Determining Factors Affecting the Soil Water Content and Yield of Selected Crops in a Field Experiment with a Rainout Shelter and a Control Plot in the Czech Republic

Abstract

To investigate the different responses of crops to drought stress under field conditions of Central European Climate for selected crop rotations, a field experiment was conducted at a test site in the Czech Republic from 2014 to 2021. Depending on the crop, rainout shelters were placed in late spring and early summer to study the effects of drought in the final stages of crop development. Due to these rainout shelters and the associated lower water availability for the crops during the summer, a reduction in leaf area index, biomass and yield was observed. For example, a yield decrease of more than 30% was observed for spring barley, winter rape and winter wheat compared to conditions without rainout shelters. The reduction was 25% and 18% for winter rye and silage maize, respectively, under rainout shelters. Soil moisture played a significant role in yield, where a predictive model based on monthly soil moisture explained up to 79% (winter rape) of the yield variance.

1. Introduction

In recent years, Europe has suffered from droughts, affecting agricultural production [1,2] and significantly impacting Central Europe. Drought can be defined as a natural hazard caused by deficient rainfall compared to the climatological normal. In agriculture, these can be understood as a deficit in soil moisture, usually in the root zone, which reduces the water supply to vegetation [3]. The impact of this condition on crop yield and quality depends on several factors, such as the timing of drought onset relative to the stage of crop development, the reliability of water sources, the vulnerability of individual crops and cultivars to drought stress, and socioeconomic factors [4]. For example, reduced leaf water potential, stomatal closure, reduced cell growth and enlargement are some effects of drought stress on crops [5]. The impairment of various physiological and biochemical functions, such as photosynthesis, respiration, nutrient metabolism, carbohydrate metabolism, chlorophyll synthesis, ion uptake and translocation, can reduce plant growth [5,6,7,8].

Thus, drought can directly impact crop growth and production through complex interactions with heat stress, soil properties and nutrient availability [9], and soil biota such as mycorrhizae [10]. While plants, for example, can cope well with moderate changes in total annual rainfall, increasing variability in rainfall per event and the timing of rainfall events during the growing season significantly impact aboveground biomass production [11,12,13,14]. High temperatures combined with low rainfall led to a negative feedback loop where high temperatures increase evapotranspiration in already stressed plants, resulting in a greater water deficit [15]. This deficit also reduces the leaf cooling of plants through reduced transpiration and affects their overall development [16]. Thus, drought stress can reduce wheat yields by up to 50% [17] due to the significant reduction in plant growth and shoot production [18,19]. In maize, yield losses of up to 40% have been observed, with a water reduction of approximately 40% [20].

Field experiments based on the manipulation of crop environments are essential to identify the crop response to the expected climatic conditions in the near future. To better understand ecological responses to natural droughts, they can be experimentally induced [21,22,23] and thus be used to study the effects of lower water availability on ecosystem processes. Rainout shelters are often used for this purpose. However, the main disadvantage of applying rainout shelters is the impact on the microclimate, such as altered soil and air temperature and reduced solar radiation, wind, and vapor pressure deficit for the sheltered variants [24,25,26,27]. Nonetheless, drought experiments help to provide essential insights into how different ecosystems respond to drought and what mechanisms control these responses [28].

To investigate the different responses of plants to drought stress under field conditions for selected crop rotations in Central Europe, a field experiment was conducted at an experimental site in the Czech Republic from 2014 to 2021. In crop production, the configuration and management of cropping systems play a crucial role in water and nutrient dynamics [29,30] and in the long-term development of soil carbon and nitrogen stocks and, thus, greenhouse gas emissions [31,32]. In this context, crop rotation is an important measure to adapt to climate change and mitigate its effects [33,34]. The selected crop rotation (winter oilseed rape, winter wheat, spring barley, winter rye and silage maize) is a common application in the Czech Republic and accounts for more than 60% of the arable land in the country. These crops are also an essential backbone of crop rotations in surrounding states. During the summer, rainout shelters were established on the experimental field to create drought stress for the cultivated crops and compare them with natural conditions.

The main objective of this study is to evaluate and analyze the impact of different water availability and drought scenarios (rainout shelter vs. natural environment) compared with observed weather conditions on the performance of the studied crops. The effect of drought stress is considered a limiting factor for crop production and is the key factor determining water-limited yield. However, few experiments aim to quantify the effects and mechanism of drought stress occurrence and its dynamics under field conditions and collect robust datasets to test and develop crop models in crop rotation settings. We hypothesize that rainfall exclusion during vulnerable phases of the crop growing season will reduce the water use by crops, decreases leaf area index (LAI) and above-ground biomass, and despite increased water use efficiency, will result in a significant yield decrease. The effects are expected to be most pronounced in spring-sown crops and in the case of winter crops, oil seed rape and wheat compared to rye.

