Evaluation of Climate Suitability for Maize Production in Poland under Climate Change

Abstract

Climatic conditions are the main factor influencing the suitability of agricultural land for crop production. Therefore, the evaluation of climate change impact on crop suitability using the best possible methods and data is needed for successful agricultural climate change adaptation. This study presents the application of a multi-criteria evaluation approach to assess climate suitability for maize production in Poland, for a baseline period (BL, 1981–2010) and two future periods 2041–2070 (2050s) and 2071–2100 (2080s) under two RCP (Representative Concentration Pathways) scenarios: RCP4.5 and RCP8.5. The analyses incorporated expert knowledge using the Analytical Hierarchy Process (AHP) into the evaluation of criteria weights. The results showed that maturity and frost stress were the most limiting factors in assessing the climatic suitability of maize cultivation in Poland, with 30% and 11% of Poland classified as marginally suitable or not suitable for maize cultivation, respectively. In the future climate, the area limited by maturity and frost stress factors is projected to decrease, while the area of water stress and heat stress is projected to increase. For 2050 climate projections, water stress limitation areas occupy 7% and 8% of Poland for RCP4.5 and RCP8.5, respectively, while for 2080 projections, the same areas occupy 12% and 32% of the country, respectively. By 2080, heat stress will become a limiting factor for maize cultivation; according to our analysis, 3% of the Polish area under RCP8.5 will be marginally suitable for maize cultivation because of heat stress. The overall analyses showed that most of Poland in the BL climate is in the high suitability class (62%) and 38% is moderately suitable for maize cultivation. This situation will improves until 2050, but will worsen in the 2080s under the RCP8.5 scenario. Under RCP8.5, by the end of the century (2080s), the highly suitable area will decrease to 47% and the moderately suitable area will increase to 53%.

1. Introduction

Observed climate change is seen as one of the major concerns for agriculture, which poses a challenge for the adaptation of agriculture to ongoing climate change [1]. At the same time, scientific progress has achieved considerable control over various parameters affecting agricultural production, such as biophysical soil characteristics. But when climate is added to the equation of shaping pedoclimatic conditions, things become more complicated, especially with current frequent weather perturbations that pose great challenges to agricultural science. Various stresses, such as water shortage, heat, frost, pests and diseases determine both the quantity and quality of crop yields [2]. To minimise risks to crop production in the context of climate change [3], it is crucial to evaluate changes in land suitability with a multivariable approach [4], and thus more and more detailed analyses can be carried out to support the preparation of improved adaptation strategies [3,5]. Refining land suitability assessments may help decision-making on agricultural land use, which can then be used by governments and institutions to improve the resilience of farming systems to climate change [6].

Maize, as a thermophilic plant, is used in many studies as a model crop to show the effects of climate change on land suitability [4,6,7,8]. In previous studies it has been shown that the expected temperature increase would improve conditions for maize production in Poland [9], which seems to be already being realised and is shown by statistics. In Poland, the area under grain maize cultivation has increased significantly, from 152,000 ha in 2000 to 1,196,000 ha in 2022 [10]. The yield over the last 24 years (2000–2023) shows an increasing trend from 5.6 to 7.0 t/ha. However, this has been accompanied by relatively high year–to–year yield variability, ranging from 4.2 t/ha in 2006 to 7.5 t/ha in 2021 [11]. It is mainly due to extreme weather events, with droughts being the most important factor [12,13]. These conditions require not only a review of the suitability of land for maize production, but also a different approach to analysis, not just one that is variable-based, but which is also based on multivariables. An additional condition is the need to rebalance production systems for proper adaptation actions for a more sustainable model, so-called short-term and long-term strategies, which should also be taken into account in land suitability analyses [14] and farmers’ perception [15]. In addition, agriculture should support climate change mitigation, with a particular focus on reducing emissions from agriculture. This can be achieved through fertilisation adapted to crop needs and soil conditions, care of the soil environment and rational management of resources [14].

