How different soil textures influence the infiltration rate

(1)

1

How different soil textures influence

the infiltration rate

Ines Schatz

BSc Thesis Future Planet Studies Major Future Earth

31-05-2021 Amsterdam

Supervised by: Erik Cammeraat and Anne Uilhoorn

(2)

2

Abstract

As a consequence of human-induced climate change, the hydrological cycle is changing. Especially, precipitation patterns will increase in extremity, increasing the risks of floods and droughts. A key component of the hydrological cycle is filtration. Hence, obtaining infiltration data is highly valuable to determine potential surface runoff and prevent soil erosion, or flooding events. Since infiltration is dependent on physical soil properties, such as texture, the research question became: “What are the infiltration capacities of three different soil textures, being; 1) sand, 2) loess and, 3) clay?”. For this, a rainfall simulation, a mini-disk infiltrometer, and the inverse auger hole methods were used to obtain infiltration values of these three textures. The results were analyzed with a one-way ANOVA and resulted in high infiltration values for clay, and intermediate values for both loess and sand. As for the mean infiltrations from the mini-disk infiltrometer, the one-way ANOVA resulted in significant difference between textures. A post-hoc Tukey test resulted in a significant difference in infiltration between sand on the one hand, and loess and clay on the other. These results were remarkable since sand, being the least compact texture, was expected to have the highest value followed by loess and eventually clay. Next, correlation coefficients were calculated per field to see whether sorption played a prominent role. Here too, the type of correlation and strength differed per sample area. Overall, the results showed high temporal variability for the infiltration rates and did not provide a complete answer to answer the research question. The cause of this temporal variability was found to be the agricultural activity in the area and the kind of field maintenance. To improve these results, it is recommended to collect more consistent amounts of samples per texture, assess the suitability of each method for the given purpose and analyze the type of crop being grown.

(3)

3

1. Introduction

1.1 Defining Infiltration

Human-induced climate change is increasingly changing the hydrological cycle, especially extremes (Trenberth, 2011 & Cole et al., 2014 & IPCC, 2018). As a consequence, more extreme events can occur in precipitation patterns, often leading to floods and droughts. Floods are associated with high amounts of precipitation, are fairly local, and develop in short time scales (Trenberth, 2011). Since infiltration is a key aspect of the hydrological cycle, data on this infiltration can be crucial for flood prevention and other forms of land management. Infiltration is defined as the downward movement of water into soil or sediment (Johnson, 1963 & Ferre & Warrick, 2005 & Haghnazari, Shahgholi, Feizi, 2015) Yet, the movement of water through a (dry) porous medium can also be divided into 1) flow that takes place in the absence of gravity (e.g. horizontal flow), called sorption, and 2) flow that occurs solely as a result of downward gravitational forces, often defined as infiltration. Numerous factors affect infiltration, such as the water supply rate, the elapsed time since the onset of water application, the chemical composition of the soil/liquid, initial/boundary conditions, the spatial variability/distribution of hydraulic properties within the soil, topography, temperature and different biological activity in the soil. Since infiltration velocity is dependent on aggregate size, as well as pore/crevices or cavities size the infiltration capacity and sorption go hand in hand with the different soil textures (Assouline, 2013). As the water starts to infiltrate and the wetting front is propagated in dry soil, the diffusivity and gravitational component start to play a role. The infiltration rate is the amount of water that flows into the ground during a certain period (ibid). At the beginning of the infiltration process, the diffusion accelerates the propagation of the wetting front. Diffusivity becomes less important as the wetting front sinks deeper into the soil profile, leaving the gravitational component with relatively more importance. Ultimately, the infiltration capacity has been reached when infiltration only depends on the gravitational component (Cammeraat, 2005).

1.2 The relevance of infiltration data

When the maximal rate at which the soil can absorb water is exceeded, only part of the precipitation infiltrates and the remnants start to pond on the soil surface and generate runoff (Assouline, 2013). Infiltration data is therefore extremely important to quantify possible surface run-off and is a key component of the water budget equation (Assouline, 2013). Furthermore, estimating the infiltration rate of different soil types can be crucial in hydrology, agricultural and civil engineering, irrigation design, and geophysics (Berli et al., 2008). After all, for soluble substances, such as plant nutrients or pollutants, infiltration is the only way to flow into the soil and is thus regarded as a principal variable influencing fertilization, environmental quality, and conservation (Assouline, 2013).

