Assessment of Suitability Area for Maize Production in Poland Related to the Climate Change and Water Stress

Abstract

In this study, we identify the spatial distribution of water deficits in Poland. The analyses considered expert knowledge in soil categories importance in water stress evaluation influencing the climate suitability for maize production using the analytical hierarchy process (AHP). The Climatic Water Balance was calculated from April to September, for the baseline (BL) period (1981–2010) and two future periods of 2041–2070 (2050s) and 2071–2100 (2080s) using a six-member ensemble of GCM-RCM chain simulations under two representative concentration pathways (RCP) scenarios: low emissions (LE, RCP4.5) and high emissions (HE, RCP8.5). Taking into consideration water deficiency for the BL period, about 81% of Poland proved highly suitable, 18% moderately suitable and 1% marginally suitable for maize cultivation. According to LE and HE scenarios, the area of Poland that is highly suitable for maize production would decrease to 67 and 69% by the 2050s, and to 64 and 44% by the 2080s. By the 2080s, under the HE scenario, rain-fed maize production would become risky, as 21% of Poland would be marginally suitable, while 11% would not be suitable. According to our findings, supplemental irrigation is one of the effective adaptation strategies to maintain the production potential of maize in Poland.

1. Introduction

In Poland, the area under grain maize has increased significantly over the last 30 years, from 56 thousand ha in 1992 to 998 thousand ha in 2021, accounting for 11% of the total grain maize area in the EU [1]. The average grain maize yield over the last 30 years (1992–2021) was 5.9 t per 1 ha with a stable yield increase trend (0.7 t/10 years) but characterised by relatively high yield variability from year to year [2], from 3.7 t in 1992 to 7.3 t in 2012 [1].

Previous studies on land suitability for maize cultivation in Poland have mainly focused on thermal conditions [3]; however, in recent years, there has also been growing interest in other factors, mainly drought [2]. Climate change, with projected temperature increases, will increase the number of years with extremely unfavourable drought conditions, which might result in higher interannual yield variability and constitute a challenge for crop management [4]. In view of the increasing risk associated with drought and heat stress, the suitable production area for certain crops will change [5]. Previous studies have highlighted the need for adaptation measures to optimise water use on arable land [6]. The introduction of climate adaptation measures is necessary due to the predicted increase of droughts in Central Europe in the future [7]. Our study contributes to increasing the knowledge of adaptation practices in maize cultivation in the face of climate change. The methodology used in this work will be applied for further research aimed at assessing the suitability of land for maize cultivation in Poland, considering other climate change hazards such as spring frost stress, heat stress, etc.

In the context of climate change, it is essential to assess shifts in land suitability for crop production to minimise crop production risks [8]. Land suitability assessments may also help to indicate proper adaptation strategies to optimise agricultural production for improving the resilience of farming systems to climate change [9,10]. Land suitability assessment is a technique used to determine how suitable the land is for growing a given crop in a given region [11]. It involves the combination of several parameters, such as topographic, soil, climatic and management parameters, and is, therefore, a multi-dimensional problem [12]. Various multi-criteria evaluation techniques have been used to address crop suitability, such as Analytical Hierarchy Process (AHP) [12,13], Fuzzy Analytical Hierarchy Process (FAHP) [14,15,16], Analytic Network Process (ANP) [14], Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) [17,18] and ELimination Et Choice Translating REality (ELECTRE) [19] integrated with remote sensing and Geographical Information System [11]. Some studies [14,20,21,22,23] have carried out a comparative analysis in land suitability assessment. Seyedmohammadi et al. [14] compared two methods (square root and fuzzy set) for land suitability assessment for irrigated maize cultivation. Rodcha et al. [20] compared parametric, classical AHP, and fuzzy AHP (FAP) approaches in the suitability assessment of cash crops. Elaalem [21] compared Parametric and Fuzzy Multicriteria Methods to evaluate land suitability for olive production. Kalichkin et al. [22] obtained the criteria weights based on AHP and the direct ranking method in land suitability evaluation for spring wheat. Elaalem [23] compared two approaches, Boolean and Fuzzy MCDA for land suitability evaluation of sorghum production.

One of the challenges in water deficiency assessment for crops is considering the dependence of crop water requirements and different characteristics of soil retention related to granulometric composition according to soil categories [24,25]. Such analyses are conducted as part of the sustainability assessment of production by determining the criteria weights and can also be applied in the analysis of the suitability of areas. There are various methods that can be used to determine the criteria weights, one of which is the analytical hierarchy process (AHP) [13,26]. The AHP has been used to determine the weights of the parameters in the land suitability assessment for maize [13,27,28,29] and other crops such as tea [30,31], rice and soybean [29]. The AHP has the advantage of incorporating experts’ opinion to prioritise the criteria [27]. Another advantage over other methods is the measurement and control of judgement inconsistency. One of the drawbacks of the AHP methodology is that it involves a subjective judgement of the relative importance of two criteria being compared, while there is a possibility of omitting interrelationships between other criteria [13].