Through a multicriteria approach in suitability areas assessment for agricultural production, many factors can be assessed [16,17], involving expert opinions [18]. In the view of the fact that these criteria have different levels of importance for land suitability, there are many methods used for criteria weight evaluation [19]. The AHP has been widely used to determine the weights of the parameters in the land suitability assessment for maize [20,21,22,23] and other crops such as tea [24,25], rice and soybean [23]. The AHP has the advantage of incorporating experts’ opinion in order to prioritise the criteria [21]. Another advantage over other methods is the measurement and control of judgment inconsistency. One drawback 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 different criteria [20].

In spite of many studies on the impact of climate change on the conditions for growing maize in Poland, it is difficult to obtain a clear answer as to whether these changes are beneficial or not [1]. Therefore, the overall objective of this study is to assess the climatic conditions for maize cultivation, considering both the positive effects of climate change, such as the increase in accumulated temperature leading to crop maturity, and the potential risks, such as those of water deficiency, frost and heat stress. These factors were integrated using a multicriteria assessment approach. Additionally, the experts’ opinion were also integrated during the evaluation of selected criteria importance. This study expands on earlier work [26] in which the climate suitability of areas for maize cultivation was assessed, but only in relation to water deficit stress. Using the same methodology, and the developed water stress, and adding further factors: heat stress, frost stress and reaching maturity, we re-evaluated climate suitability for maize cultivation in Poland. Furthermore, in order to adapt the crop to the climatic conditions and provide more realistic results, the climate risk indicators were assessed using possible sowing and harvesting dates based on air temperature.

2. Materials and Methods

2.1. Study Area

Poland is located in Central Europe, from 14°07′ E to 24°09′ E and from 49° N to 54°50′ N, and covers an area of 311,888 km². It is mainly a lowland area with an average altitude of 173 m above sea level, but it has mountains in the south (the Carpathians and the Sudetes). Poland’s climate is influenced by various air masses: maritime from Western Europe, continental from the East, polar from the Arctic and subtropical from the Atlantic, Mediterranean and Black Seas [27]. There are six distinct seasons in Poland: early spring, spring, summer, autumn, early winter and winter. In Poland the seasons vary from year to year due to significant natural variations in the climate [27]. In 2022, the average annual temperature in Poland was 9.5 °C, and it varied between 8.9 and 9.9 °C in different areas. In 2022, the average annual precipitation sum in Poland was 534.4 mm and it was characterised by spatial variability, ranging from over 350 mm to almost 950 mm. The 2022 year should be classified as a very warm and dry year [28]. Based on climatological norms for the period 1991–2020, the average annual temperature in Poland is 8.4 °C, while area-averaged precipitation total in Poland is 678.6 mm [29]. The growing season also varies between areas; in the northeast of Poland it is about 40 days shorter than in the southwest [27]. The average duration of the growing season in Poland for the period 1991–2020 was 224 days [29].

Natural conditions in Poland are fairly favourable for agriculture, but climate change is making droughts and floods more frequent and intense [30,31]. Decreases in yields of traditional crops are also predicted for most of the country, with their inter-annual variability. However, despite the negative impacts, there are also positive changes observed in agriculture: the length of the vegetative period will increase, and more favourable conditions for growing crops typical of warmer areas (maize, soya bean, etc.) are predicted [27].