1.3 Problem description

There are several issues to consider regarding surface runoff. Examples of social problems associated with large amounts of surface runoff are: flooding urban areas, damage to buildings/infrastructure, and overloaded sewage systems (Stolte et al., 1999). Regarding physical aspects, surface runoff can cause soil erosion leading to washing away of the topsoil and water accumulation at low-lying points (ibid). This study will therefore focus on investigating the following research question: “What are the infiltration capacities of three different soil textures, being; 1) sand, 2) loess, and 3) clay?”, with the appurtenant research aim of quantifying surface runoff to promote mitigation and land management. It is expected/hypothesized that the clayey soil will have a low infiltration capacity, the loamier texture will have an intermediate capacity, and the sandy soil will have a high capacity (Haghnazari et al., 2015).

1.4 Study Area

Since the research focuses on three different soil texture groups, the chosen study locations varied. An example of a specific study that studied/investigated the effects of surface water runoff was carried

(5)

5 out by Stolte et al (1999). Financed by the province of Gelderland, the municipality of Groesbeek, and the polder district Groot Maas en Waal, the DLO-Staring Center has studied/investigated how the southern slope area can be designed to minimize water and soil runoff (ibid). As a consequence of this and its hilly landscape and recurring floods (Waterschap Rivierenland & Gemeente Berg en Dal, n.d.), Groesbeek was the first chosen study location.

Groesbeek is located in the east of the Netherlands, close to the German border (Figure 2), and consists of mainly two soil types: loamy soils and more sandy-like soils (Stolte et al., 1999). The loamy soils consist of a thin loess layer, which covers more coarse-grained sand. The area is gently sloping and contains a variety of landscapes, including floodplains, agricultural areas, and forests (Li, 1998). Nevertheless, arable farming land is the main land use found in this study area, occupying about 49% of the total area (ibid). Crops grown are cereals, sugar beet, potatoes, and maize, among others. Besides crops, wine and dairy production is additionally taking place in the area (de Haan, 2017). For this study, five fields were chosen based on elevation and the surface runoff route of the water. The selected fields were, in order from high to low elevation: 1) a fallow field, 2) the so-called sand track, 3) a field located next to a campsite, 4) a field next to the Klein Amerika vineyard and 5) a sprayed field.

Additionally, a sandy clay soil was analyzed in Hem (field 6, Figure 3), to serve as a contrast between infiltration rates of moderately permeable soils, as found in Groesbeek. The land use found in Hem was arable farming land as well, with cauliflower being the main crop produced (Van Maanen et al., 2020). In the following chapter, the methodology and tools used will be described, after which the obtained results will be explained with the aid of a graphical representation. Finally, the discussion will address areas for improvement and put the results into context, followed by an overall conclusion.

HEM

Figure 2: Study location Groesbeek (Waterschap Rivierenland & Gemeente Berg en Dal, n.d.)

Figure 3: Study location Hem, blue areas are sandy clay

(6)

6

2. Methods and Data

2.1 Inverse auger hole method

First, a soil texture analysis was described, based on the soil texture triangle (WRB 2014, figure 4). Additionally, the inversed auger-hole method was applied to calculate the hydraulic conductivity (K) of multiple texture layers found in one soil profile, and to subsequently calculate the infiltration rates. The inversed auger hole method is based on the principles of Darcy’s law, meaning that when the wetted soil is saturated, the mean flow velocity in the wetted soil below starts to approach the hydraulic conductivity resulting in v = K (Oosterbaan, R. J., and Nuland, H. J., 1994, Determining the Saturated Hydraulic Conductivity, in Ritzema, H. P., ed., Drainage Principles and Applications, Volume 16: Wageningen, ILRI, p. 435-476.). As infiltration occurs both through the bottom and the sidewalls of the drilled hole, the final formula to calculate K is as follows (see p.463 Ritzema, 1994)

𝐾 = 1.15𝑟log(ℎ0+ 1 2𝑟)−log(ℎ𝑡+ 1 2𝑟) 𝑡−𝑡0 [1].

One hole was drilled per field, and filled with water up to the soil’s surface. Subsequently, the water level was recorded with the corresponding elapsed time with the help of a folding ruler. The infiltration velocity was determined by measuring how many cm the water level dropped in a given timespan (Ritzema, 1994).