2. Materials and Methods

2.1. Experimental Location

The field experiment was conducted at the Domanínek experimental station (49°31′42″ N, 16°14′13″ E, 560 m altitude), Czech Republic, located in the Bohemian-Moravian Highlands.

It is a relatively cool location with an annual mean temperature of 7.5 °C and annual total precipitation of 638 mm (climatological standard normal for 1991–2020) (

Table 1. Mean temperature [°C] and precipitation sum [mm] per month and year in Domanínek (1991–2020).

Month	Temperature [°C]	Precipitation [mm]
Jan	−2.6	44.2
Feb	−1.3	35.5
Mär	2.4	45.6
Apr	7.6	34.5
Mai	12.1	67.3
Jun	15.5	74.1
Jul	17.5	81.3
Aug	17.4	71.6
Sep	12.7	57.0
Okt	7.6	42.9
Nov	2.7	41.6
Dez	−1.7	42.4
Year	7.5	638.1

). Winter temperatures are often below 0 °C, and in the summer months of July and August, the average temperature is around 17.5 °C. A high potential risk of late frosts exists [35], and most precipitation falls from May to August.

The area is characterized by lower soil quality than fertile lowlands, and the soil type can be classified as dystric cambisol. Three soil profiles (Figure 1) were collected from the experimental site in August and September 2015 and analyzed (

Table 2. Soil characteristics of the soil samples from 6 August 2015 (soil 1) and 14 September 2015 (soil 2 and 3).

Soil	Depth	pH/H2O	pH/KCl	Cox (%)	hum (%)	bd (kg/m3)	p (%)	wp (%)	fc (%)	swc (%)	st Class
soil 1	5 cm	5.25	3.90	1.17	2.01	1520	44.12	11.4	26.1	39.5	loam
	15 cm	4.98	3.78	1.07	1.85	1559	42.67	11.3	25.5	38.2	loam
	25 cm	5.02	3.83	1.01	1.75	1539	43.84	11	24.9	38.9	loam
	35 cm	5.19	3.92	0.75	1.29	1654	39.64	9.8	23.9	35.2	loam
	45 cm	4.84	3.56	0.41	0.71	1589	42.44	11.6	23.1	37.6	sandy loam
	55 cm	4.75	3.60	0.29	0.50	1500		11.8	21.8	41	sandy loam
	65 cm	4.63	3.58	0.18	0.31	1440		12.6	25	43.2	loam
	75 cm	4.60	3.59	0.17	0.29	1500		11.6	22.1	41.1	sandy loam
	85 cm	4.58	3.55	0.09	0.16	1570		9.3	17.4	38.6	sandy loam
soil 2	5 cm	5.52	4.36	1.15	1.99	1472	45.89	11.4	25.1	39.5	loam
	15 cm	5.58	4.31	1.05	1.80	1572	42.22	11.3	24.6	38.2	loam
	25 cm	5.77	4.56	0.80	1.37	1523	44.01	11	22.6	38.9	loam
	35 cm	5.90	4.53	0.42	0.73	1424	47.85	9.8	24.7	35.2	loam
	45 cm	5.88	4.56	0.37	0.64	1422	48.66	11.6	26.2	37.6	silt loam
	55 cm	5.75	4.41	0.30	0.52	1320		10.8	25.5	47.5	silt loam
soil 3	5 cm	6.17	5.07	1.01	1.75	1566	43.27	12.6	30.2	37.8	silt loam
	15 cm	6.26	5.20	1.06	1.83	1572	43.04	12.7	30.5	37.8	silt loam
	25 cm	6.49	5.29	0.29	0.50	1669	39.52	11.7	28.6	34.9	silt loam
	35 cm	6.56	5.26	0.15	0.27	1731	38.16	12.9	29.1	32.8	silt loam
	45 cm	6.58	5.10	0.11	0.19	1742	37.34	13.7	29.9	32.5	silt loam
	55 cm	6.14	4.56	0.08	0.14	1340		12.9	27.5	46.9	silt loam
	65 cm	5.66	4.27	0.06	0.11	1550		12.1	21.5	39.4	sandy loam
	75 cm	5.57	4.13	0.10	0.17	1500		14.8	25.9	41.1	loam

Cox = soil organic carbon; hum = humus; db = bulk density; p = porosity; wp = wilting point; fc = field capacity; swc = saturated soil water content; st class = soil textural class.

