast could also have caused the high concentrations in chloride found in the samples (Geilfus, 2019; De Lange et al., 2014). Although saltwater intrusion mostly occurs at locations further to the west, the field sites have a low-lying position beneath sea level. As both field sites have a tidal plain or creek ridge geomorphology it is likely they consist of paleochannels within the Holocene layer of the soil. These channels can form a spillway for upward seepage, due to permeable sand layers within the compact Holocene layer (de Louw et al., 2013). Therefore vertical hydraulic gradients may move saline groundwater into surface water and, during dry conditions, by capillary rise even into the root zone. (Mulligan et al., 2007). According to the farmer at site B, the calcareous sandy layer deposited by the paleochannels reflect on crop conditions. Due to higher calcium concentration, cauliflower and grass located above such channels appear more vital than surrounding crops. This observation could possibly be an indication of the presence of paleochannels and an explanation of high chloride concentrations in surface water by upward seepage.
Predictions of nitrate loss beneath the root zone to leaching can be simulated with soil-crop models that are developed for nitrogen balance and transformations. The numerical HYDRUS-1D model is easily accessible and can analyse water flow and solute transport in unsaturated and saturated soils (Colombani et al., 2020; De Silva & Tellam, 2011). According to Colombani et al. (2020), the model could successfully provide information about nitrogen balances in terms of water balance and retention in the soil. Similar discontinuously sampling methods as within this study were used for validation of the model together with soil water probe samples. It is highly recommended to analyse element flux within the soil with this model to measure NO3- flux and eventual leaching between non-inversion and inversion tillage treatment. Nevertheless, timesteps between measurements within this study were not consistent as almost three months were not measured between November 2020 and February 2021. Long-term repeated measures with equal timesteps are necessary to control for seasonal and annual weather variability between measurements.
19
5.
Conclusion
In this study, surface water and soil solution were analysed for NO3- and NH4+ concentration during late autumn and winter from inversion tillage (T) and non-inversion tillage (NT) plots. The plots were located at two different field sites, both on a marine clay soil and fertilised with straw-rich manure. Surface water and soil solution maximum NH4+concentration was found in February 2021. The high values could possibly be induced by dissimilation of NO3- due to anoxic conditions. The mean NO3 -concentration in soil solution fluctuated over time and depth and was different between the two field sites. NT treatment showed significantly lower NO3- (p = 3.9e-4) and on the other hand significantly higherNH4+ concentrations in soil solution (p = 0.0115). No significant difference in solute concentration was found for soil horizon Ap or A/C. Nevertheless, the C horizon showed a concentration of NO3- (p = 0.00147) that was significantly lower for NT. Similar to the total result, NH4+ concentration in the C horizon for NT was significantly higher (p = 0.006836) than the concentration in the C horizon of T. Sunny and warm weather conditions resulted in active nitrification in November 2020. In combination with a relatively high input of organic material from crop residues, the highest NO3- concentration in either surface water and soil solution was observed in November 2020 after harvest. All NO3- values were lower than the 2018 average nitrate concentration for surface waters on clay soils in the Netherlands and remained under 806 µmol NO3- / L of the EU Nitrate Directive limit. Recommended predictions of NO3- and NH4+ loss to groundwater or run-off should be calculated with a suitable model such as HYDRUS-1D. Long-term repeated measures with equal timesteps are necessary to control for seasonal and annual weather variability between measurements and could in combination with HYDRUS-1D make successful simulations of the element flux.
20
Acknowledgements
The study was commissioned by The Province of Noord-Holland in the framework of the circular agriculture vision by the Ministry of Agriculture, Nature and Food Quality (ministerie LNV) and executed by a collaboration with Van Hal Larenstein University of Applied Sciences, Ecolana and Amsterdam Green Campus. This thesis was written during the strange times of a global pandemic where restrictions and anxiety turned the world upside down. Therefore, I am grateful for all the support I have received from many people, especially when progress was slowed down by certain limitations.
First, and foremost I would like to express my sincere appreciation to my supervisors Dr. L.H. Cammeraat and Dr. W.E. Morriën for giving me the opportunity to contribute to this interesting study. Their shared knowledge and helpful advise that was given during Future Planet Studies pillared this bachelor thesis.
Second, I would like to extend my deepest gratitude to Léon Feenstra and Rosa Boone for guiding me with the fieldwork and sharing your knowledge on lab analyses. Your patience and constructive feedback provided me with useful new skills and scientific experience.