2.2. Climate Data Dataset

Regional climate model (RCM) projections are used for regional climate change studies and impact assessments. Compared to global climate models (GCMs), they have the advantage of superior resolution and better representation of the local climate. However, relying solely on the output of RCMs can be subject to uncertainties and biases that may affect the accuracy of the results. As part of the Coordinated Regional Downscaling Experiment Project (CORDEX) project, the EURO-CORDEX initiative is providing high resolution RCM projections for the European domain that are openly available through the Earth System Grid Federation (ESGF) [32,33]. The climate projections data used in this work were obtained from the ESGF server (https://esgf-data.dkrz.de/search/cordex-dkrz, accessed on 6 July 2021). The bias-adjusted daily mean/max/min temperature and precipitation data are available from the CORDEX-Adjust project. These data are bias-corrected using a number of different methods [33]. For analysis, the daily precipitation, minimum, maximum and mean temperature bias-adjusted outputs of RCMs were obtained from the CORDEX Adjust database. These outputs were generated by the Swedish Meteorological and Hydrological Institute (SMHI) with Distribution-Based Scaling (DBS45) method [34] and were calibrated by SMHI according to the MESoscale ANalysis system (MESAN) reanalysis data [35]. These datasets had an original spatial resolution of 0.11°. CDO (Climate Data Operators; v. 1.7.0) software [36] was used to bilinearly interpolate the original gridded data to a regular grid and for further calculations. Due to the fact that there are considerable uncertainties in the projections of future climate by climate models, previous studies suggest that an averaged ensemble of several climate models, rather than a single one whose output is unreliable, should be used to assess future climate [30,32]. Therefore, in this and our previous study [26], we decided to use a mean ensemble of six GCM-RCM chain simulations (combination of 4 GCMs and 3 RCMs) to obtain a reliable projection of future climate variables (

Table 1. Combinations of GCMs and RCMs that were used in this study.

Global Climate Models	Regional Climate Models
	SMHI-RCA4	KNMI-RACMO22E	DMI-HIRHAM5
CNRM-CERFACS-CNRM-CM5	X
ICHEC-EC-EARTH	X	X	X
IPSL-IPSL-CM5A-MR	X
MPI-M-MPI-ESM-LR	X

). In this research, the change in climate suitability and climatic conditions for maize cultivation in Poland was analysed under two Representative Concentration Pathways (RCP) scenarios: RCP4.5 (low emissions, LE) and RCP8.5 (high emissions, HE). RCP scenarios correspond to radiative forcing by 2100 relative to pre-industrial values of +4.5 Wm⁻² (RCP 4.5), and to +8.5 Wm⁻² (RCP 8.5) [37,38]. We selected equal 30-year periods for the reference period (1981–2010) and two future periods: the 2050s (2041–2070) and the 2080s (2071–2100) for our analyses.

Table 2. The area-averaged annual temperature, annual precipitation sum and their changes (in brackets) in Poland [26].

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.

shows the area-averaged annual temperature and annual precipitation sum in Poland and their projected changes in future climate conditions. According to the ensemble mean of six analysed RCMs, the mean annual air temperature in Poland was 8.2 °C for BL, which increased by 2.1 °C and 3.8 °C for LE and HE scenarios, respectively, by the 2080s. Furthermore, the mean annual precipitation sum was about 622 mm in BL, with projected increase of 9–14% and 11–19% in the 2050s and the 2080s, respectively. When using climate change scenarios, it is important to note that the results of numerical simulations may differ from observations. It is therefore essential to make comparisons between the two in order to assess the performance and capability of the climate models used [30,39]. The detailed information related to the comparison between observations and climate scenarios data used in this study can be found in our previous study [26]. The RMSE (root mean square error) of the annual mean values was 0.3 °C for the temperature and 36 mm/year (6%) for the precipitation [26].

2.3. Climate Suitability Assessment

The selection of different criteria for climate suitability assessment was based on a literature review in respect of their relation to maize growth conditions. The weights for the selected criteria were calculated using the AHP. Then, the suitability map was generated by application of a weighted overlay. The methodology employed is presented in the form of a flowchart in Figure 1.