).

2.2. Experimental Design

The field experiment (Figure 2) was conducted on an area of 1.357 ha (115 m × 118 m), laid out as a strip plot design. Two rows were grown side by side in parallel, each with six plots of 480 m² (40 m × 12 m). The distance between the individual plots was 6 m within the same row, and the space was between 8 and 10 m to the edge. A distance of 10 m separated the two rows. This study used six plots (A–F) on the south side (Figure 2a).

From 2014/2015 to 2020/2021, the following crop rotation was cultivated on the six plots: rape, winter wheat, spring barley, silage maize, and legume mixture. Initially, winter wheat was cultivated twice and replaced by winter rye from the 2017/2018 season onward (Figure 3). In some cases, different varieties were selected (see Appendix A); winter oilseed rape was sown in spring rather than autumn in 2016/2017 and 2018/2019 (

Table 3. Overview of the grown crops, sowing and harvesting dates, and days of the growing period [d] for the six plots.

Crop	Plot A	Plot B	Plot C	Plot D	Plot E	Plot F
Spring barley	11/04/2017–24/08/2017 135d	05/04/2016–15/08/2016 132d	26/03/2015–06/08/2015 133d 12/04/2021–18/08/2021 128d	08/04/2020–20/08/2020 134d	05/04/2019–12/08/2019 129d	12/04/2018–02/08/2018 112d
Silage maize	02/05/2019–01/10/2019 152d	03/05/2018–05/09/2018 125d	10/05/2017–09/10/2017 152d	11/05/2016–29/09/2016 141d	05/05/2015–30/09/2015 148d 30/04/2021–12/10/2021 165	27/04/2020–06/10/2020 162d
Winter rape	25/08/2014–31/07/2015 340d 25/08/2020–09/08/2021 349d	23/08/2019–30/07/2020 342d		22/08/2017–17/07/2018 329d		26/08/2015–26/07/2016 335d
Summer rape			11/04/2019–07/08/2019 118d		11/04/2017–24/08/2017 135d
Winter rye	02/10/2017–26/07/2018 297d			24/09/2020–13/08/2021 323d	18/09/2019–21/08/2020 338d	27/09/2018–02/08/2019 309d
Winter wheat	25/09/2015–26/07/2016 305d	30/09/2014–06/08/2015 310d 30/09/2016–17/08/2017 321d 06/10/2020–13/08/2021 311d	25/09/2015–13/08/2016 323d 04/10/2019–13/08/2020 314d	30/09/2014–06/08/2015 310d 01/10/2018–02/08/2019 305d	07/10/2017–26/07/2018 292d	30/09/2016–17/08/2017 321d

). The previous crop of the experimental area was oats in 2014, and the plots were uniformly fertilized according to the methods of the Central Institute for Monitoring and Testing in Agriculture in Domanínek. More detailed information on management, such as tillage, fertilizer timing and dosage, can be found in Appendix A.

Each plot was set up for a field experiment in two variants. The first variant was carried out under natural climatic conditions (referred to as “control”, plot size of 1.5 m × 8.0 m); in the second one, drought stress was induced using rain protection devices (referred to as “shelter” plot size of 3.1 m × 8 m). Each variation was replicated three times. The position of the shelters was recorded through precise GPS coordinates and marked in the terrain to ensure that the positions of the control and rain-excluded plots did not differ. The rainout shelters (3.1 m × 8 m) comprised two connected flexible and mobile parts (3.1 m × 4 m). Each segment weighed approximately 100 kg and had wheels, allowing them to be easily moved by two people without damaging the crops. The height of the shelter was adjusted to the height of the vegetation. Between May and August (

Table 4. Overview of the crops grown, date of installation and removal of the shelters, and number of sheltered days (d) for the six plots.

Crop	Plot A	Plot B	Plot C	Plot D	Plot E	Plot F
Spring barley	15/06–24/08/2017 70d	23/05–15/08/2016 84d	11/06–18/08/2021 68d	03/06–20/08/2020 78d	30/05–12/08/2019 74d	11/06–02/08/2018 52d
Silage maize	20/06–30/07/2019 40d	27/06–27/07/2018 30d	21/06–01/08/2017 41d	23/05–03/08/2016 72d	30/06–29/07/2021 29d	23/06–28/07/2020 35d
Winter rape	12/05–09/08/2021 89d	24/04–30/07/2020 97d		03/05–12/07/2018 70d		21/04–26/07/2016 96d
Summer rape			06/06–07/08/2019 62d		10/06–24/08/2017 75d
Winter rye	15/05–26/07/2018 72d			12/05–13/08/2021 93d	24/04–21/08/2020 119d	30/04–02/08/2019 94d
Winter wheat	15/04–26/07/2016 102d	15/05–06/08/2015 83 15/05–17/08/2017 94d 24/05–13/08/2021 81d	14/05–13/08/2016 91d 14/05–13/08/2020 91d	07/05–02/08/2019 87d	25/05–26/07/2018 62d	12/05–17/08/2017 97d