Figure 4: The soil texture triangle (WRB, 2014) Figure 5: the soil texture triangle: main texture groups

2.2 Rainfall simulation

Secondly, a rainfall simulator, designed based on the methods described in Imeson (1977) and Kamphorst (1987) was used to calculate the amount of surface runoff and infiltration.

Control measurements

To determine the mean control values of the rainfall intensity, several measurements were carried out after from the fieldwork. First, a large water collection container was placed under the rainfall simulator, without a specific measuring surface. The simulation was carried out for one minute, and the amount of precipitation was collected in a gray container to calculate the precipitation intensity in precipitation per hour. The rain intensity was regulated by the diameter of the tube which was inserted into the jerry can. It should be noted that a greater diameter results in greater rainfall intensity. Three repetitions were carried out, resulting in a mean value of 81.19 mm/hr.

(7)

7

Fieldwork measurements

On each field, two simulations were performed. First, the runoff collection ring (Area = 706.86 cm2₎ was inserted into the ground with the collecting notch facing the lowest part of the soil surface. Here, the soil was cleared to make room for the plastic bottles that would collect the runoff. Then, the sides of the ring were sealed to prevent leakage. Subsequently, two gray crates were placed aside of the runoff collection ring on which the dripping plate was positioned, and the water was leveled. The jerrycan was placed on top of this dripping plate, after which the initial water height was marked. As soon as the rainfall simulator started to drip, the stopwatch was turned on to determine the time to runoff (TTR) (when water starts to flow out of the ring). After the TTR was recorded, the plastic bottles were replaced every two minutes, until the runoff was constant. Only when the runoff had become approximately constant did the simulation stop, and the final water height was again marked to calculate the final difference between start and end value. Consequently, the infiltration rate was calculated by subtracting the surface runoff from the rainfall intensity (Cammeraat, 2005).

Figure 6: Setup Rainfall Simulator: two gray crates, runoff collection ring, dripping plate, jerry can and collection bottles (Cammeraat, 2021)

2.3 Mini Disk Infiltrometer

Next, the infiltrometer was used to determine the unsaturated hydraulic conductivity. Hydraulic conductivity is defined as the rate at which water can move through the soil (Decagon Devices, 2016). Because the mini-disk infiltrometer is a tension meter, it measures the hydraulic conductivity at different applied tensions. The suction rate for different soil textures was determined according to the instructions of Decagon Devices (2016) (Table 1).

Groesbeek1: Sandy soil

Groesbeek2: Loess soil Hem: Clayey soil

Suction Rate (cm) 6 2 0.5

Time Interval (s) 2-5 30 180-360

Table 1: Mini disk infiltrometer adjustment: suction rate & time interval per soil texture (Decagon Devices, 2016)

Figure 7: Example of the vertical cross section of a Rainfall Simulator (measurements and parts

might differ from used exemplar) (Kamphorst, 1987).

(8)

8

Instructions of use

First, both the upper and lower chambers of the infiltrometer were filled with water, after which the suction rate was adjusted in the bubble chamber according to each soil texture (p. 8-11, Decagon Devices, 2016). Then the device was placed on a solid soil surface and the water volume (in ml) was recorded at time zero. Finally, the water level continued to be documented at random time intervals as water started to infiltrate. Per field, five infiltrometer measurements were conducted, each with at least five observations. After data collection, sorption and infiltration were calculated with the help of the accompanied excel spreadsheet (p. 13, Decagon Devices, 2016).

(9)

9

3. Results

In this section, the results will be addressed. First, the mean infiltration values of each texture were visualized per method, after which a one-way ANOVA was conducted to note whether infiltration differed significantly between textures (1. Sand, 2. Loess, 3. Clay). If an ANOVA resulted in p < 0.05, a Tukey-Kramer Post-Hoc Test was performed to see which textures differed significantly.. Finally, correlation coefficients were calculated to analyze if the variable sorption affected infiltration.

FIELD 1 FIELD 2 FIELD 3 FIELD 4 FIELD 5 FIELD 6 MAIN TEXTURE Sandy Loam Sand Loamy Sand Loamy Sand Loamy Sand Sandy Clay

LAND USE Arable Farmingland Arable Farmingland Arable Farmingland Arable Farmingland Arable Farmingland Arable Farmingland VEGETATION

COVER None None None grass/weeds A little bit of

A little bit of grass/weeds

None

Table 2: Texture, surface cover, and land use per field

3.1.1 Inverse auger hole method

For sand, the mean infiltration rate was 169.73 mm/hr (Table 3), which is close to the infiltration rate of clay, being 157.38 mm/hr (Table 3 & Figure 8). Infiltration in fields with loess on the top layer, on the other hand, rose remarkably above the other two textures with a mean value of 529.53 mm/hr due to the coarse sand texture found in the bottom of the profile. The different textures within this profile should have been measured separately to keep the infiltration rates of sand and loess separate. However, validations like this will be addressed in more detail in the discussion.