As climate change may impact the land suitability for crop production, in some studies climate change projections with adaptation options were considered. According to an analysis in Europe, the most important factor for cultivated crops is the spatially distributed water availability from rainfall [32]; however, an increasing risk of late frost will also be very important [33]. Lopez-Blanco et al. [9] identified suitable areas for rainfed maize production in Mexico for current and future climate conditions projected under different climate change scenarios. Jia et al. [34] evaluated the suitable areas for maize cultivation in China and their change under future climate conditions. Tan et al. [10] evaluated suitable areas for maize and wheat in Ethiopia under current and future climate conditions using the GIS-MCE Planting Ecological Adaptability model.

In the last years, there has been a tendency for the growing intensity of the agricultural drought in Poland [35]. While assessing drought risk, it is important to distinguish between the different soil categories, crops and geographic regions [24]. Wójcik et al. [35] determined Central Poland (among others the Dolnośląskie Voivodeship) as an area with the highest frequencies of agricultural droughts in spring cereals, while in the Małopolskie Voivodeship drought occurred only twice. Łabędzki et al. [36] noted that the tendency of water shortage for crops will increase with the foreseen climate warming. Somorowska [37] indicated the increase of drought severity in relatively large area extending from the south-west to Central Poland over the 60-year period that was analysed. Among the adaptation strategies proposed to minimise possible climate related yield losses, there are various adaptation practices, including changing the timing of cultivation, introducing irrigation, selecting varieties with both a shorter and a longer maturity [38].

Therefore, this study attempts to analyse the distribution of water deficiency parameters that influence climate suitability for maize production in Poland using the analytical hierarchy process (AHP). The study considers irrigation as an agricultural practice for drought risk reduction. The next section presents the description of the study area and the climate data used in the study. The methodology for water condition evaluation and land suitability assessment is also presented. Section 3 shows temperature and water conditions in the current and future climate, as well as the results of the assessment of the climatic water balance in Poland. At the end of this section, the results of the sustainability assessment considering water deficiency on different soil categories are presented with testing the addition of different amounts of water supply through irrigation. Section 4 presents a discussion of the results obtained, which are summarised in Section 5.

2. Materials and Methods

2.1. Study Area

Poland is situated in a transitional warm temperate climate zone. A mild oceanic climate and a dry continental climate have an influence on the Polish climate. The mean annual air temperature varies from 6 to 9 °C and decreases from south-west to north-east [36,39]. The growing season in Poland varies from 180 to 225 days while in the mountains it is up to 150 days [40]. In Poland, there is spatial and temporal variability of precipitation [41]. The annual precipitation sum is about 550–650 mm in Central Poland, 650–850 mm in Northern Poland, 700–900 mm in Southern Poland and more than 1000 mm in the mountains [40]. Natural conditions in Poland are rather favourable for agriculture, but droughts may appear every 4–5 years and become an important risk factor [42].

2.2. Climate Data

The gridded climate data was derived from the regional downscaled dataset generated by the EURO-CORDEX project [43]. In the following study, we used EURO-CORDEX simulations for Europe with horizontal resolutions of 0.11° (EUR-11, ~12.5 km) [44]. The original gridded data were bilinearly interpolated to a regular grid by using CDO (climate data operators; v. 1.7.0) software [45]. In the study, a six-member ensemble of GCM-RCM chain simulations (

Table 1. GCM-RCM chain simulations used in the study.

N	GCMs	RCMs	Institution
1	CNRM-CERFACS-CNRM-CM5	SMHI-RCA4	SMHI
2	ICHEC-EC-EARTH	SMHI-RCA4	SMHI
3	ICHEC-EC-EARTH	KNMI-RACMO22E	KNMI
4	ICHEC-EC-EARTH	DMI-HIRHAM5	DMI
5	IPSL-IPSL-CM5A-MR	SMHI-RCA4	SMHI
6	MPI-M-MPI-ESM-LR	SMHI-RCA4	SMHI

) were used, and the ensemble mean was used as final indicator value. The bias adjustment model output performed under the CORDEX-Adjust project (SMHI-DBS45-MESAN, 1989–2010) was used [46,47]. The EURO-CORDEX simulations uses the RCP scenarios (Representative Concentration Pathways) used in the IPCC fifth assessment report [48]. For the present study, we applied two anthropogenic forcing scenarios, namely with low (LE, RCP4.5) and high emissions (HE, RCP8.5). The RCP4.5 is an intermediate stabilisation pathway that assumes radiative forcing stabilisation at 4.5 Wm⁻² after year 2100 while RCP8.5 is a high pathway for which radiative forcing after year 2100 still continues to rise above 8.5 Wm⁻² [49]. The analyses were performed for the three time periods: (1) baseline period (BL) 1981–2010, and two future periods, (2) 2050s (2041–2070) and (3) 2080s (2071–2100).