), these mobile rainout shelters were installed on each plot. A corrugated material (Suntuf CS—clear polycarbonate with a UV filter on both sides; trapezoid 76/16 and thickness 0.8 mm) was used to make the shelters [36]. In the middle of each covered plot, a 1.5 m wide strip was harvested to eliminate the edge effect. Thus, the protected area was large enough to take the samples needed to determine the growth parameters and soil moisture [35]. The water from the shelters was collected through a system of gutters and through pipes that were channeled outside the experimental area. Access paths [~0.25 m wide] were created through mechanical weeders between the individual plots (Figure 4).

2.3. Data Collection and Monitoring

Adjacent to the experimental field, an agrometeorological weather station was set up over a grass plot, measuring in 10 s intervals and collecting 10-min averages/sums/extremes of 2 m air temperature, relative humidity, precipitation, total radiation, and wind speed. Additionally, under the shelters and in the control, plots were sensors (Minikin THi, EMS Brno, The Czech Republic) that measured the temperature and humidity of the air. The sensors were always at the current height of the shelters. Soil moisture was measured at a 0–30 cm depth in the different rainout shelters and control plots using TDR (Time Domain Reflectometry, CS 616 Campbell Scientific Inc., Logan, Utah, USA) sensors during the sheltered period (Figure 4a). Two to four TDR sensors were placed under each variant in the center of the harvested strip. By measuring the soil water content, it was possible to confirm the reduction in precipitation under the rainout shelters.

Phenology (emergence, tillering, shooting, heading, flowering, and maturity) and management (tillage, construction of rainout shelters, fertilization) were recorded for the different plots. The LAI was measured with a SunScan instrument (Delta-T Devices, Cambridge, UK) during the vegetation period. Aboveground biomass and soil moisture were sampled on 1/3 of a 1.5 m strip at least six times during the season for the spring crops and nine times for winter crops.

2.4. Statistical Data Analysis

Initially, the measured and observed values were analyzed using descriptive statistics in R statistics software, such as the mean, standard deviation, percentiles, minimum and maximum. This allowed the information in the sample data to be summarized in tables, charts, and statistical measures clearly and unbiasedly. Bivariate analysis was performed using a one-way variance analysis (one-way ANOVA) after testing the normal distribution and Person correlation coefficient (Pearson’s r). The various measurements (precipitation, soil moisture, yields, LAI, etc.) of the two studied variants, control, and shelter, were tested for linear correlation. With the help of linear regression analysis (r²), the effect of drought on the yields of each crop under different natural and weather conditions was determined.

3. Results

3.1. Precipitation

Precipitation variability (i.e., anomalies), defined as deviations in annual rainfall from long-term averages [37], helps to classify a year as dry or wet. Figure 5 shows the monthly precipitation anomalies during the study period compared to the normal period (1991–2020). Notably, in 2015 and 2018, precipitation during the vegetation period was below the long-term average, while higher precipitation was recorded in 2020 and 2021. At the monthly scale, differences in precipitation totals are pronounced and indicate strong with-in seasonal variability, which may affect specific crop drought stress vulnerability at the various phenological phases.

At the experimental site, crops were exposed to a water deficit between May and August (see Table 4 for exact timing) using mobile rainout shelters, and these plots were then compared to the control plots exposed to natural rainfall. Figure 6 summarizes (top) the amount of precipitation (mm) from sowing to harvest for each crop under natural conditions (control) and the lower amount of precipitation (mm) due to the presence of the rainout shelters (shelter) for each year, as well as (bottom) the distribution of the yearly total rainfall over the entire investigation period.

The summer crops of spring barley and silage maize had the greatest differences in precipitation amounts—over 68% more rainfall was available for the crops during the growing season under the control conditions. For maize, this was most pronounced in 2016 and 2017, and for spring barley, it was most evident in 2020 (>80% difference) (Figure 6 above). Rape, planted in the spring of 2017 and 2019, showed an average of 58% less precipitation under the sheltered conditions in the two years. Due to the longer growing season (despite the fact that precipitation during the winter was lower than in the summer; see Table 1), the differences were not as strong for winter crops, reaching an average of 45% lower rainfall totals under the rainout shelter conditions for winter wheat and winter rape and −55% for winter rye (Figure 6 bottom).