Sand Loess Clay

Mean Inf. (mm/hr.) 169.73 529.53 157.38

SD 0 235.01 42.83

(10)

10 Figure 8: Mean infiltration rate per texture for the inverse auger hole method. The error bars indicate the standard deviation and the letters indicate whether there was a significant difference between textures.

3.1.2 One-Way ANOVA

To investigate between-group differences, a one-way ANOVA was performed. Results demonstrated that the p-value is 0.26 (Table 4), which is greater than the alpha level of 0.05, meaning that there is no sufficient evidence to reject the null hypothesis. Therefore, it can be stated that there is no significant difference between infiltration means for the different textures, when using the inverse auger hole method. This was visualized in Figure 8 by giving all groups (textures) the same letter (A).

ANOVA

Source of Variation SS df MS F P-value F crit

Between Groups 156782.5 2 78391.26 1.646791 0.259334 4.737414

Within Groups 333217 7 47602.43

Total 489999.5 9

Table 4: Results from one-way ANOVA for the inverse auger hole method

3.2.1 Rainfall simulation

Regarding the rainfall simulation infiltration means, values differed slightly less compared to the inverse auger hole method. Here, however, the two textures most equal to each other were sand and loess in Groesbeek, with the former having an infiltration rate of 59.78 mm/hr, and the latter having an infiltration rate of 57.97 mm/hr. In Hem, the infiltration rate was highest since all precipitation infiltrated for each simulation, with a value of >81.19 mm/ (Figure 9).

Mean Inf. (mm/hr.) 59.78 57.97 81.19

SD 12.68 17.43 0

Table 5: Mean infiltration and standard deviation per texture for the rainfall simulations

0 100 200 300 400 500 600 700 800 900

Sand Loess Clay

Inverse Auger Hole

Mean Infiltration

A A A I (mm/hr) T

(11)

11 Figure 9: Mean infiltration rate per texture for the rainfall simulations. The error bars indicate the standard deviation and the letters indicate whether there was a significant difference between textures.

For the rainfall simulation data, the one-way ANOVA also resulted in a p-value > 0.05 (0.19, Table 6). Thus, again the null hypothesis could not be rejected, meaning that there is no statistical evidence for a significant difference between infiltration means.

ANOVA

Between Groups 1015.091 2 507.5455 1.958709 0.191517 4.102821

Within Groups 2591.224 10 259.1224

Total 3606.315 12

Table 6: Results from one-way ANOVA for the rainfall simulation

3.3 Mini Disk Infiltrometer

At first glance, a difference between the mean infiltration rate of sand and the mean rates of both loess and clay was observed using the mini-disk infiltrometer. The mean values found were 39.94 mm/hr. for sand, 15.57 mm/hr. for loess and 15.95 mm/hr. for clay (Figure 10). Here, sand had the highest infiltration, followed by clay, and loess.

Mean Inf. (mm/hr.) 39.94 15.57 15.95

SD 11.82 8.16 10.55

Table 7: Mean infiltration and standard deviation per texture for the mini-disk infiltrometer

0 10 20 30 40 50 60 70 80 90

Sand Loess Clay

Rainfall Simulation

Mean Infiltration

A A A I (mm/hr) T

(12)

12 Figure 10: Mean infiltration rate per texture for the mini-disk infiltrometer. The error bars indicate the standard deviation and the letters indicate whether there was a significant difference between textures.

The resulting p-value was < 0.05 (p = 4.77E-05, Table 8), indicating a significant difference of at least one group when using the Mini Disk Infiltrometer. Therefore, the Tukey-Kramer post-hoc test was performed to indicate where the exact dissimilarity occurred. Calculations resulted in two q values larger than the corresponding critical value of 3.49 (Table 9, Appendix A: Combined_Data/Extra/Statistics/Ines), which indicated that significant differences between means were found when comparing sand to loess (1 vs 2) and sand to clay (1 vs 3) (Table 9). The comparison of loess and clay did not result in a significant difference, indicated by giving these two textures a letter B (Figure 10).