While using climate change scenarios, it should be mentioned that the results of numerical simulations differ from observations, therefore it is necessary to compare them [50]. Ensemble mean of six bias-corrected simulations were evaluated for annual mean temperature and annual precipitation sum. Following other studies [51], the averaged error values and RMSEs were calculated for 42 Institute of Meteorology and Water Management (IMGW-PIB) meteorological stations [52] for 1981–2010 time frame (Figure 1). The RMSE in annual means was 0.3 °C for temperature and 36 mm/year (6%) for precipitation. For each meteorological station, the error for precipitation was less than 15% that is an acceptable model bias of up to 20% of true precipitation due to a systematic rain gauge undercatch [50,53].

2.3. Water Condition Evaluation for Maize Cultivation

Climatic water balance (CWB) is one of the indicators used in Poland for meteorological drought monitoring as well as for assessment of its intensity [24]. CWB is the difference between precipitation (P) and potential evapotranspiration (PET) [54]. Since there is a tendency for increasing the frequency of extreme weather phenomena such as drought [55], water deficiency indicated by climatic water balance sum for the period from April to September was indicated as suitability criteria for maize cultivation. In the literature, there are many methods for PET calculation, however, some of them require numerous input data. Depending on the application, a number of simplifications is possible to implement in PET formula [56]. Doroszewski and Górski [56] elaborated an index of PET as a function of sunshine duration, daylength and temperature. In the following study, the calculation of an index of potential evapotranspiration was carried out on the basis of the simplified Penman formula presented by Doroszewski and Górski [56] and Żyłowska et al. [57].We calculated CWB based on an indicator of moisture conditions as a function of mean monthly air temperature and monthly precipitation sum [57].

2.4. Suitability Assessment

AHP models were used in several studies for weighting factors in land suitability assessment [13,27,29,31]. Thus, the selected criteria were rated using the AHP methodology introduced by Saaty [58]. Steps of the AHP are given as follows:

Step 1: Construction of pairwise comparison matrix. In this step, two criteria are evaluated at the time in terms of their relative importance regarding the suitability of the land for maize cultivation, considering water deficiency occurrence [59]. In AHP 9 points importance scale is used to express individual preferences or judgments (

Table 2. Scale of relative importance of selected criteria based on Saaty [58].

Intensity of Importance	Definition
1	Equal importance
3	Moderate importance
5	Strong importance
7	Very strong importance
9	Extreme importance
2, 4, 6, 8	Compromise between above values
Reciprocals of above	Values for inverse comparison

) [31,58]. The pairwise comparison matrixes were created by consulting the opinions of the team experts from the Institute of Soil Science and Plant Cultivation (IUNG) and workers of Agricultural Experimental Stations (RZD IUNG) leading long-term experiments.

Step 2: Calculation of the eigenvectors and eigenvalue of the judgment matrix. After construction of pairwise comparison matrix, the eigenvectors ${\mathrm{w}}_{\mathrm{i}}$ (weights of criteria) and eigenvalue ${\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}$ were calculated as follows [29]:

$${\mathrm{w}}_{\mathrm{i}}=\frac{\sqrt[\mathrm{n}]{{\prod }_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{a}}_{\mathrm{i}\mathrm{j}}}}{{\sum }_{\mathrm{i}=1}^{\mathrm{n}}\sqrt[\mathrm{n}]{{\prod }_{\mathrm{j}=1}^{\mathrm{n}}{\mathrm{a}}_{\mathrm{i}\mathrm{j}}}},$$

(1)

$${\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}=\frac{1}{\mathrm{n}}\cdot {\sum }_{\mathrm{i}=1}^{\mathrm{n}}\frac{{\left(\mathrm{A}\mathrm{W}\right)}_{\mathrm{i}}}{{\mathrm{w}}_{\mathrm{i}}},$$

(2)

where $\mathrm{i}$ and $j$ is number of rank and column of $\mathrm{A}$.

Step 3: Consistency test. By using AHP methodology, the inconsistency in pairwise comparison matrix is identified and taken into account [60,61]. The consistency ratio ($\mathrm{C}\mathrm{R}$) proposed by Saaty [58] is used to test the consistency of pairwise comparisons. If the $\mathrm{C}\mathrm{R}$ is below 0.10, it is considered that the inconsistency of the judgments is acceptable [31,58,60]. The consistency index ($\mathrm{C}\mathrm{I}$) and the consistency ratio ($\mathrm{C}\mathrm{R}$) is given by [62]:

$$\mathrm{C}\mathrm{I}=\frac{{\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}-\mathrm{n}}{\mathrm{n}-1},$$

(3)

$$\mathrm{C}\mathrm{R}=\frac{\mathrm{C}\mathrm{I}}{\mathrm{R}\mathrm{I}},$$

(4)

where $\mathrm{R}\mathrm{I}$ is Random Consistency Index which depends on $\mathrm{n}$ (number of elements), listed in

Table 3. Random Consistency Index ($\mathrm{R}\mathrm{I}$) values [58].

$$\mathbf{n}$$	1	2	3	4	5	6	7	8	9	10
$$\mathrm{R}\mathrm{I}$$	0.00	0.00	0.52	0.89	1.11	1.25	1.35	1.40	1.45	1.49

.