3.2. Soil Moisture

The measured soil moisture at a soil depth of 0 to 30 cm is compared in Figure 7 during the sheltered period. It is shown that soil moisture averaged between 15 (summer rape) and 22 (maize) vol.% under natural conditions, while it was significantly lower at 10 (summer and winter rape, winter rye, winter wheat) to 18 (maize) vol.% for the sheltered variants. The control and sheltered conditions were significantly different from the monthly mean soil moisture (ANOVA p-value > 0.001), except for silage maize (ANOVA p-value = 0.32). Silage maize had the most soil water available in the summer months but also showed the highest standard deviation. The mean differences between the two variants were the smallest, whereas winter rye showed the largest differences.

The soil moisture under natural conditions and shelters will be considered individually for each crop.

3.2.1. Soil Moisture: Spring Barley

Spring barley was cultivated in the experimental field over a period of six years, and mobile rainout shelters were used to cover the plots for an average of 66 days between June and August (the longest period was 80 days in 2016, and the shortest period was 51 days in 2017, Table 4). In 2018 and 2019, the difference in soil moisture between the sheltered and control plots during the investigated period was approximately 15% due to very low rainfall in spring and summer (Figure 6), and in 2017 and 2020, the differences were over 50% (Figure 8). 2017 the shelters were installed on 15 June 2017, but soil measurements did not begin until 28 June 2017. Thus, in 2017, the soil water content under the rainout shelters was significantly lower at the beginning than under the control variant.

3.2.2. Soil Moisture: Silage Maize

In the case of silage maize, the mobile rainout shelters were set up for the shortest period, with 43 days on average in the five years studied, mainly in June and July. Thus, the soil moisture was approximately 14% lower than that under natural conditions during this period. The largest deviation was found in 2018, with a difference of 28% (Figure 9).

3.2.3. Soil Moisture: Oilseed Rape

The sheltered period of winter oilseed rape averaged 88 days in the four years surveyed, 2016, 2018, 2020 and 2021, and approximately 69 days for summer oilseed rape (2017, 2019). Comparable to spring barley but slightly more pronounced, the highest differences in soil moisture were measured in 2020 and 2021 at over 60%, while 2018 and 2019 showed approximately 30% lower soil moisture under the sheltered conditions. In 2016 and 2017, the shelters were installed a few days before soil moisture measurements began (2016 = 15 d; 2017 = 6 d), which explains the higher values under natural conditions at the beginning of the plot (Figure 10).

3.2.4. Soil Moisture: Winter Rye

Winter rye was cultivated during four growing seasons. The effects of drought due to rainout shelters were tested on an average of 86 days. During this period, soil moisture under the sheltered conditions was approximately 50% less than under the control conditions (Figure 11). Soil moisture levels at 0–30 cm depths were very low in 2018, averaging 9 vol.% and 5 vol.% under natural and rain-protected conditions, respectively.

3.2.5. Soil Moisture: Winter Wheat

Winter wheat had the longest investigation period of seven years. On average, rainout shelters were set up for 84 days at the end of the growing season, showing a mean soil moisture reduction of 37%. In particular, 2016 and 2020 presented large differences of up to 50%. In 2015, 2018 and 2021, when the precipitation anomaly was negative during winter and spring (Figure 5), deviations of less than 30% were measured (Figure 12). In the first three years, soil moisture was measured after shelter installation (2015 = 3 d, 2016 = 15 d, 2017 = 10 d).

3.3. Leaf Area Index

The LAI measures half of the total green leaf area per unit of horizontal ground area. It forms an important structural property of vegetation. As leaf surfaces are the primary limit for mass exchange and energy, key processes such as evapotranspiration, canopy screening or gross photosynthesis are directly proportional to the LAI [38].

For each plot, the LAI was measured several times (2 to 11 times) during the sheltered period under natural and sheltered conditions. The largest variation between the control and sheltered conditions was found for spring barley (−18%) and winter oilseed rape (−17%), while the LAI difference was minimal for silage maize (−3%).

Looking at the LAI values in more detail (

Table 5. LAI for the control vs. sheltered conditions of spring barley (2016–2021), silage maize (2016–2021), summer rape (2017, 2019), winter rape (2016, 2018, 2020, 2021), winter rye (2018–2021) and winter wheat (2015–2021): mean, standard deviation (sd), variance (var), maximum (max), minimum (min), and percentiles (per; 10, 25, 75 and 90%).