ANOVA

Between Groups 2473.191 2 1236.596 14.71294 4.77E-05 3.354131

Within Groups 2269.3 27 84.04815

Total 4742.491 29

Table 8: Results from one-way ANOVA for the mini-disk infiltrometer.

Tukey-Kramer Post Hoc Test

Comparison Abs. Mean Diff n(Group A) n(Group B) SE q signifficant diff Critical Q 1 vs 2 24.3675 5 20 3.241299 7.51782 yes 3.49 2 vs 3 0.0215 20 5 3.241299 0.006633 no 1 vs 3 24.346 5 5 4.099955 5.938114 yes

Table 9: Tukey-Kramer Post-Hoc Test results

0 10 20 30 40 50 60

Sand Loess Clay

Infiltrometer

Mean Infiltration

I (mm/hr) A B B T

(13)

13

3.4 Correlations

Finally, Pearson’s correlation coefficient was determined to investigate the relation between sorption and infiltration. To calculate the sorption rate, data from the rainfall simulation was used, given that the soil here is wetted gradually. Per field, the rainfall simulation with most observations was chosen, after which sorptivity was calculated using the formula: S = (It – Kt)/sqrt(t). The chosen t was calculated in hours, based on the two-minute intervals per bottle switch. The calculations were carried out as long as K-sat was not yet reached, since only then sorption plays a role (Appendix A: Combined_Data/Extra/Sorptivity). The correlation between sorption and infiltration differed significantly per field measurement. Field one demonstrated a low negative value, indicating a weak inverse association between the variables, as did field four, where the association is close to zero. Additionally, field two had a slightly more moderate inverse association. Fields three and five, on the other hand, both pointed at a strong positive association. For field 6, the infiltration values resulted in numbers greater than the low precipitation intensity 81.19 (I > 81.19), because the precipitation infiltrated completely during the rainfall simulations. The infiltration was therefore estimated to be 81.19 for each repetition, resulting in a standard deviation of zero. As a result, no correlation coefficient could be calculated for the location Hem.

FIELD 1 FIELD 2 FIELD 3 FIELD 4 FIELD 5 FIELD 6

SORPTION & INFILTRATION

-0.23 -0.62 0.78 -0.03 0.89 NA

Table 10: Correlation coefficients of sorption and infiltration per field

Summarizing, it can be concluded that only the results of the mini-disk infiltrometer indicate a significant difference between texture groups sand and clay/loess. As for the correlations, both the strength as well as the direction varies from field to field. More possible causes for this will be discussed in the next chapter.

(14)

14

Discussion

4.1 Results in previous studies

Previous studies demonstrated that the temporal variation of infiltration values can be very high. For example, a study conducted by Elliot & Efetha (1999) found that infiltration rate, soil structure, and organic matter content are highly affected by tillage and cropping systems. It has been proven that conventional tillage reduces soil organic matter, and consequently also aggregate stability, aggregate size, pore continuity, and infiltration rate (Arshad et al., 1990; Carter 1992). Additionally, Table 1 (Appendix 1, p. 64) from Li (1998) demonstrates the saturated hydraulic conductivity for various soils in Groesbeek, with sometimes strikingly large differences in hydraulic conductivity between sand and loess soils, with deviations up to a factor 10. In addition, differences between months, due to field activity, were observed as well. Take for example the hydraulic conductivity of maize on a loess soil: on July 1, the average measured value was 81.56 cm/day, while this value was reduced to only 6.16 cm/day on August 6 (p. 64, Li, 1998). This points out that the conventional tillage activities on the field are of great importance for saturated hydraulic conductivity. Since the fieldwork experiments of our study were carried out during the end of March, all fields examined were fallow fields, with little to no vegetation. During this period, the land is most susceptible to water erosion, due to the lack of vegetation that the soil can retain. Concerning the fallow fields, Li (1998) found a saturated hydraulic conductivity of 200.19 cm/day (= 83 mm/hr.). This was a considerably higher number than the K values that followed from the mini-disk infiltrometer calculations (Appendix A: Location_Groesbeek/Processed_Data/Mini_Disk_Infiltrometer) and most mean infiltration values from the results section. These deviations can be explained by various other factors, since the texture is a fixed soil characteristic. Variable factors affecting infiltration such as aggregate stability, soil organic matter content (SOM), the type of crop grown and the number of minerals present in a profile could therefore also be taken into account in future research (Haghnazari et al., 2015).