2.4. Identification of Evaluation Criteria

The planting and growth of crops are affected by many factors, and therefore it is impossible to take all of them into account [23]. In accordance with maize growth conditions [4,40,41], and their potential changes due to climate change, the following criteria were selected for the evaluation of land suitability for maize cultivation in Poland: maturity, spring frost stress, heat stress and water stress. By means of FAO [42] and the literature review [20,21,23,43,44] the suitability classes are classified as follows: highly suitable (S1), moderately suitable (S2), marginally suitable (S3) and not suitable (N). Subsequently, a specific level of suitability for each criterion was defined and presented in

Table 3. Suitability level per selected criteria.

Characteristics		Suitability Rating Score *				References
Criteria	Sub-criteria	S1 100–85	S2 85–60	S3 60–40	N 40–0	[23]
Maturity of maize	FAO210 [number of years]	>27	24–27	18–24	<18	[45,46]
	FAO240 [number of years]	>27	24–27	18–24	<18	[45,46]
	FAO290 [number of years]	>27	24–27	18–24	<18	[45,46]
Frost stress	Tmin < 0 °C [number of years]	<3	3–6	6–12	>12	[46]
Heat stress	Tmax > 35 °C [number of years] Tmax > 40 °C [number of years]	<3	3–6	6–12	>12	[46]
		<3	3–6	6–12	>12	[46]
Water stress	CWB on very light soils [mm]	>−150	−200–−150	−250–−200	<−250	[47]
	CWB on light soils [mm]	>−210	−260–−210	−310–−260	<−310	[47]
	CWB on medium soils [mm]	>−250	−300–−250	−350–−300	<−350	[47]
	CWB on heavy soils [mm]	>−290	−340–−290	−390–−340	<−390	[47]

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

.

It is evident that climate exerts a significant influence on agricultural production potential [23,40]. Indeed, the necessity of agricultural management adjustment to local climatic conditions has been pointed out for the best use of these potentials. Due to changes in these conditions over time and space, it is important to consider them in the planning of adaptation measures [40]. Climatic suitability is largely influenced by heat, moisture and sunshine [23]. In agricultural practice, the temperature is typically regarded as the primary indicator that determines the sowing date [48]. The yield-forming effect of temperature is reflected in both the heat sum necessary to ensure all phenological phases occur and the temperature thresholds for critical phenological phases [49]. It is anticipated that extreme temperature events resulting from global warming will occur with greater frequency and greater intensity, which will have an impact on the growth and development of crops [50]. Therefore, frost and heat stress parameters are considered to be limiting factors for grain maize [40].

2.4.1. Maturity of Maize (GDD)

Maize is a thermophilic plant, and therefore the vegetation period for this crop should be relatively long with high values of degree-days [41]. The effective temperature sum that is needed for maize maturity depends on the type of earliness that is determined by the FAO number and the utilisation mode (for grain, CCM and silage). The choice of a variety with a desired length of the growing season ensures that maize can be cultivated on the entire area of Poland, with the exception of mountain regions [51]. It was assumed that the recommended average date of maize sowing is the first day when the average temperature over the last five days exceeds 11 °C. As the end of the maize growing season, the first day when the average temperature drops below 10 °C for five consecutive days was determined. A 6 °C threshold was used to calculate Growing Degree Days (GDD) for maize. By comparing its thermal requirements with GDD, it is possible to determine the possibility of reaching maize maturity [41]. In the following study, the number of years in which maize reaches maturity within a 30-year period is used as a suitable criterion for maize cultivation for three different maize maturity types: (1) FAO210, (2) FAO240 and (3) FAO290 as examples of early, medium-early and medium-late varieties.

2.4.2. Late Spring Frost Stress (LSF)

Maize is tolerant of a wide range of environmental conditions, although it is highly susceptible to frost [50,52,53]. Frost is a phenomenon that is difficult to predict and which can cause partial or complete damage to maize [4,54]. Frost stress occurs when the daily minimum temperature drops below 0 °C [40,54]. One of the most critical extreme events occurring in temperate regions, among others, is late spring frost (LSF), which affects plant performance. When the leaves are young, plants are most susceptible to frost, so the onset of freezing temperatures in late spring has a major impact on the growth, competitiveness and range of plants [55]. Consequently, frost occurrence was defined as one day with a minimum temperature below 0 °C from the sowing date until the end of June, and the frequency of years with at least one frost occurrence in a 30-year period was calculated.