Crop	Cond	Mean	Sd	Var	Max	Min	Perc 10	Perc 25	Perc 75	Perc 90
spring barley	control	2.5	0.5	0.3	3.9	1.3	2.0	2.2	2.8	3.1
	shelter	2.1	0.5	0.2	3.2	1.0	1.7	1.8	2.3	2.6
sillage maize	control	2.1	0.7	0.5	3.4	0.8	1.1	1.6	2.5	3.0
	shelter	2.0	0.8	0.6	3.4	0.9	1.2	1.5	2.5	2.9
summer rape	control	2.6	0.9	0.8	4.5	1.4	1.7	2.1	3.2	3.4
	shelter	2.5	0.8	0.7	4.1	0.9	1.7	2.0	2.9	3.3
winter rape	control	3.4	1.3	1.8	6.6	0.7	2.0	2.6	4.0	5.3
	shelter	2.8	1.5	2.2	6.6	0.8	1.7	1.8	3.4	5.3
winter rye	control	2.4	0.5	0.2	3.5	1.3	1.9	2.1	2.7	2.9
	shelter	2.2	0.5	0.2	3.1	1.0	1.7	1.8	2.4	2.8
winter wheat	control	3.2	1.2	1.4	6.0	0.8	1.9	2.3	3.9	4.9
	shelter	2.8	1.2	1.5	5.6	1.0	1.6	2.0	3.6	4.9
all crops	control	2.8	1.0	1.1	6.6	0.7	1.7	2.1	3.3	4.0
	shelter	2.4	1.0	1.0	6.6	0.8	1.5	1.8	2.8	3.6

), the two summer crops and winter rye had the lowest mean values and standard deviations. Meanwhile, winter wheat and rape showed LAI values exceeding six and a significantly higher standard deviation.

The relationship between the sheltered and non-sheltered LAI values was very high, as seen from the scatter plots in Figure 13 (r² > 0.65). The LAI values were measured for each crop under the two studied conditions, and the values per year are shown with various colors.

3.4. Crop Yield

Considering the yield variations between natural and sheltered conditions from 2015 to 2021, an apparent decrease in yield with less rainfall can be seen (Figure 14). In 2015, shelters were installed only on the winter wheat plots and were included in the analysis. Winter rape did not emerge well in 2017. As a result, summer rape was sown instead of winter rape. This also failed to emerge well, and the canopy was uneven. The results were thus removed from the analysis. On 10 August 2021, before harvest, there were yield failures due to hail for spring barley, winter rye and winter wheat under the control condition; the vegetation under the shelters was not damaged by hail. Excluding these years, on average, a 30% yield reduction in spring barley, 18% in silage maize, 39% in winter oilseed rape, 25% in winter rye and 31% in winter wheat was calculated for the plots with rainout shelters. For summer oilseed rape, data for only one year were available (2019), with a yield reduction of 39%.

3.5. Comparison of the Different Measured Values

3.5.1. Precipitation and Soil Moisture

A Pearson correlation analysis was used to test the correlation between meteorological factors and soil moisture (

Table 6. Pearson’s correlation analysis for monthly mean soil moisture, total precipitation and mean temperature in the summer months June and July, for summer crops also August.

Crop	Soil Moisture vs.	Pearson’s r
spring barley	precipitation	0.63 *
	temperature	−0.19
silage maize	precipitation	0.67 *
	temperature	−0.6 *
summer rape	precipitation	0.34
	temperature	0
winter rape	precipitation	0.67 *
	temperature	0.05
winter rye	precipitation	0.71 *
	temperature	0.09
winter wheat	precipitation	0.71 *
	temperature	−0.17

* p value < 0.001.

). While soil moisture was strongly positively correlated with precipitation (from 0.63 to 0.71, except for summer rape), a significant negative correlation between soil moisture and temperature was found only for maize (r = −0.6). No significant relation was detected between the last two variables mentioned.

The scatter plots in Figure 15 show the monthly mean soil moisture and total precipitation during the summer months of June and July, for summer crops also August, under the control and sheltered conditions. The regression analysis was significant across all crops (except summer rape), and the highest r² was found for winter wheat and rye (r² ≥ 0.5).

3.5.2. Leaf Area Index and Soil Moisture

Examining the relation between the monthly summer LAI values and soil moisture, a positive correlation was found for spring barley, summer (not sig.) and winter rape, winter rye (not sig.) and winter wheat (not sig.) (

Table 7. Pearson’s product-moment correlation between the LAI and SM for June and July, for summer crops, August, and the combination of all summer months.