Secondly, another study about soil erodibility in the Netherlands by Kamphorst (1987) demonstrated that aeolian sandy loam, aeolian loamy sand, and coarse loess had the highest runoff values compared to other, more compact textures, such as fine loess, riverine sandy loam, and riverine clay (Table 2, p. 413, Kamphorst, 1987). It was concluded that several texture groups, such as aeolean loamy sand, aeolean sandy loam and coarse loess did not differ significantly from each other in terms of surface runoff. In this particular case, this could also indicate similar infiltration values for the different soils, since high runoff is an indicator for low infiltration. Table 3 (p. 413, Kamphorst, 1987) demonstrates that texture group a (texture groups with sand and much loam) had the highest runoff values (indicating low infiltration), while texture group b (only sand, and (clay) loam) was below that. Van Es et al. (1991) confirmed this by finding a positive correlation between clay content and the infiltration rate, whereas the silt content had a negative correlation. Moreover, soil horizon thickness and texture may also have significant effects on spatial infiltration variability (Haghnazari et al., 2015).

The high infiltration rates in Hem for both the RS and the mini-disk infiltrometer can also be explained by the land use present. A study by Van Maanen et al. (2017) demonstrated that plowing and other agricultural management strategies disrupt the soil structure of the top layer. This causes, among other things, (coarse) loose aggregates, through which water flows very easily and cannot be retained. In addition, it was noted that the soil is becoming increasingly depleted by the excess application of fertilizers, especially that of manure (ibid). During the fieldwork, this was be seen by the dry soil composition, loose aggregates, and deep cracks that appeared on the surface. According to Haghnazari et al., (2015) some clayey soils develop shrinkage cracks as they dry, resulting in high infiltration rates under dry conditions. However, it should be noted that during long wet periods the soil pores become saturated with water, reducing both sorptivity and infiltration. Thus, under saturated conditions, clayey soils should have slow infiltration rates. Since the work was carried out on a dry day, it is still unknown what the infiltration values of this soil would be in a saturated situation.

(15)

15 Future research should be conducted to determine the infiltration rates for saturated clayey soils in this area.

4.2 Answer to research question and hypothesis

The hypothesis that sand would have the highest infiltration capacity, followed by loess and finally clay, was not confirmed by the obtained results. This can be due to unequal amounts of measurements per texture, a defective quantity of repetitions per method, and small errors during the fieldwork, which will be elaborated on in the next paragraph. In particular, the results of the inverse auger hole method could provide valuable data if a distinction is made between textures within a soil profile. The soils with loess as the top layer had a thick layer of coarse sand at the bottom of the profile, which led to extremely high infiltration rates of the total profile. These measurements should have been added separately to the texture sand for more realistic results, instead of being part of the texture loess. In addition, the well in Hem reached a less permeable clayey horizon, which is reflected in the results due to a relatively lower infiltration compared to the other two textures. Here too, distinctions between more compact layers will increase validity and reliability.

The research question: “What are the infiltration capacities of three different soil textures, being; 1) sand, 2) loess, and 3) clay?” is not covered completely by the outcome of the results. However, what is evident is the high temporal variation in infiltration, which is strongly influenced by land use and soil surface properties. The sandy soil had strong crusting on the surface, followed by a thick layer of sand, which explains why sand did not have the highest infiltration value since the water did not easily pass through that initial crust. As for the clayey soil, there were so many cracks and macropores that the situation was reversed and all the water disappeared immediately. Regarding loess, this texture had the greatest number of observations, leading to moderate results since any outliers are compensated by multiple other values.

4.3 Bottlenecks per method

Furthermore, it should be noted that each measurement method has its bottlenecks, which is why it is recommended for follow-up studies to evaluate cautiously which method best suits the ultimate research objective, soil type, and situation. Starting with the mini-disk infiltrometer, the amount of infiltration is quite accurate as water is in constant contact with the sample surface. However, it is less realistic, since it does not take into account the kinetic energy with which raindrops normally hit the ground. In addition, the mini-disk infiltrometer also lacks the effect of surface crusting due to mechanical disturbance and slaking (Cammeraat, 2005). Furthermore, the infiltrometer should not be used to estimate sorptivity, as previous studies showed considerable differences between this method and the rainfall simulator (ibid).