2.4.3. Heat Stress (HS)

Maize is tolerant of high temperatures, although extremely high temperatures can reduce yield [4]. The optimum temperature differs among development processes, but is usually determined as 28–34 °C [56]. Daily maximum temperatures exceeding 35 °C can affect maize grain [40,57]. We defined heat stress as five consecutive days with a maximum temperature above 35 °C. Given that maize development can cease at temperatures of more than 40 °C, the heat stress temperature was additionally defined as one day with a temperature above 40 °C [4,56]. HS criteria were evaluated within sowing and end of growing season date.

2.4.4. Water Stress (WS)

The climatic water balance (CWB), employed in Poland as an indicator for meteorological and agricultural drought monitoring represents the difference between precipitation and potential evapotranspiration [58]. Given the observed increase in the frequency of water stress, the CWB was selected as one climate suitability criterion for maize cultivation [26,47]. Numerous methods for potential evapotranspiration (PET) calculation exist in the literature, with varying data requirements. The numerous data requirements allow for the potential for implementing simplifications in PET formulae based on specific application needs. The following study utilises the results of a previous study [26] on the calculated CWB sum for the period from April to September, based on a simplified Penman formula presented by Doroszewski and Górski [59] and Żyłowska et al. [60].

2.5. Determination of Weights (AHP)

An Analytical Hierarchy Process (AHP) method is used to determine the weights of selected main criteria and sub-criteria. The weights need to be chosen carefully as they reflect the relative importance of each criterion. Thanks to the AHP method, qualitative data, assessed by experts, can be converted into quantitative data [61].

In this study, a team of ten crop specialists (comprising five experts and five farmers) was asked to determine the relative importance of each criterion. The group of experts was selected on the basis of their extensive knowledge and experience in agriculture. It included scientists from the Institute of Soil Science and Plant Cultivation (IUNG-PIB), who conduct research in crops and agrometeorology, as well as head of crop units from the Agricultural Advisory Centre and the Regional Centre of Agricultural Advisory Services. The group of farmers comprised five directors and workers from the Agricultural Experimental Stations (AES IUNG-PIB), selected on the basis of their experience in leading long-term agricultural experiments across a range of voivodeships in Poland. A questionnaire was designed to compare the four main criteria pairwise regarding their importance in achieving maize suitability. Based on the proposed hierarchy (Figure 2), the sub-criteria are then compared pairwise in terms of their importance in achieving their upper-level criteria.

Table 4. Analytic Hierarchy Process linguistic pairwise comparison scale based on Saaty [62].

Degree 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

shows Saaty’s [62] linguistic scales for assessing the relative importance for evaluation of the selected criteria using AHP methodology.

Using the AHP as previously discussed, each specialist's opinion was transformed into a pairwise comparison matrix using a scale of 1–9 (Table 4). This scale was employed with 1 representing equal importance and 9 representing extreme importance between two compared criteria. Once the pairwise matrix had been generated, the weights for each criterion were computed using the eigenvector corresponding to the largest eigenvalue of the matrix (${\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}$), and then the sum of the components was normalized to unity [23,61]. The AHP methodology involves the identification and consideration of inconsistencies in decision-maker judgment through the evaluation of consistency relationship ($CR$), which depend on the Consistency Index ($CI$) and Random Index ($RI$) [61]. The $CI$ was calculated using the formula: $CI=({\mathsf{\lambda }}_{\mathrm{m}\mathrm{a}\mathrm{x}}-n)/(n-1)$ [63]. The random consistency index ($RI$) related to the number of elements ($n$) provided in

Table 5. Random Consistency Index ($\mathrm{R}\mathrm{I}$) values [62].