Third, I would like to extend my sincere thanks to Titus Rombouts for always making time to discuss thesis problems, practice academic skills and inspire me and the other students from the UvA Green Campus Student Lab to stay positive and active throughout the project. The optimism and sincere curiosity you posses have helped me enormously. Finally I would like to express my deepest gratitude to my family and friends who supported me selflessly during the sometimes difficult moments of the project.
21
References
Adeyemo, A. (2010). Effects of tillage and manure application on soil physicochemical properties and yield of maize grown on a degraded intensively tilled alfisol in southwestern Nigeria. 1, 205–216.
Britto, D. T., & Kronzucker, H. J. (2013). Ecological significance and complexity of N-source preference in plants. Annals of Botany, 112(6), 957–963. https://doi.org/10.1093/aob/mct157
Cavaliere, E., & Baulch, H. M. (2019). Winter nitrification in ice-covered lakes. PLOS ONE, 14(11), e0224864. https://doi.org/10.1371/journal.pone.0224864
Colombani, N., Mastrocicco, M., Vincenzi, F., & Castaldelli, G. (2020). Modeling Soil Nitrate Accumulation and Leaching in Conventional and Conservation Agriculture Cropping Systems. Water, 12(6), 1571.
Cui, M., Zeng, L., Qin, W., & Feng, J. (2020). Measures for reducing nitrate leaching in orchards:A review. Environmental Pollution, 263, 114553. https://doi.org/10.1016/j.envpol.2020.114553
De Lange, W. J., Prinsen, G. F., Hoogewoud, J. C., Veldhuizen, A. A., Verkaik, J., Oude Essink, G. H. P., van Walsum, P. E. V., Delsman, J. R., Hunink, J. C., Massop, H. Th. L., & Kroon, T. (2014). An operational, multi-scale, multi-model system for consensus-based, integrated water management and policy analysis: The Netherlands Hydrological Instrument. Environmental Modelling & Software, 59, 98–108. https://doi.org/10.1016/j.envsoft.2014.05.009
de Louw, P. G. B., Eeman, S., Oude Essink, G. H. P., Vermue, E., & Post, V. E. A. (2013). Rainwater lens dynamics and mixing between infiltrating rainwater and upward saline groundwater seepage beneath a tile-drained agricultural field. Journal of Hydrology, 501, 133–145. https://doi.org/10.1016/j.jhydrol.2013.07.026
De Silva, C. & Tellam. (2011, October 1). Simulation of nitrate leaching and groundwater pollution in vegetable cultivation using HYDRUS-1D model.
Derpsch, R. (2003). Conservation Tillage, No-Tillage and Related Technologies. In L. García-Torres, J. Benites, A. Martínez-Vilela, & A. Holgado-Cabrera (Eds.), Conservation Agriculture: Environment, Farmers Experiences, Innovations, Socio-economy, Policy (pp. 181–190). Springer Netherlands. https://doi.org/10.1007/978-94-017-1143-2_23
EU, 1991. (1991, December 12). EUR-Lex—01991L0676-20081211—EN - EUR-Lex. https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A01991L0676-20081211
Geilfus, C.-M. (2019). Chloride in soil: From nutrient to soil pollutant. Environmental and Experimental Botany, 157, 299–309. https://doi.org/10.1016/j.envexpbot.2018.10.035
Grizzetti, B., Bouraoui, F., & Aloe, A. (2007). Spatialised European Nutrient Balance (JRC36653) [EUR - Scientific and Technical Research Reports]. https://publications.jrc.ec.europa.eu/repository/handle/111111111/8344
Hansen, E. M., & Djurhuus, J. (1997). Nitrate leaching as influenced by soil tillage and catch crop. Soil and Tillage Research, 41(3), 203–219. https://doi.org/10.1016/S0167-1987(96)01097-5
Jain, H. K. (2010). Green Revolution: History, Impact and Future. The Green Revolution: Histor.