Secondly, the rainfall simulations have fewer disturbances and can better imitate a true-to-nature situation, since not all droplets will reach the soil surface and a larger sample area enables the observation of soil heterogeneity (Cammeraat, 2005). However, a disadvantage of the RS is that the infiltration strongly depends on the chosen rainfall intensity, something that would not occur in an actual situation, since the intensity can fluctuate during a shower. Another point of unreliability is the estimation of the ponding time (tp), as different parts of the sample surface in the ring can form poodles before reaching the saturation point.

With the inverse auger hole method, the decreasing water level in the well over time is assumed to be equal to the water flowing through the bottom and walls of the well. The disadvantage of this method is therefore that sorption is not taken into account, because the hole is filled up with water to the surface, resulting in no gradual wetting (Verbist et al., 2013). In addition, Verbist et al. (2013) concluded that borehole methods are not recommended for soils with high stoniness, because the construction of a borehole is limited by random rock fragments per field. Although the investigated

(16)

16 soils in Groesbeek were not officially stony, an impermeable layer was often hit with gravel / larger stones, confirming this complication.

4.4 Limitations during fieldwork

In addition to the unequal amounts of measurements per method and per texture, there were other inconsistencies, that could have been prevented. For example, with the mini-disk infiltrometer, the recommended time intervals between observations per texture from Decagon Devices (2016) were not taken into account. Furthermore, regarding the rainfall simulations, the runoff should have been recorded for another 30 minutes after reaching the constant point and we ran out of water before finishing one of the final repetitions (Cammeraat, 2005). Also, the pressure of both the mini-disk infiltrometer and the rain simulator was manually adjusted, allowing for small deviations between repetitions.

Finally, there were also some external circumstances beyond the control of the researchers that could have altered the results. These were on the one hand weather conditions such as a lot of wind, which prevented the precipitation from reaching the sample surface during the rainfall simulations. On the other hand, there were geomorphological differences between Groesbeek and Hem, such as the elevation of several fields in Groesbeek, contrasting with the plain field in Hem. At last, uneven surfaces caused measurement tools, such as the mini-disk infiltrometer, to not stay in place properly.

(17)

17

Conclusion

Although clay had the highest mean infiltration rates for the rainfall simulations, texture groups did not differ significantly from each other. Contrarily, the results from the mini-disk infiltrometer resulted in significant differences between sand, on the one hand, and loess and clay, on the other hand. Regarding the inverse auger hole method, practical mistakes during fieldwork produced invalid and unreliable data, which could be prevented by making a distinction between textures within one profile. The correlation coefficients differed in strength and direction per field. Field three and five had both a strong positive correlation. Field two had a moderate negative relation, while all other field values were close to zero, resulting in no significant correlation.

Despite the fact that results did not confirm the hypothesis, nor did they completely answer the research question, the overall analysis concluded that infiltration did have a high temporal variability. This was mainly caused by the different kinds of land use. As experiments of this research were conducted on fallow fields, the type of crop grown remained unknown, leaving a knowledge gap in the exact type of land use. The inclusion of more detailed arable activity can therefore provide further insight for future research. Additionally, several previous studies did not find significant differences between textures and unexpected infiltration patterns for different percentages of silt, sand and clay compositions as well. Variable factors such as SOM, aggregate stability and mineralization could thus offer supplementary information to explain infiltration patterns. Finally, it is recommended to go through the bottlenecks of each method and take extensive time for fieldwork, to avoid any inaccuracies.

(18)

18

References

Arshad, M. A., Schnitzer, M., Angers, D. A. and Ripmeester, J. R. 1990. Effects of till vs. no till on the quality of soil organic matter. Soil Biol. Biochem. 22: 595–599

Assouline, S. (2013). Infiltration into soils: Conceptual approaches and solutions. Water Resources

Research, 49(4), 1755-1772.

Batey, T. (2009). Soil compaction and soil management–a review. Soil use and management, 25(4), 335-345.

Berli, M., Carminati, A., Ghezzehei, T. A., & Or, D. (2008). Evolution of unsaturated hydraulic conductivity of aggregated soils due to compressive forces. Water Resources Research, 44(5).

