Effects of Minimal Tillage Management on Low-flying Insect Abundance and Meadow Bird Nests and Behaviour in Dutch Cauliflower Fields

(1)

1

Effects of Minimal Tillage Management

on Low-flying Insect Abundance and

Meadow Bird Nests and Behaviour in

Dutch Cauliflower Fields

Douwe de Maijer

30/05/2021, Amsterdam Thesis supervisor: Elly Morriën Daily Guidance: Anne Uilhoorn

Abstract

Agricultural intensification is globally leading to soil degradation and a decline of soil ecosystem services. Intensive tillage-based agricultural production systems are associated with disturbance of the soil fauna and deteriorated physical and chemical soil properties. Several insectivorous meadow bird species are dependent on the soil fauna. Adults mostly feed on worms, and chicks are highly

dependent on insects and insect larvae. As the stability of the soil ecosystem is jeopardized by

intensive tillage, no-tillage methods have the potential to improve it, and increase insect populations in and above the soil; subsequently increasing the food supply for meadow bird chicks. Therefore, this research aimed to examine the effects of no-tillage manure application on insect abundance, meadow bird behaviour and meadow bird nesting. Furthermore, the effects of different types of manure were inspected. Insects were sampled using malaise traps on two farms, both containing tilled (T) and untilled (NT) fields. Birds were observed on both farms for one hour. NT fields had significantly higher insect biomass than T fields, directly after manure application. This effect disappeared in the weeks after fertilization. Meadow birds appeared to prefer NT fields for foraging. This is potentially explained by higher food availability. The higher biomass per insect as found on NT fields is favourable for meadow birds, as the feeding yield per unit of time is higher. Effects of different manure types on insect biomass and individuals were insignificant. Insect abundance did increase as the season developed. This increase was partially caused by collembola, which appeared in the malaise traps during the last measurements. There is potentially a relationship between collembola capture in malaise traps and precipitation. In general, it is concluded that non-incorporation of manure into the soil is related to increased insect biomass. Therefore this type of manure application is also favourable for meadow birds and their chicks.

Keywords: Agriculture, no-tillage, manure, low-flying insects, meadow birds, meadow bird chicks. Word count (abstract, references, figures and tables excluded): 4728 Words.

(2)

2

Introduction

Agricultural intensification and agricultural land use are increasing as a consequence of rising global food demand (FAO, 2017). Intensive farming systems often cause structural soil changes by intensive tillage practices, such as ploughing and harrowing. These changes predominantly encompass reduced soil cohesiveness and organic matter content, along with increased soil compaction. All of these degradation processes are known to reduce plant growth (Corsi & Muminjanov, 2015). Besides reduced growth, such soil degradation processes negatively affect soil biodiversity and the metabolic capacity of its ecosystem, which in turn aggravate the soil degradation (Bini, 2009). The soil

ecosystem supports several insectivorous meadow bird species; consequently, the abundance of ground-dwelling insects and larvae partially determine the survival rate of meadow bird chicks (Schekkerman et al., 2005; Schekkerman & Beintema, 2007). In this research, an alternative farming system that potentially improves the soil ecosystem – and thereby the capacity to support meadow bird chicks – is explored. The explanation of the specific research is preceded by an introduction to the general problem of tillage in agriculture, and its effects on soil and ecosystem properties.

Tillage, No-Tillage & Conservation Tillage

Soil degradation in agricultural fields is associated with intensive tillage-based production systems. Tillage is an important aspect of intensive agricultural systems. In these systems, tillage encompasses mechanical lifting and/or mixing of the soil in order to homogenize, fertilize and prepare it for sowing. Intensive tillage methods, however, often result in soil degradation, reduced water quality and

quantity, loss of terrain and reduced above- and belowground biodiversity. Therefore, minimal implementation of tillage in agricultural systems is of increasing importance for food production (Friedrich et al., 2017). Such tillage systems are often referred to as zero- or no-tillage. However, in many no-tillage systems, minor rotary tillage practices are applied during or before seeding to ensure germination, as in the research of Wang et al. (2016). For that reason, conservation tillage is often a more appropriate term when describing minimal tillage systems (Derpsch & Benites, 2003). When using the term no-tillage in agricultural research, planting and seeding equipment and minimal tillage applications should be mentioned explicitly, in order to describe the no-tillage system transparently (Derpsch et al., 2014). As most literature – for example that of Wang et al. (2020) – that is used in this research refers to no-tillage, when actually minimal tillage is implemented, the term no-tillage is often used in this research as well. Specifications on the tillage implementation in the field plots of this research are mentioned in the Materials and Methods section.

New Farming Approaches & the Obstructions of Dutch Law

No-tillage (i.e. minimal tillage) is one of the fundamentals of an upcoming agricultural approach known as Conservation Agriculture (CA). CA farms are spreading exponentially since the 1990’s and scientific research into CA has increased as well (Friedrich et al., 2017). The FAO and the

international scientific community state that CA can be implemented in all agricultural landscapes by implementing (i) minimal soil disturbance, (ii) permanent soil organic covering by the use of cover crops or by leaving crop residues after harvesting (or both) and by (iii) crop rotations to diversify crop species (Corsi & Muminjanov, 2015; Friedrich et al., 2017).

Despite the FAO’s (2017) recommendation to increase the implementation of CA, it is obligated by law to incorporate animal manure into agricultural soils in the Netherlands. The overall purpose of this obligation is to reduce ammonia (NH3) emissions into the atmosphere. The incorporation has to be

executed by either injecting manure, or mixing it with the soil directly after application (RVO, 2020); therefore, no-tillage manure application is forbidden. Such tillage practices potentially have negative effects on SOM, water holding capacities and soil structure (Cerdà et al., 2020; Corsi & Muminjanov, 2015; Friedrich et al., 2017). Another effect of intensive tillage is the potential destruction of bird nests located in tilled agricultural fields; as heavy machinery destroy the bird nests (McLaughlin & Mineau, 1995). Additionally, Oosterveld (2006) demonstrated that applying liquid manure, or injecting liquid manure in trenches in agricultural fields has a negative effect on worm populations, which are an

(4)

4 important source of feed for meadow birds. Conversely, the application of solid manure directly onto the soil surface, has a slight positive effect on worm populations (Oosterveld, 2006). Therefore, this method of manure application potentially has an indirect positive effect on meadow bird populations.

No-tillage, Insects & Meadow Birds

In the Netherlands, meadow bird populations such as the Black-tailed Godwit are strongly declining, despite the implementation of several conservational laws (Schekkerman & Beintema, 2007; Sovon, 2020). This continuation of the meadow bird population decline, is mainly because those laws mostly targeted the survival of clutches; they were mostly aimed to reduce mowing and trampling losses in grasslands ( Schekkerman & Beintema, 2007). However, chick survival rates are determinative for the population dynamics as well (Schekkerman & Beintema, 2007). The food intake of meadow bird chicks largely consists of ground-dwelling insects and low-flying insects and their larvae (mainly invertebrates and arthropods) present in the topsoil; therefore, their abundance is an important

contributor to the survival rate of chicks ( Schekkerman et al., 2005; Schekkerman & Beintema, 2007). Lahr & Van Der Pol (2007) outlined that manure injection methods may disturb the soil fauna,

including ground dwelling insects and low-flying insects and their larvae. On top of that, Oosterveld (2006) stated that for proper meadow bird management the only manure application should be solid manure. The type of manure (for example: pig or goat manure) that is used, additionally affects insect survival rates (Khan et al., 2012). Therefore, no-tillage solid manure application and the type of manure used potentially affect insect abundance, which can subsequently affect the survival rate of meadow bird chicks.