$\mathit{n}$	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

was used to determine the $CR$ based on the formula $CR=CI/RI$. A $CR$ value of 0.1 or less is deemed acceptable, although in some studies a value of 0.2 is considered an acceptable consistency level [64].

AHP can be used for priorities. According to Forman & Peniwati [65], there are two basic ways to aggregate individual opinions into a group opinion, depending on whether the group wants to act together as a unit (e.g., a group of department heads meeting to set corporate policy) or as separate individuals (e.g., a group consisting of representative constituencies with stakes in welfare reform, such as taxpayers, politicians, etc.). Several researchers find agreement in applying the AIJ method in case the target group is homogenous and assumed to act as one individual, while the AIP method is recommended when the group is heterogeneous and is seen as a collection of independent agents maintaining their own identities.

Given that the evaluation of criteria importance involved ten respondents with disparate opinions, the AHP group decision-making process was employed to aggregate individual judgments. There were alternative methods used to aggregate individual opinions applied in AHP methodology, depending on whether the group wanted to act together as a unit assuming consensus or as separate individuals. The Aggregation of Individual Judgments (AIJ) method is suggested in case the target group is homogenous whereas for heterogeneous groups the AIP method is recommended [65]. Calculated by the geometric mean to aggregate comparison matrices (AIJ) or to aggregate priority vectors (AIP) [66]. Srdjevic [67], who goes through a GDM process by utilizing a group of experts, split into interest subgroups, claims that AIJ is possible to apply only at the subgroup level. Among different subgroups only AIP is possible for final decision.

2.6. Weighted Overlay

There is a need to standardize the selected criteria as they are expressed in different units [61]. All selected criteria were reclassified into four suitability classes (S1, S2, S3, N) according to Table 3. Standardised criteria maps were generated using QGIS 3.28.4. Climate suitability for maize production was then assessed by combing all standardised criteria maps and weights of criteria. The final map was generated using the weighted overlay method in QGIS [44,61].

3. Results

3.1. Growing Season of Maize for Baseline and 2050s and 2080s Climates

The spatial distribution of the start (11 °C) and end (10 °C) dates of the growing season is presented in Figure 3 and Figure 4. The beginning and end dates of the growing season are delayed in colder regions (north-east) compared to warmer parts of Poland (south-west). This is seen in both the baseline timeframe and in the future climate. The projected warmer climate leads to an acceleration in the occurrence of the beginning and to a delay of the end dates of the growing season, with the rate of change depending on the time horizon and the RCP analysed. For the 2050s, the projected rate of acceleration of the start and delay of the end dates of the growing season is about 5–10 and 10–15 days for the LE and HE scenarios, respectively. In the 2080s, a higher advancement of about 5–15 days for LE and approximately 15–25 days for HE was projected.

The mean accumulated temperature sum above 6 °C (AT) during the maize growing season was higher than 1200 °C in the baseline period, except in southern Poland where the mean AT was lower (Figure 5). Due to climate change, the mean AT is projected to increase. According to the LE scenario, the AT during the maize growing season is projected to range from 1600 to 2000 °C in the 2050s and from 1800 to 2200 °C in most areas, whereas the AT will be much lower in the northern and southern parts of Poland. According to the HE scenario, the rise in AT is much higher, and by the end of the century, the mean AT is projected to rise up to 2600 °C in the west-central part of Poland.

3.2. Evaluation of Climate Criteria under Baseline and 2050s and 2080s Climates

The probability of the maturity of the different maize varieties is shown in Figure 6 as the reference number of years in which maize reached maturity per analysed 30-year period. The base line time period (BL) is highly diverse. In the northern and southern parts of Poland, the probability of reaching maize maturity is low (<60%) even for early-maturing varieties, and this area increases for later varieties. Furthermore, for early varieties, the probability of maturation is higher than 90% in 44% of Poland, but for medium-early varieties, this area decreases to 25% of Poland, while for medium-late varieties, it is almost non-existent (2%). The projected temperature increase results in greater thermal resources, allowing the cultivation of even medium-late varieties in a large part of Poland. However, there will still be regions where choosing a particular variety should involve considering the risk of a cold year, but for most of the area (from 65 to 98%) this criterion will not limit maize cultivation.