Crop	June	July	August	Summer +
spring barley	0.24	0.15		0.50 *
silage maize	−0.06	−0.01	−0.01	−0.47
summer rape	0.2	0.11	0.26	0.49
winter rape	0.18	0.21		0.61 *
winter rye	0.11	0.09		0.53
winter wheat	0.15	0.04		0.23

⁺ Winter crops = June and July, sommer crops = June, July, and August; * p-value < 0.05.

, Figure 16). For winter crops, August was omitted from the analysis because maturity occurred in late July to mid-August (Table 3).

The behavior of silage maize, however, was different. A negative correlation (not sig.) was observed, which was attributed to the fact that maize was still in the early period of the growing season when the available soil water reserves were still at a high level in June (see Figure 9) in all years; the other crops already claimed most of the available soil water in June during dry spring seasons. Furthermore, increasing the rooting depth of maize beyond 30 cm soil depth during that period allows significant water stress to be avoided [39]. While spring barley, winter rape, winter rye and winter wheat mature in July and August, silage maize actively grows until September and October.

3.5.3. Ratio between Control and Shelter of Leaf Area Index, Biomass and Yield

An overview of the maximum measured LAI (LAI_max), total above-ground biomass and yield of the individual crops, non-stressed and stressed, for those years unaffected by, e.g., hail is given in Figure 17.

The ratio between non-stressed and stressed (control:protection) crops for LAI_max, biomass, and yield are summarized in

Table 8. The ratio between control and shelter of LAI_max, above-ground biomass and yield for spring barley (2016–2020), silage maize (2016–2020), winter rape (2016, 2018, 2020, 2021), winter rye (2018–2020) and winter wheat (2015–2020).

Crop	LAImax	Biomass	Yield
Spring barley	1.2	1.1	1.4
Silage maize	1.3	1.2	1.2
Winter rape	1.2	1.3	1.6
Winter rye	1.1	1.1	1.3
Winter wheat	1.1	1.4	1.4

. The highest ratios are observed for yield in all crops except maize. All three ratios of silage maize, LAI_max, biomass and yield, show comparable values (1.2–1.3). The yield of winter rape stands out with a ratio of 1.6, but winter wheat and spring barley have high ratios of 1.4.

3.5.4. Parameters Affecting the Yield

The differences in the yields of the individual crops can be attributed to water availability as a main limiting variable and other related factors. In our case, we measured the influence on yield by selected potential predictors with the help of correlation and linear regression analyses. Monthly and total precipitation over the growing season, monthly average soil moisture data in the first 0–30 cm and LAI values were used as independent variables.

Except for winter rye, all crops showed a high positive correlation between yield and soil moisture in June and July. The LAI in July showed a significant positive yield impact for spring barley, winter oilseed rape and winter rye, while rainfall also played a significant role for spring barley and winter rape (

Table 9. Pearson’s product-moment correlation for yield vs. total precipitation (vegetation period, June, July, August * and summer months), mean soil moisture (June, July, August * and summer months), and mean LAI (June, July, August * and summer months) for spring barley, silage maize, winter rape, winter rye and winter wheat; * only for summer crops.

Yield	Spring Barley	Silage Maize	Winter Rape	Winter Rye	Winter Wheat
Precipitation vegetation period	0.69 *	0.37	0.41	0.79	0.38
Precipitation June	0.60	0.25	0.74 *	0.72	0.55
Precipitation July	0.37	0.47	0.84 **	0.68	0.32
Precipitation August		0.33
Precipitation summer month	0.51	0.46	0.85 *	0.71	0.48
Soil moisture June	0.87 **	0.85 **	0.82 *	0.65	0.63 *
Soil moisture July	0.67 *	0.85 **	0.88 *	0.72	0.51
Soil moisture August	0.62	-	-	-	0.57
Soil moisture summer month	0.81 **	0.83 *	0.85 *	0.72	0.59 *
LAI June	0.65	−0.74	0.78	0.97 *	0.26
LAI July	0.79 **	0.01	0.76 *	0.9 *	0.41
LAI August		−0.15	-
LAI summer month	0.74 *	0.12	0.61	0.88 *	0.41

** p-value < 0.01; * p-value < 0.05.

).

To extend the analysis, a regression analysis per crop was performed. Soil moisture in June was the most important variable for spring barley (r² = 0.75), and winter wheat (r² = 0.4), and soil moisture in July was the most important variable for silage maize (r² = 0.72) and winter rape (r² = 0.78). For winter rye, the LAI in June, which can also be determined by available water, was the best predictor (r² = 0.94) (Figure 18). However, only three years, i.e., six values, were available for the analysis. Summer rape was not used in the analysis, as only two years were available, one of which failed to emerge.