Step 4: Group decision making. Since decision making frequently involves contribution of many respondents with different individual opinions, group decision making process has been applied to aggregate them [63,64]. There are alternative methods suggested by the literature to aggregate individual opinions applied in AHP methodology, depending on whether the decision group is assumed to be a synergistic unit or simply a collection of individuals. When the individuals act in concert, judgments should be pooled in such a way that the group ‘merges’ the different individuals into a new ‘individual’ representing it. Aggregation of Individual Judgments (AIJ) is suggested for this purpose, that requires satisfaction of reciprocity condition for the synthesis alike the individual judgments. This is provided by the geometric mean, thus in AIJ, the elements of aggregated comparison matrix ${\left({\mathrm{a}}_{\mathrm{i}\mathrm{j}}\right)}_{\mathrm{g}}$ would be calculated as follows [63,64,65]:

$${\left({\mathrm{a}}_{\mathrm{i}\mathrm{j}}\right)}_{\mathrm{g}}=\sqrt[\mathrm{N}]{{\prod }_{\mathrm{n}=1}^{\mathrm{N}}{\left({\mathrm{a}}_{\mathrm{i}\mathrm{j}}\right)}_{\mathrm{I}\mathrm{n}\mathrm{d}\_\mathrm{n}}},$$

(5)

where ${\left({\mathrm{a}}_{\mathrm{i}\mathrm{j}}\right)}_{\mathrm{I}\mathrm{n}\mathrm{d}\_\mathrm{n}}$ is the judgment of individual $\mathrm{n}$, $\mathrm{N}$ is number of individuals that belong to the group.

The selected suitability criteria (CWB values for four soil types) were reclassified into four suitability classes (S1, S2, S3, N) (

Table 4. Suitability level per selected criteria [68].

Suitability Class	Description
S1—highly suitable	No significant limitations or only minor limitations
S2—moderately suitable	Moderately severe limitations that reduce productivity or benefits and rise inputs
S3—marginally suitable	Severe limitations that reduce productivity or benefits and rise inputs to the extent of being marginally justifiable
N—not suitable	Severe limitations that may preclude sustained use of the land in a specific manner

and

Table 5. Factor ratings of climate suitability for maize regarding water deficiency.

Characteristic		Suitability Class 1				References
Criteria		S1	S2	S3	N
Rating scale		85–100	60–85	40–60	0–40	[29,68]
Climatic water balance	on very light soils [mm]	>−150	−200–−150	−250–−200	<−250	[69]
	on light soils [mm]	>−210	−260–−210	−310–−260	<−310	[69]
	on medium soils [mm]	>−250	−300–−250	−350–−300	<−350	[69]
	on heavy soils [mm]	>−290	−340–−290	−390–−340	<−390	[69]

¹ S1—highly suitable, S2—moderately suitable, S3—marginally suitable, N—not suitable.

). The soil types reflect their agronomic category according to their granulometric composition [66,67]. The soils are divided into four categories reflecting their different susceptibility to drought in relation to the amount of available water (AW): very light soils—the most susceptible (AW less than 127.5 mm), light soils—susceptible soils (AW between 127.5 and 169.9 mm), medium soils—moderately susceptible soils (AW between 170 and 202.5 mm) and heavy soils—soils with low susceptibility to drought (AW more than 202.5 mm) [66]. The range of field water capacity in the soil profile from 0 to 100 cm of these categories of soils is as follows: 110–145 mm, 146–210 mm, 211–270 mm and 271–460 mm [67]. The final suitability map was generated by using weighted overlay method, by combining the standardized criterion maps with the criterion weights [59,61] using QGIS 3.28.4.

$$\mathrm{Weighted} \mathrm{Overlay}={\sum }_{\mathrm{i}=1}^{\mathrm{n}}{\mathrm{C}}_{\mathrm{i}}\cdot {\mathrm{W}}_{\mathrm{i}},$$

(6)

where ${\mathrm{C}}_{\mathrm{i}}$ is each reclassified criterion ($i$), ${\mathrm{W}}_{\mathrm{i}}$ is the weight of the selected criteria, $\mathrm{n}$ is the total number of criteria.

The final suitability map was divided into four suitability classes according to FAO [29,68]. The flowchart of the methodology is presented in Figure 2.

2.5. Irrigation as Potential Adaptation Strategy

Nowadays irrigation in Poland is treated as a supplemental practice, it is usually used in short time spans during the vegetation period. However, it can mitigate the effects of droughts on crop production locally. During the period from April to September, potential evapotranspiration usually exceeds precipitation causing a water deficit. It is seen especially on light soils with low water-holding capacity. The net irrigation water requirement in the growing period in Central Poland for maize is 50–70 mm [70]. Łabędzki [71] identified the need to assess the impact of climate change on irrigation requirements in Poland. Due to temperature increases with lower precipitation during summer, it is expected that irrigation needs will increase [71]. Therefore, in the current study, we tested irrigation by 30, 80 and 140 mm by adding three different amounts of irrigation water that impacted CWB for the growing season from April to September. The chosen amount of water supply (30, 80 and 140 mm) during irrigation is an example to test three types of irrigation requirements: (1) small (less than 50 mm—it is about 3 irrigation treatments per season), (2) medium (between 50 and 100 mm) and (3) high (more than 100 mm) [72].