KNMI - Overzicht van de neerslag en verdamping in Nederland. (2021). Retrieved March 25, 2021, from https://www.knmi.nl/nederland-nu/klimatologie/gegevens/monv
Leitner, S., Dirnböck, T., Kobler, J., & Zechmeister-Boltenstern, S. (2020). Legacy effects of drought on nitrate leaching in a
temperate mixed forest on karst. Journal of Environmental Management, 262, 110338.
https://doi.org/10.1016/j.jenvman.2020.110338
Liu, E., Yan, C., Mei, X., He, W., Bing, S. H., Ding, L., Liu, Q., Liu, S., & Fan, T. (2010). Long-term effect of chemical fertilizer, straw, and manure on soil chemical and biological properties in northwest China. Geoderma, 158(3), 173–180. https://doi.org/10.1016/j.geoderma.2010.04.029
22
Mary, B., Beaudoin, N., Justes, E., & Machet, J. (2008). Calculation of nitrogen mineralization and leaching in fallow soil using a simple dynamic model. European Journal of Soil Science, 50, 549–566. https://doi.org/10.1046/j.1365-2389.1999.00264.xMcLaughlin, A., & Mineau, P. (1995). The impact of agricultural practices on biodiversity. Agriculture, Ecosystems & Environment, 55(3), 201–212. https://doi.org/10.1016/0167-8809(95)00609-V
Meisinger, J. J., Palmer, R. E., & Timlin, D. J. (2015). Effects of tillage practices on drainage and nitrate leaching from winter wheat
in the Northern Atlantic Coastal-Plain USA. Soil and Tillage Research, 151, 18–27.
https://doi.org/10.1016/j.still.2015.02.007
Mulligan, A. E., Evans, R. L., & Lizarralde, D. (2007). The role of paleochannels in groundwater/seawater exchange. Journal of Hydrology, 335(3), 313–329. https://doi.org/10.1016/j.jhydrol.2006.11.025
Oades, J. M. (1984). Soil organic matter and structural stability: Mechanisms and implications for management. Plant and Soil, 76(1), 319–337. https://doi.org/10.1007/BF02205590
Philips, S., Laanbroek, H. J., & Verstraete, W. (2002). Origin, causes and effects of increased nitrite concentrations in aquatic
environments. Reviews in Environmental Science and Biotechnology, 1(2), 115–141.
https://doi.org/10.1023/A:1020892826575
RIVM, 2020. (n.d.). Agricultural practice and water quality in the Netherlands: Status (2012-2014) and trend (1992-2014). 192. Schimel, J. P. (2018). Life in Dry Soils: Effects of Drought on Soil Microbial Communities and Processes. Annual Review of Ecology,
Evolution, and Systematics, 49(1), 409–432. https://doi.org/10.1146/annurev-ecolsys-110617-062614
Stenberg, M., Aronsson, H., Lindén, B., Rydberg, T., & Gustafson, A. (1999). Soil mineral nitrogen and nitrate leaching losses in soil tillage systems combined with a catch crop. Soil and Tillage Research, 50(2), 115–125. https://doi.org/10.1016/S0167-1987(98)00197-4
Tesoriero, A. J., Terziotti, S., & Abrams, D. B. (2015). Predicting Redox Conditions in Groundwater at a Regional Scale. Environmental Science & Technology, 49(16), 9657–9664. https://doi.org/10.1021/acs.est.5b01869
The R Foundation for Statistical Computing (2021). R version 4.0.4 (2021-02-15)—"Lost Library Book".
Wood, W. W. (1999). Use and misuse of the chloride-mass balance method in estimating ground water recharge. Ground Water, 37(1), 2–5.
Zhang, W., Ricketts, T. H., Kremen, C., Carney, K., & Swinton, S. M. (2007). Ecosystem services and dis-services to agriculture. Ecological Economics, 64(2), 253–260. https://doi.org/10.1016/j.ecolecon.2007.02.024
23
Appendices
Appendix A – p-value tables
Table A1 p-values NO3- concentration per horizon and per locationTable A2 p-values NH4+ concentration per horizon and per location
Location subset
Ho
ri
zo
n
site A site B site A - B site A-B Total
Treatment T- NT T-NT T -T NT-NT T-NT Ap 1 0.0625 1 0.6875 0.1099 A/C 0.3125 0.2188 0.3125 0.09375 0.07715 C 0.0625 0.03125 0.03125 0.03125 0.001465 Location subset Ho ri zo n
site A site B site A - B site A-B Total
Treatment T- NT T-NT T -T NT-NT T-NT
Ap 0.09375 0.4375 0.09375 0.09375 0.07715
A/C 1 0.8438 1 0.84375 0.9697
C 0.03125 0.2188 0.15625 0.84375 0.006836
Wilcoxon signed rank test p-values NO3- per horizon at p < 0.05. The test was calculated separately for each site (T-NT), between site A
and B (for T or NT treatment) and for total values of site A and B together (T-NT).