For these reasons, this research aims to examine the effects of no-tillage solid manure application on low-flying insect abundance (i). Additionally, the research is aimed to identify whether there is a relation between no-tillage and meadow bird nests and meadow bird behaviour, with emphasis on foraging preference (ii). Furthermore, investigating the effects of different types of manure (goat, pig and horse) on insect abundance (iii) is the final aim of this research.

Expected Results

As explained above, manure incorporation methods disturb the soil fauna (Lahr & Van Der Pol, 2007). As the no-tillage (NT) treatments are considered to be less disturbing for the soil than the tillage (T) treatments, it is expected that insect quantities on “NT fields” are significantly higher than on “T fields”. Lahr & Van Der Pol (2007) indicated that the abundance of manure feeding insect larvae and beetles increases in the first 5 days after excretion; after 10-30 days their presence decreases quickly. Therefore, it is hypothesized that insect activity is highest after the first manure application, after which it will decrease over time. However, insect abundance is highly dependent on previous and current weather conditions, such as wind and rainfall (Williams, 1961); therefore, the insect activity may vary strongly.

As insects and their larvae serve as a food source for meadow bird chicks (H Schekkerman et al., 2005), it is expected that the presence of bird nests on NT fields is significantly higher. However, the sampling site may be too small to draw representative conclusions.

The difference in insect abundance between the different manures, is dependent on manure quality (Lahr & Van Der Pol, 2007). This is partially influenced by medication intake of the animals

providing the manure. The specifics on these medication intakes of the different animals are unknown; therefore, no concrete hypothesis is formulated on the effect of manure type on insect abundance.

Materials and Methods

Study Sites

In order to examine the effects of no-tillage management and different manure types on low-flying insects and meadow birds, two cauliflower farms (from now on referred to as farm Slagter and farm

(5)

5 Reus, which are the farmer’s last names) – located in Hem, Noord-Holland, the Netherlands – were examined. In general, each farm was divided in tilled (T) and untilled (NT) areas.

Farm Slagter

Farm Slagter (figure 1) is located at 52° 39’5.59’’ N, 5° 12’3.201’’ E. Two fields formed the research area: A (± 9 ha) and B (± 3 ha). Both fields were treated with a disc harrow (35-40 cm) in advance of fertilizing. Manure used on these fields was goat manure mixed with straw. Manure was applied on 20 April. On field A the manure is not incorporated in the soil after application (NT); on field B the manure is directly incorporated into the soil after application (T). Nutrient shortages are compensated with artificial fertilizer (mostly urea and phosphate). Approximately 75% of the fertilizer is artificial and 25% is solid manure.. The soil was milled (11-12 cm) before planting. During this research crops were planted in field B on 26 April 2021. At that point there were no crops yet on field A.

Figure 1: The research area farm Slagter, including trap locations and bird spotting sites. Ento Sphinx and NHBS represent two trap types used for insect sampling. On field A, manure is not incorporated after application (NT). On field B manure is incorporated in the soil after application (T).

Farm Reus

Farm Reus (figure 2) is located at 52° 38’51.56’’ N, 5°12’03.47’’ E. The research area consists of five strips (A-E) of ± 22 m by ± 200 m. Each manure type was mixed with straw. Manure was applied on 4 April. Before manure application, the T fields were milled. The NT fields were untreated before manure application. During this research, the crops were not yet planted.

(6)

6 Figure 2: The research area farm Reus, including trap locations and bird spotting site. Ento Sphinx represents trap type used for insect sampling. On strip B and C, manure is not

incorporated after application (NT). On strip A, D and E manure is incorporated in the soil directly after application (T). Distance between bird spotting location and research area is approximately 200 meters.

Nutrient shortages are compensated with artificial fertilizer (the type is yet unknown). Approximately 50% of fertilization is done with animal manure, the other 50% with artificial fertilizer.

Sampling Design and Measurements

Insect Sampling

In order to examine the difference in insect abundance between NT and T fields, malaise traps (figure 3) of the brand Ento Sphinx (farm Reus and Slagter) and NHBS (farm Slagter) were used as sampling device. A malaise trap is a scientifically recognized instrument for sampling and collection of diverse insect groups, but especially for low-flying insects (Sheikh et al., 2016). On both farms, traps were assigned to specific locations to obtain representative results for each field or strip (figure 1 & figure 2). On farm Reus, the samples were additionally used to compare the effects of different manure types on insect abundance.

Figure 3: Ento Sphinx Malaise trap

(left) and NHBS malaise trap (right). Low-flying insects are intercepted in the fabric barrier in the middle. As insects are expected to move

upwards when trying to escape, they get trapped in the collecting jar on the upward end of the tent (Sheikh et al., 2016). Images retrieved from Veldshop.nl (n.d.).

The captured insects were collected in a bottle with 70% ethanol solution. These bottles were emptied in cardboard boxes and air dried for 2-3 days. After drying, the insects were sieved for the fractions <4 mm, 4-10 mm and >10 mm. These fractions were weighed, counted and identified for the dominant groups.

(7)

7

Timeframe

Above mentioned sampling was conducted in different timeframes on both farms. On farm Slagter, a before and an after treatment measurement was conducted (table 1). These measurements were conducted before and after manure application in order to obtain a baseline measurement and an after treatment measurement.

Table 1: Sampling dates farm Slagter, 2021. Each column represents a time frame in which insects

were collected. “Before” represents the baseline measurement. “After” is the first measurement after manure application. Manure was applied on 20 April.

Time frame Before After

Date 29/03-01/04 (± 72 hr) 23/04-26/04 (± 72 hr)

On farm Reus, five measurements were executed (table 2). The first measurement was the baseline measurement. After that, four more measurements we executed in order to obtain after treatment data, and to obtain seasonal development data.

Table 2: Insect Sampling dates farm Reus, 2021. Each column represents a time frame in which

insects were collected. Date row describes the sampling time. Traps were removed at T0 and replaced at T1, after which the traps were emptied weekly, without removal. T0 represents the baseline

measurement. T1 is the first measurement after manure application. Manure was applied on 4 April.

Time frame T0 T1 T2 T3 T4 Date 29/03-01/04 (±72 hr) 09/04-12/04 (± 72 hr) 12/04-19/04 (± 168 hr) 19/04-26/04 (± 168 hr) 26/04-04/05 (± 192 hr)

Bird Nest and Behaviour Sampling

Bird nests were counted in all fields during the measuring time (29/03/2021-04/05/2021). Bird behaviour was analysed as followed: From a minimally disturbing distance birds were observed for one hour. This hour was divided into timeframes of five minutes. Each five minutes bird species and behaviour was noted. Behaviour types that were noted were: foraging, brooding and courtship and territorial behaviour. On farm Slagter the observations were conducted separately on each field from different locations (figure 1) on the same date (4 May 2021). On farm Reus these observations were conducted on 15 April. All strips were observed at the same time from the same location (figure 2), as the complete research area is easily surveyable. On this farm, the location (strip) of the birds was noted as well in each timeframe. By noting both location and behaviour, preferred foraging areas on farm Reus could be identified. Identifying the preferred foraging strips allowed to examine whether there was a difference in foraging preference between T and NT strips.