3. Results

3.1. Mean Annual Temperature and Mean Annual Precipitation Sum for Baseline Period and by 2050s and 2080s

According to the ensemble mean of six analysed RCMs, the mean annual air temperature in Poland for 1981–2010 (BL) was 8.2 °C, while the mean annual precipitation sum was about 622 mm (

Table 6. Mean values of annual temperature and annual precipitation sum and their changes (in brackets) in Poland.

Period 1	BL	2050s		2080s
Emission Scenario 2		LE	HE	LE	HE
Mean annual temperature [°C]	8.2	9.7 (+1.5)	10.3 (+2.1)	10.3 (+2.1)	12.0 (+3.8)
Mean annual precipitation [mm]	622	676 (+9%)	710 (+14%)	692 (+11%)	741 (+19%)

¹ BL—baseline; ² LE—low emissions scenario; HE—high emissions scenario.

). The spatial distribution of mean annual air temperature (Figure 3) shows its decrease from the southwest to the northeast of Poland under both current and future climate conditions. The projected increase in the annual temperature from the BL to the 2050s according to the LE scenario is about 1.5 °C and 2.1 °C to the 2080s (Table 6). For the HE scenario, projected mean annual temperature changes can reach about 2.1 °C and 3.8 °C by the 2050s and 2080s, respectively. The spatial distribution of mean annual precipitation is presented in Figure 4. Both in the baseline period and under future climate conditions, the highest annual precipitation occurs in mountainous areas (southern Poland) and the lowest in central Poland. According to the LE scenario, the annual precipitation will not change significantly in the 2050 perspective, whereas the increase in precipitation will be greater in the 2080 perspective. The HE scenario predicts a higher increase in precipitation than the LE scenario. Projected changes in mean annual precipitation sum according to the analysed ensembles of models are about +9% and +14% in the 2050s and +11% and +19% in the 2080s for the LE and HE scenarios, respectively.

3.2. Mean Climatic Water Balance in the Period from April to September for Baseline Period and by 2050s and 2080s

The growing season sum of CWB (from April to September), averaged over Poland for the BL, was −213 mm (

Table 7. Mean values of the climatic water balance sum and its changes (in brackets) from April to September in Poland.

Period 1	BL	2050s		2080s
Emission Scenario 2		LE	HE	LE	HE
CWB sum (Apr–Sept) [mm]	−213	−237 (−11%)	−214 (0%)	−240 (−13%)	−230 (−8%)

¹ BL—baseline; ² LE—low emissions scenario; HE—high emissions scenario.

). According to the analysed LE scenario, the CWB will decrease by about 11% in the 2050s and will not change according to the HE scenario. In the 2080s, the seasonal CWB change will decrease by 13% according to the LE scenario and by 8% according to HE. The spatial distribution of the CWB for the growing season (from April to September) is shown in Figure 5. For most of the area of Poland, the CWB was negative as the PE exceeded the sum of precipitation. The highest positive values of CWB (more than 100 mm) were found in the south of Poland, where the sum of precipitation exceeded the PE. In north-western part of Poland (near the Baltic Sea) and in the southern part of Poland the values of CWB were more than −150 mm. The lowest values of the climatic water balance (less than −250 mm) were observed in the central Poland. The area with CWB less than −300 mm is projected to increase in the future, and according to the RCP8.5 scenario this area will dominate in 2071–2100.

3.3. Suitability Assessment

While assessing drought risk, it is important to distinguish water deficiency between the different soil categories. In the following study, we reclassified CWB values into four suitability classes varying according to soil types (Table 5, Figure 6). According to BL, the not suitable areas was observed for very light and light soils in central Poland. On very light soils, not suitable areas are going to increase in the future from 43% for BL to 57% for LE and 72% of Poland for HE, while for light soil the area is projected to increase from 10% for BL to 26% for LE and 48% of Poland for the HE scenario. On heavy soils, there is no extreme deficiency for CWB observed in neither baseline nor futures scenarios apart from 2080s for the HE scenario. The highly suitable areas are observed on all kinds of soil in north-west and south of Poland, however the largest area is observed for medium and heavy soils.

Four soil categories (very light, light, medium and heavy) were weighted using farmers’ (5) and agricultural advisors/experts’ (5) opinion using AHP pairwise comparison matrix (in total, 10 individuals). The purpose of the weighting is to determine the importance of each soil categories relative to others that influence land suitability for maize production, considering the occurrence of water deficiencys on these soil categories. The respondents indicated what type of soil they would choose for growing corn in the context of water stress (drought). The mean judgement of the individuals derived through pairwise comparisons is presented in

Table 8. AHP pairwise comparison matrix for the main criteria and resulted weight vector.