Wilcoxon signed rank test p-values NH4+ per horizon at p < 0.05. The test was calculated separately for each site (T-NT), between site
24
Appendix B - R script analyses and visualisation
# This code is to replicate the analyses and visualisation from my bachelor thesis:
# "Effect of non-inversion tillage and straw-rich manure on water quality" # Code developed by Natalia Dreijer
# Future Planet Studies, UvA, April 2021
install_github("kassambara/easyGgplot2") install.packages("PairedData") install.packages("devtools") library(ggplot2) library(ggrepel) library(ggthemes) library(scales) library(PairedData) library(devtools) library(easyGgplot2) library(tidyr) #### LOAD DATA ####
data <- read.table(file="clipboard", sep = "\t", header = TRUE)
#### Parametric test ####
logNO3 <- log(data$NO3)
logNH4 <- log(data$NH4)
logCl <- log(data$Cl)
para_test_NO3 <- shapiro.test(logNO3)
para_test_NH4 <- shapiro.test(logNH4)
para_test_Cl <- shapiro.test(logCl)
# Holm-Bonferonni correction (prevents Type I errors) # p.adjust(p, method = "holm")
################################## SUBSETTING ################################ #### SURFACE WATER ####
surf_water <- subset(data, sample== "surfacewater")
ggplot(surf_water[order(surf_water$date),], aes(fill=location, x=date,y=NH4)) + geom_bar(position="dodge", stat = "identity",color="black") +
geom_hline(yintercept=28, linetype="dashed", color = "red") +
labs(x = "Month",y = "NH4+ [mol/L]",
title = "NH4+ concentration surface water")+
scale_fill_discrete(name = "Treatment")+
geom_bar(stat="identity", position="dodge", color="black")+
annotate("text", x = "mrt-21", y = 29,
label = "NL average 2018", size=3, vjust = -0.5)+
scale_x_discrete(limits=c("nov-20", "feb-21", "mrt-21"))
#### T& NT total (Site A and Site B ##### ####### ALL
all <- subset(data, sample=="soil water")
outliers <- boxplot(all$NO3)$out
ggplot(all, aes(x=date, y=NH4, fill= Ttreatment)) + geom_boxplot(position=position_dodge2(reverse=TRUE))+
labs(x = "Month", y = "NH4+ [mol/L]",
title = "NH4+ concentration soil moisture", subtitle = "Total")+
scale_fill_discrete(name = "Treatment")+
scale_x_discrete(limits=c("nov-20", "feb-21", "mrt-21")) +
annotate("text", x = "feb-21", y = 36,
label = "NL average 2018", size=3, vjust = -0.5,hjust=0.2)+
#facet_wrap(~Ttreatment, scales= "free_y")+
#facet_grid(~factor(Ttreatment,'T','NT'))+
guides(fill=guide_legend(reverse=TRUE))+
geom_hline(yintercept=36, linetype="dashed", color = "red")
25
A_all <- subset(data, sample=="soil water" & location == "Site A")
ggplot(A_all, aes(x=date, y=NH4, fill= Ttreatment)) + geom_boxplot(position=position_dodge2(reverse = TRUE))+
guides(fill = guide_legend(reverse = TRUE))+
labs(x = "Month", y = "NH4+ [mol/L]",
title = "NH4+ concentration soil moisture", subtitle= "Site A")+
scale_fill_discrete(name = "Treatment")+
scale_x_discrete(limits=c("nov-20", "feb-21", "mrt-21")) +
annotate("text", x = "feb-21", y = 36,
label = "NL average 2018", size=3, vjust = -0.5,hjust=0.2)+