Statistical Analysis

For the statistical analysis the obtained data (Appendix 1) were processed in Rstudio (script: Appendix 2.2). For insect abundance a repeated measures ANOVA was executed for both biomass (Appendix 2.1.1) and individuals (Appendix 2.1.3) separately. If the ANOVA test showed significant results, the test was proceeded by a paired t-test for T and NT groups and for different timeframes (Appendix 2.1.2 & Appendix 2.1.4). The latter test was performed to examine whether there was a significant difference in insect abundance over time. Another two-way ANOVA test was performed to identify differences in insect abundance between different manure types (Appendix 2.1.5 & Appendix 2.1.6). Finally, another two-way ANOVA test was performed to identify the foraging preference difference between T and NT fields (Appendix 2.1.7).

(8)

8

Results

Effects of Tillage on Insect Abundance

Farm Reus

The graphs representing the measurements of insect biomass (figure 4) and insect individuals (figure 5) on farm Reus show different results. Whereas the first “after treatment” (t1) reveals a significantly higher value with NT in comparison to T (p < 0.01) on insect biomass (figure 4), this is not the case when only individuals (figure 5) are considered. In the individual counts, there is no significant difference between T and NT at any time-interval. The mean NT biomass on t1 is 0.0326 g. For T strips this is 0.0135 g. In total, 132 individuals constitute the NT biomass at t1, where 165 individuals constitute the T biomass. This indicates that, on average, insect individual weight is approximately twice as high on NT (0.494 mg) fields than on T (0.245 mg) fields. The captured individuals and their biomass do show similarities in their overall development over the different time intervals. In both graphs, the quantities are relatively similar from t0-t3. At t4, both figures present a high increase.

Figure 4: Boxplot of insect biomass (g) of Tilled (T) and untilled (NT) strips on farm Reus over time intervals t0-t4. t0 is measurement before treatment, t1 is first

measurement after treatment. On t1 difference between T and NT is significant (p < 0.01).

Figure 5: Boxplot of insect

individuals of Tilled (T) and untilled (NT) strips on farm Reus over time intervals t0-t4. t0 is measurement before treatment, t1 is first

measurement after treatment. There is no significant difference on any time interval.

This increase is significant, especially for the T fields (figure 6). When individuals only are considered it is also significant for NT fields at t2 and t0.

(9)

9 Figure 6: Boxplots of insect biomass (left) and insect individuals (right) of T and NT with significance levels over time. After adjustments, there were only significant differences in the T group for insect biomass. Within this group, insect biomass was significantly higher at t4 than at t0 (p<0.05), t1 (p<0.01) and t2 (p<0.01). For the individuals, there was a significant difference between the different timeframes for the NT and the T group. For the T group, these differences were significant between t4 and all other time frames (p<0.01). For the NT group these differences were significant between t4 and t2 (p<0.05) and between t4 and t0 (p<0.05).

The cause of this increase is possibly explained by weather conditions (table 3). This relationship is further discussed in the Discussion section.

Table 3: Means of wind speed and precipitation in Hoorn during different time frames.

Timeframe t0 t1 t2 t3 t4

Wind speed(km/h) 25,7 31,2 22.075 22.15 27.9

Precipitation (mm) 0 4.68 0.29 0 9.54

Farm Slagter

The results of insect biomass (table 4) and individuals (table 5) at farm Slagter are of a different order than the results from farm Reus. At farm Slagter, there was only a before and an after treatment measurement. The biomass capture (table 4), shows a decrease for NT and an increase for T after manure is applied. As there is only one sample of each treatment at each point in time, it is impossible to prove significance as the data do not contain enough power for statistical testing.

Table 4: Biomass (g) of insects captured on farm Slagter. “Before” shows results of measurement before the treatment, “After” are the after treatment results.

Field Tillage Biomass (g) before Biomass (g) after

A NT 0.05 0.0404

B T 0.0293 0.0596

The results for the individuals present a small increase for both T and NT after manure application. Identically to the biomass results, significance is impossible to prove due to inadequate sample size.

(10)

10 Table 5:Insect Individuals captured on farm Slagter. “Before” shows results of measurement before the treatment, “After” are the after treatment results.

Field Tillage Individuals before Individuals after

A NT 169 182

B T 116 120

On farm Slagter the 120 individuals on the T field constitute a higher biomass (0.0596 g) than the 182 individuals on the NT field, where biomass is 0.0404. This indicates that on average, the individuals have a higher weight on T fields after the treatment.

Effects of Manure on Insect Abundance

The effects of manure on both biomass (figure 7) and individuals (figure 8) are insignificant at all measurement time intervals. Although insects on “Horse” strip appear to be significantly higher at t0 and t3 (figure 7 & figure 8), this is not the case. At each time-interval, there is only one sample available of insects present on Horse manure, whereas there are two samples available of Goat and Pig manure at each time-interval. Therefore, the higher biomass of Horse at t0 and t3 (figure 7) is not considered as significantly higher than the biomass of Goat and Pig at those measurements. The progress over time of both insect biomass and individuals on the different types of manure shows a similar development as in the graphs where T and NT are compared (figure 7 and figure 8).

Figure 7: Boxplot of insect biomass (g) in different manure types (Goat, Pig and Horse) on farm Reus over time intervals t0-t4. t0 is

measurement before treatment, t1 is first

measurement after treatment. There is no significant difference on any time interval.

Figure 8: Boxplot of insect Individuals in different manure types (Goat, Pig and Horse) on farm Reus over time intervals t0-t4. t0 is measurement before treatment, t1 is first

measurement after treatment. There is no significant difference on any time interval.

(11)

11

Bird Nests and Behaviour

Farm Reus

Behaviour

At farm Reus the following birds were present within the five strips: 4 lapwings (vanellus vanellus), 2 oyster catchers (Haematopus ostralegus) and 1 wood pigeon (Columba palumbus). Furthermore, one hare (Lepus europaeus) was present on the strips (mostly strip A) for the whole measuring hour. After 20 minutes (timeslot 5) a predatory bird, most likely a common kestrel, (Falco tinnunculus) flew over and most birds responded by initiating flight combined with warning noises. The high abundance of airborne birds made it impossible to count all of them, but at least three pairs of airborne lapwings were spotted.

During the measuring hours, these birds were exhibiting various types of behaviour. The lapwings were mostly brooding, interchanged by distanced guarding and foraging. Both the oyster catchers and the pigeons were solely foraging. The foraging (table 6) was significantly more practiced in the NT strips (p<0.05).

Table 6: Total minutes of foraging per strip on farm Reus in one hour. If, for example, in one timeframe (5 mins) two birds were foraging on the same strip, this is counted as 10 minutes (2*5 minutes). Foraging was significantly higher in NT strips (p<0.05)

Strip A(T) B(NT) C(NT) D(T) E(T)

Foraging minutes

10 25 25 10 10

Nests

The strips on farm Reus contained 4 lapwing nests during the measuring time: one nest on strip A, one nest on strip C and two nests on strip E. The nests were discovered around 9 April 2021 and hatched about a month later. It is assumed all nests hatched. Chick survival rate is unknown.