Criteria	Very Light	Light	Medium	Heavy	Weights
Very light	1.0000	0.2194	0.1305	0.1386	0.0433
Light	4.5581	1.0000	0.1909	0.2437	0.1169
Medium	7.6617	5.2382	1.0000	1.3286	0.4655
Heavy	7.2157	4.1038	0.7526	1.0000	0.3743

. The weights of each soil categories calculated by the AHP method are listed in the last column of Table 8, with the higher value indicating that the criterion is more important. Based on the calculated values of the criteria weights (Table 8), medium soil categories are the most important factor, followed by heavy soils, while light soils are significantly less important. The least important factor in considering land suitability for maize cultivation due to water deficiency occurrence are very light soils. In this suitability analysis, the consistency ratio (CR) value is 0.0535, which is below the threshold value of 0.1 indicating a reasonable level of consistency of the experts’ judgment.

The final suitability map (Figure 7) was generated by using the weighted overlay method, by combining the reclassified criterion maps (Figure 6) with the criterion weights (Table 8). The suitability maps of the water deficiency indicators for maize cultivation were categorised into four suitability classes, namely: highly suitable (S1), moderately suitable (S2), marginally suitable (S3) and not suitable (N). The percentage of the suitability classes in Poland is presented in

Table 9. Percentages of different water deficiency suitability levels for maize in Poland for the baseline (1981–2010) and projected according to low emissions (LE, RCP4.5) and high emissions (HE, RCP8.5) scenarios for the 2050s (2041–2070) and 2080s (2071–2100).

RCP					RCP4.5								RCP8.5
Period	BL				2050s				2080s				2050s				2080s
Rating scale 1	S1	S2	S3	N	S1	S2	S3	N	S1	S2	S3	N	S1	S2	S3	N	S1	S2	S3	N
Percentage	81	18	1	0	67	26	7	0	64	24	12	0	69	22	8	0	44	24	21	11

¹ S1—highly suitable, S2—moderately suitable, S3—marginally suitable, N—not suitable.

. Water deficiency is one of the key climate variables for maize, in consequence, climate change could significantly impact its production. Central Poland is the only area of Poland that has moderate suitability (18%) for maize cultivation due to the occurrence of water deficiency in the current climate. Subsequently, these conditions will extend to 22–26% of Poland depending on the period and scenario analysed. By the end of the century, water deficiency will increase and, according to HE, 11% of the area is projected to be not suitable while 21% will be only marginally suitable for rainfed maize cultivation (0 and 1% respectively in current conditions).

3.4. Adaptation Strategies Increasing Land Suitability for Maize—Irrigation

Considering water deficiency, climate suitability for maize will decrease over time. The distribution with water limitations is shown in Figure 8. According to water stress, 19% of Poland, mainly in central Poland, was moderately or marginally suitable for maize cultivation under current climate conditions (

Table 10. Percentages of different water deficiency suitability levels for different amounts of water supply through irrigation (30, 80 and 140 mm per season) in maize cultivation in Poland for the baseline (1981–2010) and projected according to low emissions (LE, RCP4.5) and high emissions (HE, RCP8.5) scenarios for the 2050s (2041–2070) and 2080s (2071–2100).

RCP					RCP4.5								RCP8.5
Period	BL				2050s				2080s				2050s				2080s
Rating scale 1	S1	S2	S3	N	S1	S2	S3	N	S1	S2	S3	N	S1	S2	S3	N	S1	S2	S3	N
0 mm	81	18	1	0	67	26	7	0	64	24	12	0	69	22	8	0	44	24	21	11
30 mm	94	6	0	0	84	16	0	0	79	20	2	0	82	16	1	0	57	23	18	1
80 mm	100	0	0	0	100	0	0	0	98	2	0	0	99	1	0	0	80	18	1	0
140 mm	100	0	0	0	100	0	0	0	100	0	0	0	100	0	0	0	100	0	0	0

¹ S1—highly suitable, S2—moderately suitable, S3—marginally suitable, N—not suitable.

). Furthermore, under future climate conditions, this area will increase to 33 and 30% in the 2050s and 36 and 45% in the 2080s for the LE and HE scenarios, respectively. However, in the 2080s for the RCP8.5 scenario according to water deficiency, 11% of the analysed area will not be suitable for maize production. Under current climate conditions, according to the results of reclassification of water deficiency, the highly suitable area would increase by 13%, 19% and 19% if 30, 80 and 140 mm of water were applied in irrigation, respectively. In the 2050s, the application of 30 mm of water will help to avoid marginally suitable areas for maize cultivation, while irrigation with 140 mm of water will create highly suitable conditions for the whole of Poland regarding water deficiency. In the 2080s, according to the HE scenario, the application of 30 mm of water will help to eliminate not suitable areas for maize cultivation, but to avoid marginally suitable areas, irrigation with 80 mm of water will be necessary. If 80 mm of water is applied by irrigation, about 18% of the analysed area will remain moderately suitable for maize production.