geom_hline(yintercept=36, linetype="dashed", color = "red")
######### Site B
B_all <- subset(data, sample=="soil water" & location == "Site B")
ggplot(B_all, aes(x=date, y=NO3, fill= Ttreatment)) + geom_boxplot(position=position_dodge2(reverse = TRUE))+
guides(fill = guide_legend(reverse = TRUE))+
labs(x = "Month", y = "NO3- [mol/L]",
title = "NO3- concentration soil moisture", subtitle= "Site B")+
scale_fill_discrete(name = "Treatment")+
scale_x_discrete(limits=c("nov-20", "feb-21", "mrt-21")) +
annotate("text", x = "feb-21", y = 600,
label = "NL average 2018", size=3, vjust = -0.5,hjust=0.2)+
geom_hline(yintercept=629, linetype="dashed", color = "red")
############ NT & T per horizon (& Site A or/and Site B) #########
TOP <- subset(all, horizon == "TOP")
ggplot(TOP, aes(x=date, y=NH4, fill= Ttreatment)) + geom_boxplot(position=position_dodge2(reverse = TRUE))+
guides(fill = guide_legend(reverse = TRUE))+
labs(x = "Month", y = "NH4- [mol/L]",
title = "NH4- concentration soil moisture", subtitle= "A horizon")+
scale_fill_discrete(name = "Treatment")+
scale_x_discrete(limits=c("nov-20", "feb-21", "mrt-21")) +
annotate("text", x = "feb-21", y = 36,
label = "NL average 2018", size=3, vjust = -0.5,hjust=0.2)+
geom_hline(yintercept=36, linetype="dashed", color = "red")
############## WILCOXON MEAN RANK TESTS ##########################
NT = data[data$Ttreatment == "NT", ] tT = data[data$Ttreatment == "T", ]
res_NO3 <- wilcox.test(tT$NO3, jitter(NT$NO3), paired = TRUE)
res_NH4 <- wilcox.test(tT$NH4, NT$NH4, paired = TRUE)
res_NO3 res_NH4
###################################################################### ### total for each site -- Within groups
# Site A
res_NH4_A <- wilcox.test(tT_A$NH4, NT_A$NH4, paired = TRUE)
res_NO3_A <- wilcox.test(tT_A$NO3, NT_A$NO3, paired = TRUE)
# Site B
res_NH4_B <- wilcox.test(tT_B$NH4, NT_B$NH4, paired = TRUE)
res_NO3_B <- wilcox.test(tT_B$NO3, NT_B$NO3, paired = TRUE)
26
##### Location A - Horizon -- Differences within site A
NT_A_C = NT_A[NT_A$horizon == "C", ] NT_A_AC = NT_A[NT_A$horizon == "A/C", ] NT_A_TOP = NT_A[NT_A$horizon == "TOP", ] tT_A_C = tT_A[tT_A$horizon == "C", ] tT_A_AC = tT_A[tT_A$horizon == "A/C", ] tT_A_TOP = tT_A[tT_A$horizon == "TOP", ]
res_NH4_A_C = wilcox.test(tT_A_C$NH4, NT_A_C$NH4, paired = TRUE)
res_NO3_A_C = wilcox.test(tT_A_C$NO3, NT_A_C$NO3, paired = TRUE)
res_NH4_A_AC = wilcox.test(tT_A_AC$NH4, NT_A_AC$NH4, paired = TRUE)
res_NO3_A_AC = wilcox.test(tT_A_AC$NO3, NT_A_AC$NO3, paired = TRUE)
res_NH4_A_TOP = wilcox.test(tT_A_TOP$NH4, NT_A_TOP$NH4, paired = TRUE)
res_NO3_A_TOP = wilcox.test(tT_A_TOP$NO3, NT_A_TOP$NO3, paired = TRUE)
##### Site B - Horizon -- Differences within site B
NT_B_C = NT_B[NT_B$horizon == "C", ] NT_B_AC = NT_B[NT_B$horizon == "A/C", ] NT_B_TOP = NT_B[NT_B$horizon == "TOP", ] tT_B_C = tT_B[tT_B$horizon == "C", ] tT_B_AC = tT_B[tT_B$horizon == "A/C", ] tT_B_TOP = tT_B[tT_B$horizon == "TOP", ]
res_NH4_B_C = wilcox.test(tT_B_C$NH4, NT_B_C$NH4, paired = TRUE)