Farm Slagter

Behaviour

At farm Slagter, at least 8 lapwings (Vanellus vanellus) were observed on field A, together with 5 wood pigeons (columba palumbus), 2 oyster catchers (Haematopus ostralegus) , 2 eurasian jackdaws (Corvus monedula) and two undefined small brown birds, most likely in the family of prunellidae. Alongside field A, a common kestrel was hovering in the wind. However, it was never actually above the field or hunting in the field. The lapwings appeared to exhibit mostly courtship and territorial behaviour. All other birds were foraging. At this measuring, field B was covered with a black sheet to prevent pigeons, feeding from the seedlings. During the whole measuring time, there were

approximately 40 adult and juvenile black-headed gulls (Larus ridibundus) and 2 oyster catchers foraging on the field.

Nests

On farm Slagter, four bird nests were counted on field A and zero nests on field B during the measuring time. During the last observation, the nests had hatched. Chick survival rate is unknown.

Discussion

The effects of manure on insect abundance are highest shortly after manure is exposed, according to Lahr & Van der Pol (2007); and the effects steadily decrease after 10-30 days. This is also

demonstrated by the biomass results (figure 4). Directly after manure application, insect biomass was significantly higher on untilled strips than on tilled strips. This is in line with the hypothesis that the effects are strongest directly after manure application. In order to carefully draw a conclusion of the effects of no tillage on biomass, other factors that could influence the malaise trap capture must be ruled out. The most important factor that decides whether the malaise trap sample represents the wide

(12)

12 or close surroundings is wind speed. If wind speed is high, it is likely that the captured insects are not originating from close distance to the trap. Therefore, it is important to examine wind speed during the measuring time frames (table 3). Wind speed is higher on t1 than on t0. This would indicate that the capture of the malaise traps at t1 represents a larger area than the capture at t0. However, the difference between insect abundance on NT and T is high at t1, but there is almost no difference in insect capture at the T fields between t0 and t1. This indicates that the high biomass observed at t1, is in fact a result of no-tillage management.

This effect is solely visible on biomass and not on individuals. It is the insect biomass that constitutes the nutritional value for meadow bird chicks. Therefore, the finding that mean individual biomass was approximately twice as high on NT fields (0.494 mg) than on T fields (0.245 mg) is very relevant for these chicks. In his research, Beintema (1989) examined, inter alia, the dietary needs of captive meadow bird chicks such as the black-tailed godwit and the lapwing. He indicated that for a prey weight of 0.5 mg – which is comparable to the mean insect biomass on NT fields at t1 – these meadow bird chicks have to consume approximately 1500 insects per hour to meet their energy demands. At a prey biomass of 0.25 mg – which is comparable to the mean biomass of T fields at t1 – this quantity more than doubled to an amount of 3500 insects per hour. In the wild, meadow birds and their chicks have to be alert for predators; therefore, the time for feeding is shorter for wild meadow birds than for birds living captivity. Furthermore, there are other factors that negatively influence feeding time in the wild such as water shortage, which causes meadow bird chicks to inhibit feeding (Beintema, 1989). As for these reasons feeding time in the wild is limited, it is even more favourable for chicks to consume insects with a higher biomass. So, the observations and remarks presented here allow to conclude that feeding on insects is more favourable on NT fields than on T fields, as the so-called profit is higher on these fields.

The bird spotting findings at farm Reus are mostly in line with the expected results. It appears that the meadow birds are more comfortable foraging in the untilled area. It is unclear whether this is caused by a higher food supply only, or that the higher vegetation density at the NT fields also contributes to the preference for foraging in NT fields. The relationship between foraging preference and vegetation cover is highly species dependent. Some species actually prefer bare ground over covered ground (Korniluk et al., 2020). Generally, the more structurally complex a foraging habitat is (i.e. the higher the vegetation grows), the harder it is for lapwings to find food (Devereux et al., 2004). However, vegetation on the NT fields was relatively low, and completely absent on the T fields due to milling before fertilizing. As found in literature and supported by the results of this research, this milling affects insect population diversity and abundance (Müller, 2000). Therefore, it can be concluded that the absence of soil conditioning in NT fields results in a preferable foraging area, even though the structure is more complex than on T fields. Furthermore, Oosterveld (2006) indicated that direct solid manure application on the soil is favourable for worm populations, which are an important source of feed for adult meadow birds. As the observed birds in this research were adult, it is likely that the NT fields were preferred for foraging because of a higher worm abundance. Data on worm populations on farm Slagter and Reus are known. Therefore, this research could be extended to inspect whether there is a relationship between meadow bird foraging behaviour and worm populations in T and NT fields. Insect sampling could be improved by increasing the sample size. For this research this means that more of the same malaise traps should be used, especially on farm Slagter, where there is ample surface area on both fields. During the set-up of the after measurements on farm Slagter, the NHBS traps were placed in the same direction as the wind direction, whereas the correct placement is

perpendicular to the wind. Due to this erroneous placement of the NHBS traps, the after measurements of both field A and field B were only correctly sampled by the Ento Sphinx tents. The catches of the NHBS traps were so low that those samples were useless for this study. Therefore, the results of farm Slagter presented in the results section are completely described in the tables (table 4 & table 5). Furthermore, pitfalls and stick-traps could contribute to capturing a more local population.

(13)

13 traps captures fluctuate less than other traps (Butler et al., 1999), this type of sampling is most likely the best sampling option, especially on farm Slagter in its large surface area.

There were three nests in tilled strips and one nest in the untilled strip C. As the research area on farm Reus is relatively small, it was not expected that T or NT affects bird nest location choice there. On farm Slagter, however, there were four nests present on the untilled field and zero on the tilled field. This higher nest abundance is not necessarily a result of minimal tillage, as the untilled field is spread over a larger area than the tilled field. Field A is approximately 9 ha, while field B is approximately 3 ha. However, when size is compensated, field A still houses more nests, which might be an effect of no-tillage manure application. Concerning he bird behavioural observations at farm Slagter, it is impossible to compare field A with field B. At the time of observation, field B had just been planted and the seedlings were protected from pigeon feeding by a cover sheet. Field A was still bare land at that point. As there were only foraging gulls and two oyster catchers present on field B at the time of observation, it can be concluded that meadow birds prefer untilled bare land for foraging over a tilled field covered with a black sheet. This conclusion is highly unsurprising.

Another remarkable result is the high increase in insect individuals, as well as an increase in insect biomass – although to a smaller extent – at t4. These individuals were mostly collembola. These collembola were mostly captured on all strips on farm Reus. The t4 measurement was approximately a month after fertilizing. According to Lahr & Van der Pol (2007), it is during a period of 10-30 days that collembola become more abundant in manure, as they feed on decomposing plant material and the associated fungi. Therefore, the presence of collembola in the manure on farm Reus are a reasonable occurrence. However, the capture of collembola in malaise traps is an unexpected result, as these hexapods lack wings and are therefore not considered as low-flying insects. It is likely that the collembola entered the malaise traps because they attempted to escape from the heavy rains that occurred during timeframe t4. On each tilled strip the number of individuals was higher than on the untilled strips. There is a possibility that this higher capture is the result of a faster infiltration rate at the tilled strips. As the tilled strips contain large cracks and holes, water is likely to infiltrate deep in the soil quite rapidly. However, as the soil below the tilled layer is dense clay, it is likely that the water quickly rises, initiating an escape reaction of the collembola. However, this is merely speculation. The relationship between tillage, precipitation and collembola behaviour is matter for further research. The different types of manure did not show significant differences. Specifics on the properties of the different types of manure were unknown; therefore, no hypothesis was formulated on the effects of manure type on insect abundance. Further research could be aimed to relate manure properties with insect abundance, when these properties are known.