4. Discussion

4.1. Climate Change Impact on Maize

The growth and yield of maize is significantly affected by climate change. There has been a tendency for grain maize area and yield to increase in Poland over the last 30 years [1]. This study presents the spatial distribution of selected climatic variables and how they change over time related to climate change. According to the climate change scenarios analysed, the increase in both annual temperature and precipitation sum is projected for Poland. The highest increase (by almost 4 °C) in annual mean temperature is projected in HE scenario for 2080s. Similarly, warmer and wetter conditions are projected for Poland on an annual scale by Mezghani et al. [51] and Strużewska et al. [73] together with an intensification of the changes expected by the end of the 21st century.

The obtained results show that despite the fact that annual precipitation sum is projected to increase in both LE and HE scenarios, the growing season sum of CWB is projected to decrease. The potential evapotranspiration exceeds precipitation over most of Poland during the period of April to September, resulting in a water deficit. The highest water deficit was observed in the Central Poland and this area is projected to increase under future climate conditions. Similar spatial distribution of climatic water balance is presented in [36]. Other studies highlight the uncertainty of summer precipitation projections depending on the models analysed, which leads to considerable uncertainty in the assessment of the impacts of climate change on the water balance [74].

Ziernicka-Wojtaszek [75] evaluated changes in thermal resources and CWB in Poland for two time periods: 1971–2000 and 1981–2010. It was shown that the very warm and dry areas increased from 10% to 48% and from 34 to 48%, respectively. The novelty of this paper is that we have also considered the expert opinion on soil categories importance in water deficiency evaluation, influencing climate suitability for maize production, that was not concerned in previous research on adaptation of maize production in Poland. Our study focused on water deficiency determined based on the climatic water balance and tested an adaptive irrigation practice that can mitigate the effects of increasingly frequent droughts.

In our study, soil categories were weighted using AHP methodology based on farmers’ and experts’ opinions. The aim of weighting the soil categories was to determine their relative importance in influencing land suitability for maize production due to the occurrence of water deficiency. According to the mean opinion of farmers and experts, medium and heavy soils were defined as the most important soil category in assessing land suitability evaluation for maize production, while light and very light soils were much less important. However, in Poland, the share of light and very light soils with low water capacity is predominant (60%) [76] and water deficit is most frequently observed on these soils [35].

Observed changes in climatic conditions indicate that increasingly frequent water shortages will be the main constraint for maize production in Poland, while thermal conditions will no longer limit production [77]. Therefore, we decided to evaluate the climate suitability for maize production in relation to occurrence of water stress under current and future climate conditions. Under current climate conditions, Poland has highly and moderately suitable water conditions for maize cultivation. In the future, water conditions for maize cultivation will worsen and about 10% of Poland will become marginally suitable. However, the biggest changes are observed in 2080s according to HE scenario, where 21% of Poland will have marginally suitable and 11% not suitable water conditions for maize production. The worsening of water conditions for maize production will be in central Poland, where the climatic water balance is the lowest. The previous studies [77,78] show an improvement of climatic conditions for maize cultivation in Poland. With climate warming the areas suitable for maize cultivation have increased significantly [77]. Ziernicka-Wojtaszek [79] introduced new agroclimatic regionalisation of the Polish area. It was simulated that increase in temperature by 1 °C and 2 °C with unchanged precipitation will result in a decrease in the area of optimal humidity from 70 to 48% and 26% respectively [79]. Ramirez-Cabral et al. [80] modelled the potential climate distribution of maize at global level in current and future climate conditions using the CLIMEX model taking into account stress factors that reduce climate suitability. The medium suitability for maize cultivation in Poland was mostly projected under current climate conditions, except in central Poland where marginal climate suitability occurs. For further climate conditions, different results were obtained for different general circulation models (GCM) CSIRO and MIROC. CSIRO projected a slight increase in the marginal suitability category, especially for the year 2100. However, for 2050 MIROC projected a change from suitable to optimal suitability in the northern and southern part of Poland while in 2100 almost the whole of Poland is projected to be optimal for maize cultivation except for the central part where suitable areas occur [80].

It should be noted as limitation of this research that the analyses carried out mainly focused on water deficiency factors and only include climatic factors and their changes in the future climate conditions, while the potential genetic progress of the species was not considered. Other research [80] suggests that it would be desirable to include non-climatic factors such as topography, land use, soil taxonomy and its physico-chemical properties in further studies. While including expert opinion in the identification of criteria relevance, especially when many criteria are evaluated, it is important to provide simple and intuitive techniques to obtain it, due to the possibility of errors occurring in the evaluation of criteria importance. There is interest in the literature for methods that better handle expert hesitation that can arise in subjective weighting methods. Więckowski et al. [81] proposed the RANCOM method, which is based on the ranking of criteria to obtain criteria weights on the basis of expert knowledge. RANCOM could be a valuable method for evaluating criteria weights or its comparison with the AHP methodology used in our study for future research since RANCOM is a simple method which can be adapted for both less and more experienced users. The RANCOM method can provide more repeatable results than the AHP method. This is especially true for a larger number of criteria [81].