4. Discussion

Given the climate expected in the near future, a more frequent summer drought will occur in Europe [2,40], one of the major yield-limiting climate factors [41]. Drought is an abiotic factor that causes significant yield losses, especially in rain-fed agriculture. A deficit in soil moisture, usually in the root zone, reduces the water supply to vegetation. The impact of this condition on crop yield and quality depends on several factors, such as timing, duration and intensity, which make it very difficult to apply simpler screening tools and selection procedures to develop drought-resistant genotypes. Field experiments based on the manipulation of crop environments are critical to determining crop responses to climate conditions expected in the near future. To better understand ecological responses to natural droughts, these responses can be experimentally induced and thus be used to study the effects of reduced water availability on ecosystem processes. For example, rainout shelters at experimental sites, where drought-like conditions are established, can be used to study the effects of water stress on crops. This involves reducing the amount of precipitation available to the crop to induce water stress and observing the resulting changes in crop growth, yield, and other physiological traits.

To investigate the different responses of plants to drought stress under field conditions for selected crop rotations, a field experiment was started at an experimental site in the Czech Republic in 2014. The results for 2014–2021 are introduced in this study, with the experiment envisaged to continue for three full crop rotations, if possible, i.e., ~19 years. This crop rotation experiment in Central Europe includes current crops and can be considered in more detail for this specific climate zone. Thus, depending on the crop, rainout shelters were placed in late spring and early summer to study the effects of drought in the final stages of crop development.

Lower LAI, biomass, and yield were noted for all crops due to these rainout shelters and lower water availability during the summer months. The ratio between control and shelter LAI_max ranges from 1.1 to 1.3 and is most pronounced for silage maize. For above-ground biomass, the ratio is strongest in winter rape and wheat, with 1.3 and 1.4, respectively. Looking at yield, spring barley, winter rape and winter wheat show a more than 30% yield reduction. Here, winter rape, in particular, shows a very high value with a ratio of 1.6. For winter rye, the reduction was 25% and for silage maize 18%. The smallest decrease in maize is since the rain shelters were mainly placed in July and had the fewest days with shelter on the plot. However, the growing season of this crop extends through September and October, but the artificial drought stress conditions were in the early part of the growing season when available soil water reserves were still high in June. Maize also had the least difference between natural and rain-protected conditions, with about 14% less soil moisture. Looking more closely at the measured and observed values, soil moisture significantly affected yield. Regression analysis created a predictive model based on monthly soil moisture that explained up to 79% (winter rape) of the yield variance.

However, it must be kept in mind that permanent cover by the rainout shelter has strong effects on microclimate factors and, therefore, on the water balance of plants. They may also limit the amount of light reaching crops, impacting photosynthesis and crop yield [42]. To take this influence into account, sensors were installed under the shelters and in the control plots to measure temperature and humidity. The sensors were located at the current height of the shelters and showed no differences on average. However, situations affecting the experiment, such as reduced windspeed or the canopy shading, could not be addressed. Therefore, it is unclear whether the shaded variants benefit from the roofing compared to the unshaded variants. Despite these sources of error, overall, rainout shelters can help researchers and farmers better understand the effects of drought and water stress on crops and develop strategies to mitigate the negative impacts. For example, controlled experiments with water stress conditions enable the development of crop breeding programs and should help improve crop resilience to future drought events. It can also help farmers make more informed decisions about crop management and how to deal with environmental risks.

5. Conclusions

Summer drought, which is expected to increase in Central Europe in the near future, significantly impacts the yields of various crops. To investigate the different responses of plants to drought stress under field conditions for selected crops in this Central European climate zone, a field trial was conducted at an experimental site in the Czech Republic over six years (2014/2015 to 2020/2021). In the process, we hypothesized that a failure to receive precipitation during the critical part of the growing season would result in reduced crop water use, LAI, and above-ground biomass, as well as a significant reduction in yield despite increased water use efficiency. In this regard, the effects should be most pronounced for spring-sown crops (i.e., spring barley and silage corn) and rape and wheat compared to rye. Our study essentially confirmed our hypothesis with the exception of silage maize, where the results have been less marked due to shorter drought exposure and the better ability of C4 species to withstand shorter droughts.

Soil moisture played a crucial role in the experimental results. Regression analysis created a predictive model based on monthly soil moisture that explained up to 79% (winter rape) of the yield variance. However, it should be mentioned that shelter effects on canopy microclimate could not be accounted for in the analysis. The experimental plot is still in operation so that the new data can be used for further evaluation.