Conclusion

In conclusion, no-tillage manure application is favourable for meadow bird foraging. Fields where manure is not incorporated into the soil directly after application have significantly higher insect biomass. This is favourable for meadow birds and their chicks, as the time available for foraging is limited in the wild; therefore, prey with a higher biomass is advantageous. These effects were not found on farm Slagter, as data was limited due to sampling errors on this farm. The main conclusions drawn for this farm with the available data, is that meadow birds do not seem to be attracted to a sheet-covered soil, contrarily to gulls, who seemed to forage on these sheets. Furthermore, the untilled field on this farm housed four lapwing nests, whereas the tilled field housed zero nests. This allows to carefully conclude that untilled fields are more preferable for nesting. Wind can have an effect on the capture of malaise traps. The more wind, the more likely it is that the insect capture is a representation for a larger area. This did not seem to affect the results of this research. Precipitation did seem to affect the results. As in the last measurement timeframe there were days with high precipitation, collembola unexpectedly entered the malaise traps. These collembola were captured in all strips on farm Reus, but mostly in the tilled strips. Collembola behaviour in relation to tillage and precipitation is hardly known, and could therefore be further inspected. Furthermore, as worms constitute a high fraction of

(14)

14 adult meadow bird’s diets, worm population dynamics as a consequence of (no-) tillage management could be connected to the bird data as obtained in this research. Manure types did not have a

significant effect on insect abundance. In the coming years, this research will continue. It is

recommended that on farm Slagter the insect sampling devices are of the same type, both to prevent errors and to obtain data with better interpretability. Furthermore, implementation of other devices besides malaise traps to obtain more local insect samples is recommended for extension of this research.

References

Beintema, A. J. (1989). Breeding Ecology of Meadow Birds (Charadriiformes); Implications for Conservation and Management.

Bini, C. (2009). Soil: A precious natural resource. In Conservation of Natural Resources (pp. 1–48). Butler, L., Kondo, V., Barrows, E. M., & Townsend, E. C. (1999). Effects of weather conditions and

trap types on sampling for richness and abundance of forest macrolepidoptera. Environmental Entomology, 28(5), 795–811. https://doi.org/10.1093/ee/28.5.795

Cerdà, A., Rodrigo-Comino, J., Yakupoğlu, T., Dindaroğlu, T., Terol, E., Mora-Navarro, G., Arabameri, A., Radziemska, M., Novara, A., Kavian, A., Vaverková, M. D., Abd-Elmabod, S. K., Hammad, H. M., & Daliakopoulos, I. N. (2020). Tillage versus no-tillage. soil properties and hydrology in an organic persimmon farm in eastern Iberian Peninsula. Water (Switzerland), 12(6). https://doi.org/10.3390/W12061539

Corsi, S., & Muminjanov, H. (2015). Conservation agriculture: Training guide for extension agents and farmers in Eastern Europe and Central Asia. FAO, 1–665. https://doi.org/10.1007/978-3-319-11620-4

Derpsch, R., & Benites, J. (2003). Situation of conservation agriculture in the world. World Congress on Conservation Agriculture, 2, 67–70. http://www.madadoc.mg/8148_Article Igassu 2003 Derpsch CA in the world.pdf

Derpsch, R., Franzluebbers, A. J., Duiker, S. W., Reicosky, D. C., Koeller, K., Friedrich, T., Sturny, W. G., Sá, J. C. M., & Weiss, K. (2014). Why do we need to standardize no-tillage research? Soil and Tillage Research, 137, 16–22. https://doi.org/10.1016/j.still.2013.10.002

Devereux, C. L., Mckeever, C. U., Benton, T. G., & Whittingham, M. J. (2004). The effect of sward height and drainage on Common Starlings Sturnus vulgaris and Northern Lapwings Vanellus vanellus foraging in grassland habitats. In Ibis.

FAO. (2017). The future of food and agriculture–Trends and challenges (Vol. 4, Issue 4).

Friedrich, T., Derpsch, R., & Kassam, A. (2017). Sustainable Development of Organic Agriculture. Sustainable Development of Organic Agriculture, 6, 0–7.

https://doi.org/10.1201/9781315365800

Khan, H. A. A., Shad, S. A., & Akram, W. (2012). Effect of livestock manures on the fitness of house fly, Musca domestica L. (Diptera: Muscidae). https://doi.org/10.1007/s00436-012-2947-1 Korniluk, M., Białomyzy, P., Grygoruk, G., Kozub, Ł., Sielezniew, M., Swie Tochowski, P., Tumiel,

T., Wereszczuk, M., & Chylarecki, P. (2020). Habitat selection of foraging male Great Snipes on floodplain meadows: importance of proximity to the lek, vegetation cover and bare ground. https://doi.org/10.1111/ibi.12898

Lahr, J., & Van Der Pol, J. J. C. (2007). Belangrijkste groepen dieren, levenswijze & ecologische diensten.

(15)

15 Agriculture, Ecosystems and Environment, 55(3), 201–212.

https://doi.org/10.1016/0167-8809(95)00609-V

Müller, D. G. (2000). Research Collection.

Oosterveld, E. (2006). Betekenis van waterpeil en bemesting voor weidevogels. De Levende Natuur, 107(3), 134–137.

RVO. (2020). Hoe mest uitrijden. https://www.rvo.nl/onderwerpen/agrarisch-ondernemen/mest/gebruiken-en-uitrijden/hoe-mest-uitrijden

Schekkerman, H, Teunissen, W., & Oosterveld, E. (2005). Resultaatonderzoek Nederland Gruttoland; broedsucces van grutto’s in beheersmozaïeken in vergelijking mat gangbaar agrarisch

graslandgebruik. Wageningen, Alterra, Alterra-Rapport 1291.

http://www2.alterra.wur.nl/Webdocs/PDFFiles/Alterrarapporten/AlterraRapport1291.pdf Schekkerman, Hans, & Beintema, A. J. (2007). Abundance of invertebrates and foraging success of

Black-tailed Godwit Limosa limosa chicks in relation to agricultural grassland management. Ardea, 95(1), 39–54. https://doi.org/10.5253/078.095.0105

Sheikh, A. H., Thomas, M., Bhandari, R., & Meshram, H. (2016). (13) Malaise trap and insect sampling: Mini review | Request PDF. Bio Bulletin, 2(2), 35–40.

https://www.researchgate.net/publication/339139611_Malaise_trap_and_insect_sampling_Mini_ review

Sovon. (z.d.). Sovon Vogelonderzoek | Grutto. Geraadpleegd op 26 maart 2021, van https://stats.sovon.nl/stats/soort/5320

Veldshop.nl. (z.d.-a). Malaiseval. Geraadpleegd op 27 maart 2021, van https://www.veldshop.nl/nl/malaiseval.html

Wang, Z., Li, Y., Li, T., Zhao, D., & Liao, Y. (2020). Tillage practices with different soil disturbance shape the rhizosphere bacterial community throughout crop growth. Soil and Tillage Research, 197(November 2019), 104501. https://doi.org/10.1016/j.still.2019.104501

Wang, Z., Liu, L., Chen, Q., Wen, X., & Liao, Y. (2016). Conservation tillage increases soil bacterial diversity in the dryland of northern China. Agronomy for Sustainable Development, 36(2). https://doi.org/10.1007/s13593-016-0366-x

Williams, C. B. (1961). Studies in the Effect of Weather Conditions on the Activity and Abundance of Insect Populations. Philosophical Transactions of the Royal Society of London, 244(713), 331– 378.