4.2. Adaptation to Climate Change

Irrigation is a treatment that can effectively prevent the impact of drought on crops [82]. The use of irrigation in the cultivation of grain maize results in high yield increases that are difficult to achieve with other farming technique treatments [82]. The effect of agricultural drought on crops mainly depends on its size, duration and intensity [24]. Irrigation may help stabilize crop yields [82]. In assessing the impact of climate change on agriculture due to the lack of a clear picture regarding the amount of precipitation and the increase in evapotranspiration resulting in a deterioration of climatic water balances, the need for irrigation is increasingly being indicated [83]. Hence, we decided to test irrigation with 30, 80 and 140 mm of water and its impact on land suitability for maize production in Poland. According to water deficiency, Poland is highly or moderately suitable for maize production under current climatic conditions, but suitability is projected to decrease over time. In the future, marginally suitable water conditions and even not suitable water conditions are going to appear in central Poland. Under current climatic conditions, 80 mm of water should be applied to achieve highly suitable water conditions for maize production in almost all of Poland, while under future climatic conditions, 140 mm should be applied. Thanks to the application of 30 mm of water, the water conditions for maize production can be improved in future climate conditions, so that no marginally suitable areas will appear, except in the 2080s, according to the HE scenario where 80 mm of water should be applied.

Despite a considerable number of studies, it is very difficult to determine the amount of irrigation required for a given crop due to the variation in water demand caused by soil and meteorological factors [25]. The use of irrigation in maize production is essential in an area of distinct water deficits with low precipitation and soils with very low retention capacity [84]. Water requirements also vary between crops and their varieties, depending on the length of the growing season and the sensitivity to water shortages at different stages of development [25]. In Poland, agricultural drought is determined by the Agricultural Drought Monitoring System (ADMS) based on the Climatic Water Balance (CWB). In the determination of agricultural drought, the characteristics of the soil retention related to its granulometric composition constitute an important factor that is considered in ADMS. In the system, critical values of climatic water balance are differentiated for particular groups and species of crop plants [54].

The literature suggests that some adaptive agricultural practices should be introduced to counteract the negative effects of climate change, such as variety renovation, adjusting sowing dates, improving irrigation and fertiliser efficiency [85]. According to some research, irrigation in maize production can increase maize yield as well as stabilise yield under Polish conditions [82,86,87,88,89]. However, the use of irrigation in maize production has not always been cost effective [86]. The production effect is the main indicator of the desirability of irrigation. The production results are strongly influenced by weather conditions, in particular the amount and distribution of rainfall [82]. The effect of irrigation depends on precipitation sum in July and August, the lower precipitation, the higher the yield increase [82,87]. In the study developed by Żarski et al. [89] during the wet season, the maize grain yield increase was non-significant while in dry seasons the yield increased by more than half. Due to increased drought under climate change, soil water conservation practices along with more drought-resistant varieties are also proposed to adapt to these changes in maize production [90]. Furthermore, Zhao et al. [90] identified the use of cultivars that are more drought-resistant and cultivars with higher temperature requirements as effective adaptations to climate change in maize production in Poland (Central Europe).

5. Conclusions

This study assessed the potential effects of climate change on climate suitability related to water deficiency for maize cultivation in Poland using multi-criteria evaluation through the Analytical Hierarchy Process (AHP). The weights of soil category were calculated by AHP to include the expert opinion in the evaluation.

The obtained results suggest that the water stress limitations depend on the soil category and under the current climate conditions the extreme deficiency of the climatic water balance was observed on very light and light soils in central Poland, but not on heavy soils. Considering all soil categories with different relative importance according to experts’ opinion, central Poland was moderately suitable under current climate conditions, while the rest of Poland was highly suitable. However, in the future, the suitability according to water deficiency is predicted to decrease, and some areas in central Poland will become marginally suitable for maize cultivation or even not suitable in the 2080s according to the RCP8.5 scenario.

This study focused on the impact of climate change on maize suitability in Poland, therefore water deficiency factors that may harm maize production in the future were considered. Studying the climate suitability of a particular crop can help to manage the impact of climate change on crop production. However, in future studies, the combined consideration of topographic, climatic and soil factors would allow a more complex assessment of maize land suitability.

The obtained results show that due to climate change and higher temperatures, there will be problems related to water deficiency. This study indicated that water deficiency may limit climate suitability for maize cultivation in far future, but irrigation could mitigate this effect. The results of the study could be used to develop adaptation strategies to enhance maize productivity under future climatic conditions. Identification of climate limitations, their spatial variability and changes in the future may play an important role in keeping maize production risk at a minimum level. The methods applied in the study can be used as a tool for finding pathways to enhance the resilience of farming systems dealing with challenges.