Cammeraat, E., personal communication, March 18, (2021) Cammeraat, L.H. (ed), Recondes Field Protocol, vs. 1.0, (2005)

Carter, M. R. 1992. Characterizing the soil physical conditions in reduced tillage systems for winter wheat on a fine sandy loam using small cores. Can. J. Soil Sci. 72: 395–402

Cole, M. A., Elliott, R. J., & Strobl, E. (2014). Climate change, hydro-dependency, and the African dam boom. World Development, 60, 84-98.

Decagon Devices, Inc. Mini Disk Infiltrometer, 2016. Version: September 2, 2016 — 13:13:56.

Delta-T Devices Ltd. ThetaProbe SOIL MOISTURE SENSOR. TYPE ML2x USER MANUAL ML2x-UM-1.21, May 1999.

Elliott, J. A., & Efetha, A. A. (1999). Influence of tillage and cropping system on soil organic matter, structure and infiltration in a rolling landscape. Canadian Journal of Soil Science, 79(3), 457-463. de Haan, N. (2017). How governance changes affected the supply of ecosystem services in Berg en Dal. Haghnazari, F., Shahgholi, H., & Feizi, M. (2015). Factors affecting the infiltration of agricultural soils. International Journal of Agronomy and Agricultural Research, 6(5), 21-35.

IPCC, 2018: Global Warming of 1.5°C.An IPCC Special Report on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty [Masson-Delmotte, V., P. Zhai, H.-O. Pörtner, D. Roberts, J. Skea, P.R. Shukla, A. Pirani, W. Moufouma-Okia, C. Péan, R. Pidcock, S. Connors, J.B.R. Matthews, Y. Chen, X. Zhou, M.I. Gomis, E. Lonnoy, T. Maycock, M. Tignor, and T. Waterfield (eds.)]. In Press.

Kamphorst, A. (1987). A small rainfall simulator for the determination of soil erodibility. NJAS

wageningen journal of life sciences, 35(3), 407-415.

Li, T. (1998). Evaluation of land use systems for water and soil conservation in Groesbeek, The

Netherlands, using LISEM (Doctoral dissertation, Department of Theoretical Ecology, Wageningen

(19)

19 Mangala, O. S., Toppo, P., & Ghoshal, S. (2016). Study of infiltration capacity of different soils. International Journal of Trend in Research and Development, 3(2), 388-390.

Oxford Languages (n.d.) Retrieved from: https://languages.oup.com/google-dictionary-en/ (09-04-21) Ritzema, H. P. (1994). Drainage principles and applications (No. 16). Second Edition.

Stolte, J., C.J. Ritsema, T. Li. Invloed van verschillende landinrichtingsscenario 's op de bodemen waterafvoer in het zuidelijke deel van de ruilverkaveling Groesbeek, 1999. Wageningen, DLOStaring Centrum. Rapport 644. 54 blz. 24 fig.; 8 tab.; 10 ref.

Trenberth, K. E. (2011). Changes in precipitation with climate change. Climate Research, 47(1-2), 123-138.

Vaderstad, n.d. Soil Structure. Retrieved on 25-05-2021 from: https://www.vaderstad.com/en/know-how/basic-agronomy/soil-basics/soil-structure/

Van Es HM, Cassel DK, Daniels RB. 1991. Infiltration variability and correlations with surface soil properties for an eroded Hapludult." Soil Science Society of America Journal 55(2), 486-492.

Van Maanen, R., Morrien, E., & Cammeraat, E. (03-02-2020). Een vierjarig onderzoek naar het effect van niet onderwerken van strorijke stalmest in de bloemkoolteelt van Noord-Holland. Amsterdam Green Campus, duurzame innovatie in de regio.

Verbist, K. M. J., Cornelis, W. M., Torfs, S., & Gabriels, D. (2013). Comparing methods to determine hydraulic conductivities on stony soils. Soil Science Society of America Journal, 77(1), 25-42.

(20)

20

Acknowledgements

The learning process of this study has been very useful and entertaining. The independent fieldwork was made possible in part by the thesis supervisor Erik Cammeraat and all supplies were provided by the UvA. I would therefore like to thank both the institute, Erik Cammeraat and Anne Uilthoorn for the extensive guidance and feedback during the process. In addition, I would also like to express my appreciation for the municipality of Groesbeek and Waterschap Gemeente Berg & Dal, who stimulate research in this area. Finally, I also thank the local farmers who made it possible to carry out the experiments per study locations on their land properties.

(21)

21

How different soil textures influence the infiltration rate