Acknowledgements

First of all I would like to thank my supervisors Elly Morriën and Anne Uilhoorn. I thank Elly Morriën for providing me with relevant scientific insights on soil fauna and insects and their relation with meadow birds. This sublime guidance and feedback subtantially increased the quality of my thesis, and I am thankful for that. I would like to thank Anne Uilhoorn for her high responsiveness and her detailed feedback. Without her insights on data analysis I would not have been able to obtain the results that I have now. Furthermore, I want to thank Léon Feenstra for helping me with the data collection. Together, we collected most of the insect samples and conducted both the bird counts. Without his help and presence, the fieldwork would not have been nearly as pleasant as it was now. I want to thank Judith Nispeling for reading and improving my thesis and for providing overall support during this research. For providing their own farms as a research area and for counting the bird nests, I would like to thank farmers Wim Reus and Pé Slagter. There would not have been any research in Noord-Holland on no-tillage farming, if they did not offer their farms for this purpose. I would like to

(16)

16 thank the people of Amsterdam Green Campus for helping to set up this project, and for providing general guidance to me and other students. For the lab analysis I owe gratitude to the friendly lab technicians and managers of the Institute for Biodiversity and Ecosystem Dynamics; especially to Rutger van Hall for introducing me to the labs and allowing me to work individually. Finally, I would like to express my gratitude towards the province of Noord-Holland for funding this project.

Appendices

Appendix 1: Data for Rstudio

The data used for Rstudio analysis are presented here. The raw data and the ready-to-use Rstudio data are available in the onedrive repository of the “Duurzame Bloemkool” project under

Resultaten_Analyses>Insecten for insect data, and under Resultaten_Analyses>Vogels for bird data. Raw data on insects is a file named Insecten_Data_Bloemkoolproject.ods. Raw data on birds is a file called VogelData.ods. The data used in R have the same file names as the titles below. Rscript file is allocated to Resultaten_Analyses>Insecten as well. It is named DouwedeMaijerThesisScript.R. WimTotalBiomassR.xlsx: These are data used in Rstudio for biomass analysis on farm Reus.

Strip Tillage Manure t0 t1 t2 t3 t4

A T Goat 0,0207 0,0134 0,0047 0,0272 0,0548

B NT Goat 0,0069 0,0338 0,0153 0,0147 0,0333

C NT Pig 0,0247 0,0314 0,0049 0,0324 0,0487

D T Pig 0,003 0,01 0,0158 0,031 0,0453

E T Horse 0,0416 0,0171 0,0105 0,0377 0,0493

WimTotalIndivR.xlsx: These are data used in Rstudio for individual analysis on farm Reus.

Strip Tillage Manure t0 t1 t2 t3 t4

A T Goat 57 27 11 82 531

B NT Goat 38 42 22 69 173

C NT Pig 57 90 31 131 238

D T Pig 15 71 36 96 303

E T Horse 84 67 20 106 256

PeBiomassR.xlsx: These are data used for biomass analysis on farm Slagter.

Field Trap Tillage Before After

Pé A (E.S.) 2 NT 0,05 0,0404

Pé B (E.S.) 2 T 0,0293 0,0596

PeIndividualsR.xlsx: These are data used for individual analysis on farm Slagter.

Field Trap Tillage Before After

Pé A (NHBS) 1 No 236 40

Pé B (NHBS) 1 Yes 374 16

Pé A (E.S.) 2 No 169 182

Pé B (E.S.) 2 Yes 116 120

(17)

17 foerageerR.xlsx: These are data used for foraging analysis on farm Reus

Strip Til-lage 1 2 3 4 5 6 7 8 9 10 11 12 A T 0 0 10 0 0 0 0 0 0 0 0 0 B NT 0 0 0 10 10 0 0 0 0 5 0 0 C NT 0 0 0 0 0 5 10 5 5 0 0 0 D T 5 0 0 0 5 5 0 0 0 0 0 0 E T 0 0 5 0 0 0 0 0 0 0 0 0

Appendix 2: Statistical Analysis

Appendix 2.1: Statistical hypotheses

Appendix 2.1.1: Testing for difference between T and NT fields in insect biomass

Rscript: row 23 - row 82) Two way ANOVA test H0: µT = µNT

Ha: µT ≠ µNT

Where:

µT is the mean insect biomass (g) sampled on T fields;

µNT is the mean insect biomass (g) sampled on NT fields.

Post hoc: paired t-test for each different timeframe H0: µT = µNT

Ha: µT ≠ µNT

Where:

µT is the mean insect biomass (g) sampled on T fields in each timeframe;

µNT is the mean insect biomass (g) sampled on NT fields in each timeframe.

Appendix 2.1.2: Testing for difference in insect biomass between different timeframes

Rscript: row 83 – row 103. Paired t-test:

H0: µt0 = µt1= µt2 = µt3= µt4

Ha: µt0 ≠ µt1 ≠ µt2 ≠ µt3 ≠ µt4

(18)

18

Appendix 2.1.3: Testing for difference between T and NT fields in insect individuals:

Rscript: row 104 – row 168. Two-way ANOVA test H0: µT = µNT

Ha: µT ≠ µNT

Where:

µT is the mean insect individuals sampled on T fields;

µNT is the mean insect individuals sampled on NT fields.

Post hoc: paired t-test for each different timeframe: H0: µT = µNT

Ha: µT ≠ µNT

Where:

µT is the mean insect individuals sampled on T fields in each timeframe;

µNT is the mean insect individuals sampled on T fields in each timeframe.

Appendix 2.1.4: Testing for difference in insect individuals between different timeframes

Rscript: row 169 - 187 Paired t-test:

H0: µt0 = µt1= µt2 = µt3= µt4

Ha: µt0 ≠ µt1 ≠ µt2 ≠ µt3 ≠ µt4

Where µtx is the mean insect individuals in timeframe tx, grouped by T and NT.

Appendix 2.1.5: Testing for difference in insect biomass between goat, pig and horse manure

Rscript: row 190 - 214 H0: µgoat = µpig= µhorse

Ha: µtgoat ≠ µtpig ≠ µthorse

Where µgoat is the mean biomass (g) of insects on fields with goat manure;

µpig is the mean biomass (g) of insects on fields with pig manure;

µhorse is the mean biomass (g) of insects on fields with horse manure.