res_NO3_B_C = wilcox.test(tT_B_C$NO3, NT_B_C$NO3, paired = TRUE)
res_NH4_B_AC = wilcox.test(tT_B_AC$NH4, NT_B_AC$NH4, paired = TRUE)
res_NO3_B_AC = wilcox.test(tT_B_AC$NO3, NT_B_AC$NO3, paired = TRUE)
res_NH4_B_TOP = wilcox.test(tT_B_TOP$NH4, NT_B_TOP$NH4, paired = TRUE)
res_NO3_B_TOP = wilcox.test(tT_B_TOP$NO3, NT_B_TOP$NO3, paired = TRUE)
####################################################################### ### Comparison between sites (e.g. Site A T with Site B T)
# -- Baseline for further research
NT_A = NT[NT$location == "Site A", ] NT_B = NT[NT$location == "Site B", ] tT_A = tT[tT$location == "Site A", ] tT_B = tT[tT$location == "Site B", ]
# Difference between site A and B
res_tT_NH4_A_B <- wilcox.test(tT_A$NH4, tT_B$NH4, paired = TRUE)
res_tT_NO3_A_B <- wilcox.test(tT_A$NO3, jitter(tT_B$NO3), paired = TRUE)
res_NT_NH4_A_B <- wilcox.test(NT_A$NH4, jitter(NT_B$NH4), paired = TRUE)
res_NT_NO3_A_B <- wilcox.test(NT_A$NO3, NT_B$NO3, paired = TRUE)
#### Differences between site A and B per horizon ### Tilled
res_tT_NH4_C = wilcox.test(tT_A_C$NH4, tT_B_C$NH4, paired = TRUE)
res_tT_NO3_C = wilcox.test(tT_A_C$NO3, tT_B_C$NO3, paired = TRUE)
res_tT_NH4_AC = wilcox.test(tT_A_AC$NH4, tT_B_AC$NH4, paired = TRUE)
res_tT_NO3_AC = wilcox.test(tT_A_AC$NO3, jitter(tT_B_AC$NO3), paired = TRUE)
res_tT_NH4_TOP = wilcox.test(tT_A_TOP$NH4, tT_B_TOP$NH4, paired = TRUE)
res_tT_NO3_TOP = wilcox.test(tT_A_TOP$NO3, tT_B_TOP$NO3, paired = TRUE)
### Not Tilled
res_NT_NH4_C = wilcox.test(NT_A_C$NH4, NT_B_C$NH4, paired = TRUE)
res_NT_NO3_C = wilcox.test(NT_A_C$NO3, NT_B_C$NO3, paired = TRUE)
res_NT_NH4_AC = wilcox.test(NT_A_AC$NH4, NT_B_AC$NH4, paired = TRUE)
res_NT_NO3_AC = wilcox.test(NT_A_AC$NO3, NT_B_AC$NO3, paired = TRUE)
res_NT_NH4_TOP = wilcox.test(NT_A_TOP$NH4, NT_B_TOP$NH4, paired = TRUE)
27
############################################################################ #### JUST Horizon -- Comparison per horizon
NT_C = NT[NT$horizon == "C", ] NT_AC = NT[NT$horizon == "A/C", ] NT_TOP = NT[NT$horizon == "TOP", ] tT_C = tT[tT$horizon == "C", ] tT_AC = tT[tT$horizon == "A/C", ] tT_TOP = tT[tT$horizon == "TOP", ]
res_NH4_C = wilcox.test(tT_C$NH4, NT_C$NH4, paired = TRUE)
res_NO3_C = wilcox.test(tT_C$NO3, NT_C$NO3, paired = TRUE)
res_NH4_AC = wilcox.test(tT_AC$NH4, NT_AC$NH4, paired = TRUE)
res_NO3_AC = wilcox.test(tT_AC$NO3, NT_AC$NO3, paired = TRUE)
res_NH4_TOP = wilcox.test(tT_TOP$NH4, NT_TOP$NH4, paired = TRUE)
res_NO3_TOP = wilcox.test(tT_TOP$NO3, NT_TOP$NO3, paired = TRUE)
############################################################################## #### JUST Mtreatment -- Comparison between manure treatment
NT_Goat = NT[NT$Mtreatment == "Goat", ] NT_Pig = NT[NT$Mtreatment == "Pig" , ] tT_Goat = tT[tT$Mtreatment == "Goat", ] tT_Pig = tT[tT$Mtreatment == "Pig" , ]
res_NH4_Pig = wilcox.test(tT_Pig$NH4, NT_Pig$NH4, paired = TRUE)
res_NO3_Pig = wilcox.test(tT_Pig$NO3, NT_Pig$NO3, paired = TRUE) res_NH4_Goat = wilcox.test(tT_Goat$NH4, NT_Goat$NH4, paired = TRUE)