Appendix 2.1.6: Testing for difference in insect individuals between goat, pig and horse manure

Rscript: row 237 – 299 H0: µgoat = µpig= µhorse

Ha: µtgoat ≠ µtpig ≠ µthorse

Where µgoat is the mean of insect individuals on fields with goat manure;

µpig is the mean of insect individuals on fields with pig manure;

(19)

19

Appendix 2.1.7: Testing for difference in foraging time between T and NT fields

Rscript: row 354 - 374 H0: µT = µNT

Ha: µT ≠ µNT

Where µT is the mean foraging time in T fields;

µNT is the mean foraging time in NT fields

Appendix 2.2: Rscript

# Bachelor Thesis: Effect of Minimal TIllage management # on insect abundance # By Douwe de Maijer # 17/05/2021, Wageningen ######################################################## ######################################################## #

# Wim total Biomass

# Total Biomass of all captured insects on farm Wim # on 5 different times, where t0 is the "before" # and t1-t4 are the "after measurements"

# 2-way mixed anova: 1 'between' subject factor: Tillage # (T or NT) 1 'within' subject factor: time

###################################################### # load the required packages:

library(tidyverse) library(ggpubr) library(rstatix) library(datarium)

# Gather columns t0-t4 into long format and convert # time, tillage and manure into factor variable Biomass <- WimTotalBiomassR %>%

gather(key = "time", value = "biomass (g)", t0, t1, t2, t3, t4) %>%

convert_as_factor(Tillage, time, Manure) Biomass

# group data by tillage and time Biomass %>%

group_by(Tillage, time) %>%

get_summary_stats(`biomass (g)`, type = "mean_sd") # Create boxplot over time of biomass

# (comparison between tillage and no tillage)

bxb <- ggboxplot(Biomass, x = "time", y = "biomass (g)", color = "Tillage", palette = "jco")

# Show boxplot bxb

# Check wether assumptions are met: # Identify Outliers

Biomass %>%

identify_outliers(`biomass (g)`) # no extreme outliers found

# Check for Normality Biomass %>%

(20)

20 group_by(time, Tillage) %>%

shapiro_test(`biomass (g)`)

# shapirotest not working, so I'll try

ggplot <-ggqqplot(Biomass, "biomass (g)", ggtheme = theme_bw())+ facet_grid(time~Tillage)

ggplot

# Points fall approximately along the reference line

# therefore, normality is assumed. However, there are only two reference points

# for NT and three for T. Therefore, the reference line falls exactly between # these points. # Homogeneity of variance Biomass %>% group_by(time)%>% levene_test(`biomass (g)`~Tillage)

# p > 0.05. Therefore, homogeneity is assumed # homogeneity of covariances

box_m(Biomass[, "biomass (g)", drop = FALSE], Biomass$Tillage) # p > 0.001, so there is homogeneity of covariances

# assumption of sphericity is automatically tested in # anova test

# Assumptions are met, so time to compute the # two-way mixed ANOVA test

anovaresults <- anova_test(

data = Biomass, dv = "biomass (g)", wid = Strip, between = Tillage, within = time

)

anovaresults ?pairwise_t_test

# No significant of Tillage. Time does have # significant effect on biomass

# Check effect of tillage at each time point # pairwise comparisons between tillage

pairwise <- Biomass %>% group_by(time) %>%

pairwise_t_test(`biomass (g)`~ Tillage) pairwise

# create boxplot with significance levels

pairwise <- pairwise %>% add_xy_position(x = "time") pwcfilter <- pairwise %>% filter(time!= "t0, t2, t3, t4")

bxb + stat_pvalue_manual(pwcfilter, tip.length = 0, hide.ns = TRUE) +

labs(subtitle = get_test_label(anovaresults, detailed = TRUE), caption = get_pwc_label(pairwise))

# Pairwise comparisons between timeframes pairwisetime <- Biomass %>%

group_by(Tillage) %>%

pairwise_t_test(`biomass (g)`~ time) pairwisetime

# create boxplot with significance levels

pairwisetime <- pairwisetime %>% add_xy_position(x = "time") bxb + stat_pvalue_manual(pairwisetime, tip.length = 0, hide.ns = TRUE) +

(21)

21 caption = get_pwc_label(pairwisetime))

# Interpretation: Overal p-value is 0.17. Therefore # over the five measuring times, we cannot reject # the NULL hypothesis that Tillage has no significant # effect on insect biomass.However, at t1, which is # the first measurement after the treatment p < 0.01 # This means that only at t1 there is a significant # effect where we can reject the NULL hypothesis # that Tillage has no effect on insect biomass abun-

# dance. Furthermore, as the pairwisetime table presents. There is no signifcant difference

# over between time within the NT groups. This is probably because there are not enough degrees of freedom

# to demonstrate significant difference. In the T group there is a significant difference between t0-t4, t1-t4, t2, t4.

# Wim total individuals

# Let's see whether tillage has a significant effect # on individual abundance

Individuals <- WimTotalIndivR %>%

gather(key = "time", value = "Individuals", t0, t1, t2, t3, t4)%>% convert_as_factor(Tillage, time, Manure)

Individuals

# group data by tillage and time and get summary # statistics

Individuals %>%

group_by(Tillage, time) %>%

get_summary_stats(`Individuals`, type = "mean_sd") # Create boxplot over time of Individuals

bxbind <- ggboxplot(Individuals, x = "time", y = "Individuals", color = "Tillage", palette = "jco")

# Show boxplot bxbind

Individuals %>%

identify_outliers(Individuals)

# At measurement t4 the individuals contain outliers # This was not visible in biomass because these

# individuals are collembola, which have very low weight # First, all timepoints are researched for

# significance. Afterwards, we'll also examine the # on t0 and t1 only to inspect the treatment effect # Check for Normality

Individuals %>%

group_by(time, Tillage) %>% shapiro_test(Individuals)

ggplotindiv <-ggqqplot(Individuals, "Individuals", ggtheme = theme_bw())+ facet_grid(time~Tillage)

ggplotindiv

# Points fall approximately along the reference line # therefore, normality is assumed

(22)

22 Individuals %>%

group_by(time)%>%

levene_test(Individuals~Tillage)

# p < 0.05 for t3. Therefore, homogeneity is not # assumed here

# homogeneity of covariances

box_m(Individuals[, "Individuals", drop = FALSE], Individuals$Tillage)

# p > 0.001, so there is homogeneity of covariances # assumption of sphericity is automatically tested in # anova test

# Assumptions are not met, so time to compute the # two-way mixed ANOVA test

anovaindiv <- anova_test(

data = Individuals, dv = "Individuals", wid = Strip, between = Tillage, within = time

)

anovaindiv

# No significant effect of tillage on individuals. # Time does have significant effect. Let's see # when this effect is highest

pairwiseindiv <- Individuals %>% group_by(time) %>%

pairwise_t_test(Individuals~ Tillage) pairwiseindiv

pairwiseindiv <- pairwiseindiv %>% add_xy_position(x = "time")

bxbind + stat_pvalue_manual(pairwiseindiv, tip.length = 0, hide.ns = TRUE) +

labs(subtitle = get_test_label(anovaindiv, detailed = TRUE), caption = get_pwc_label(pairwiseindiv))

# Lets check if there is a significant difference over time pairwiseindivtime <- Individuals %>%

pairwise_t_test(Individuals~ time) pairwiseindivtime

pairwiseindivtime <- pairwiseindivtime %>% add_xy_position(x = "time")

bxbind + stat_pvalue_manual(pairwiseindivtime, tip.length = 0, hide.ns = TRUE) +

labs(subtitle = get_test_label(anovaindiv, detailed = TRUE), caption = get_pwc_label(pairwiseindivtime))

# No significant effects at any timepoint # However, the tillaged fields do have a much # higher capture of individuals. These are mostly # collembola. Maybe we can explain this with

# rain > higher runoff on tilled fields> earlier # saturation > so more collembola present under that # soil flee to get oxygen. On the other hand, maybe # are for some reason more collembola in tilled fields # Furthermore, time does have a signifcant effect on insect abundance

# for the NT group, this is for (t0,t1,t2,t3)-t4 and also for the T group

################################################### ###################################################

(23)

23 # Let's look at effect of manure only on farm Wim

bxbman <- ggboxplot(Biomass, x = "time", y = "biomass (g)", color = "Manure", palette = "jco")

# Show boxplot bxbman

# ANOVA test

anovaresultsman <- anova_test(

data = Biomass, dv = "biomass (g)", wid = Strip, between = Manure, within = time

)

anovaresultsman

# Again, no significant effect of manure on insect # biomass over time. Maybe there is an effect # on certain points in time. Let's find out pairwiseman <- Biomass%>%

group_by(time)%>%

pairwise_t_test(`biomass (g)`~ Manure) pairwiseman

pairwiseman <- pairwiseman %>% add_xy_position(x = "time")

bxbman + stat_pvalue_manual(pairwiseman, tip.length = 0, hide.ns = TRUE) +

labs(subtitle = get_test_label(anovaresults, detailed = TRUE), caption = get_pwc_label(pairwiseman))

# No significant difference of manure type on # biomass

# No results. Let's see what the combined effects # manure and tillage are with a three-way mixed # ANOVA. 2 between factors: tillage and manure # 1 within factor: time

Biomass <- Biomass %>% convert_as_factor(Strip) Biomass %>%

group_by(Tillage, Manure,time)%>%

get_summary_stats(`biomass (g)`, type = "mean_sd") bxbmantil <- ggboxplot(

Biomass, x = "Tillage", y = "biomass (g)", color = "Manure", palette = "jco",

facet.by = "time" )

bxbmantil

resaovmantil <- anova_test(

data = Biomass, dv = "biomass (g)", wid = Strip, within = time, between = c(Tillage, Manure) )

get_anova_table(resaovmantil)

# This is not possible; not enough degrees of freedom ###################################################

########################################################## ####################################################### # Difference between manure types on individuals

Individuals <- WimTotalIndivR %>%

gather(key = "time", value = "Individuals", t0, t1, t2, t3, t4)%>% convert_as_factor(Tillage, time, Manure)

(24)

24 # group data by tillage and time and get summary

# statistics Individuals %>%

group_by(Manure, time) %>%

get_summary_stats(`Individuals`, type = "mean_sd") # Create boxplot over time of Individuals

bxbind <- ggboxplot(Individuals, x = "time", y = "Individuals", color = "Manure", palette = "jco")

# Show boxplot bxbind

Individuals %>%

identify_outliers(Individuals)

# At measurement t4 the individuals contain outliers # This was not visible in biomass because these

# individuals are collembola, which have very low weight # First, all timepoints are researched for

# significance. Afterwards, we'll also examine the # on t0 and t1 only to inspect the treatment effect # Check for Normality

Individuals %>%

group_by(time, Tillage) %>% shapiro_test(Individuals)

ggplotindiv <-ggqqplot(Individuals, "Individuals", ggtheme = theme_bw())+ facet_grid(time~Manure)

ggplotindiv

# Homogeneity of variance Individuals %>%

group_by(time)%>%

levene_test(Individuals~Manure)

# p < 0.05 for t3. Therefore, homogeneity is not # assumed here

# homogeneity of covariances

box_m(Individuals[, "Individuals", drop = FALSE], Individuals$Manure)

# p > 0.001, so there is homogeneity of covariances # assumption of sphericity is automatically tested in # anova test

# Assumptions are not met, so time to compute the # two-way mixed ANOVA test

anovaindivman <- anova_test(

data = Individuals, dv = "Individuals", wid = Strip, between = Manure, within = time

)

anovaindivman

# No significant effect of tillage on individuals. # Time does have significant effect. Let's see # whe this effect is highest

(25)

25 group_by(time) %>%

pairwise_t_test(Individuals~ Manure) pairwiseindiv

pairwiseindiv <- pairwiseindiv %>% add_xy_position(x = "time")

bxbind + stat_pvalue_manual(pairwiseindiv, tip.length = 0, hide.ns = TRUE) +

labs(subtitle = get_test_label(anovaindiv, detailed = TRUE), caption = get_pwc_label(pairwiseindiv))

# Pé total biomass

PéBiomass <- PeBiomassR%>%

gather(key = "time", value = "biomass (g)", Before,)%>% convert_as_factor(Tillage, time, Trap, Field)

PéBiomass

# group data by tillage and time PéBiomass %>%

get_summary_stats(`biomass (g)`, type = "mean_sd") # Create boxplot over time of biomass

Pébiobxb <- ggboxplot(PéBiomass, x = "time", y = "biomass (g)", color = "Tillage", palette = "jco")

# Show boxplot Pébiobxb

PéBiomass %>%

identify_outliers(`biomass (g)`)

# There is one outlier in the before measurement. This # trap caught a lot of large flies like houseflies

# which are relatively heavy. This is probably explained # by the fact that in the before measurment, the white # trap was placed on field B, whic lures more flies # on field A this trap was placed in the "after" measure # -ment

# Check for Normality PéBiomass %>%

group_by(time, Tillage) %>% shapiro_test(`biomass (g)`)

ggplotPébio <-ggqqplot(PéBiomass, "biomass (g)", ggtheme = theme_bw())+ facet_grid(time~Tillage)

ggplotPébio

# Homogeneity of variance PéBiomass %>%

group_by(time)%>%

levene_test(`biomass (g)`~Tillage)

# p > 0.05. Therefore, homogeneity is assumed # homogeneity of covariances

box_m(PéBiomass[, "biomass (g)", drop = FALSE], PéBiomass$Tillage) # p > 0.001, so there is homogeneity of covariances

# assumption of sphericity is automatically tested in # anova test

(26)

26 # two-way mixed ANOVA test

anovaresultsPe <- anova_test(

data = PéBiomass, dv = "biomass (g)", wid = Field, between = Tillage, within = time

)

anovaresultsPe

# No significant of Tillage. Time does have # significant effect on biomass

# Data from farm Pe Slagter are not suited for anova analysis # due to inadequate sampling size

# BIRDS foraging Vogels <- VogelsR %>%

gather(key = "time", value = "Birds", t1,t2,t3,t4,t5,t6,t7,t8,t9,t10,t11,t12 )%>% convert_as_factor(Tillage) Vogels Vogels %>% group_by(time) Birdanov <- anova_test(

data = Vogels, dv = "Birds", wid = Strip, between = Tillage, within = time

)

Birdanov

Birdbox <- ggboxplot(Vogels, x = "Tillage", y = "Birds", combine = TRUE) Birdbox PairBirds <- Vogels %>% group_by(time)%>% pairwise_t_test(Birds~Tillage) PairBirds

PairBirds <- PairBirds %>% add_xy_position(x = "Tillage")

Birdbox + stat_pvalue_manual(tip.length = 0, hide.ns = FALSE) + labs(subtitle = get_test_label(Birdanov, detailed = TRUE), caption = get_pwc_label(PairBird

Effects of Minimal Tillage Management on Low-flying Insect Abundance and Meadow Bird Nests and Behaviour in Dutch Cauliflower Fields