Frost stress (FS) was evaluated from the beginning (11 °C) of the growing season until the end of June, and its spatial distribution is presented in Figure 7. In the baseline time period, FS (minimum temperature below 0 °C) occurred mainly in 3–6 years out of the analysed 30-year time period (10–20%). The occurrence of spring FS decreases in future climate conditions. The area which is highly suitable according to spring frost risk, where the minimum temperature falls below 0 °C, will increase, and, by the 2080s, it will cover a large area of Poland (70% and 51% of Poland, according to LE and HE scenarios, respectively).

Heat stress was evaluated for the growing season and its spatial distribution is presented in Figure 8. In the baseline (BL) period, the maximum temperature did not exceed 35 °C for at least five consecutive days, while the maximum temperature was higher than 40 °C only for one day in a small area on the west part of Poland (6% of the area). In the future climate scenarios, the area with a maximum temperature higher than 35 °C for five days increased. According to RCP8.5, for the 2080s timeframe, it is going to occur up to 6 times per 30 years in a large area of Poland (97%). According to the frequency of occurrence of temperatures higher than 35 °C in Poland, there is a highly suitable area for the cultivation of maize in both current and future climate conditions, except for the 2080s for the HE scenario, where a large area of Poland (71%) is defined as moderately suitable. Furthermore, the frequency of the event when the maximum temperature exceeds 40 °C is expected to increase in the future. On the baseline, in Poland, there are highly suitable conditions for maize cultivation according to this stress. However, in the 2050s, in the western part of Poland, there is a large area defined as moderately suitable (this stress occurs in 3–6 years for 41–45% of Poland). According to the HE scenarios in the 2080s, it is projected that, in Poland, there will be marginally suitable and not suitable areas for maize cultivation in the western and central parts of Poland, with this stress occurring for more than 6 years in 62% and 1% of the area of Poland, respectively.

Water stress (WS) was evaluated in our previous study [26] on the basis of the CWB sum from April to September and its spatial distribution is presented there [26]. The CWB values were negative when the potential evapotranspiration exceeded the precipitation sum, which occured in a large area of Poland. The CWB exceeding 100 mm was observed only in mountainous areas. The area with CWB less than −300 mm was observed in central Poland and it is projected to increase in the future.

3.3. Climate Criteria Weights for Land Suitability Assessment

The opinions of experts (5) and farmers (5) were used to weigh the four main climate suitability criteria for maize (maturity, frost risk, heat risk and water stress) and sub-criteria using an AHP pairwise comparison matrix.

Table 6. AHP pairwise comparison matrix for the maturity of maize sub-criteria.

Subgroup	Experts				Farmers				Mean
Criteria	FAO210	FAO240	FAO290	Weights	FAO210	FAO240	FAO290	Weights	Weights
FAO210	1.000	0.284	1.600	0.190	1.000	0.367	4.441	0.283	0.232
FAO240	3.519	1.000	6.382	0.696	2.728	1.000	6.921	0.640	0.668
FAO290	0.625	0.157	1.000	0.114	0.225	0.144	1.000	0.077	0.093
	Consistency Ratio (CR) = 0.002				Consistency Ratio (CR) = 0.034

,

Table 7. AHP pairwise comparison matrix for the heat stress sub-criteria.

Subgroup	Experts			Farmers			Mean
Criteria	Tmax > 35 °C	Tmax > 40 °C	Weights	Tmax > 35 °C	Tmax > 40 °C	Weights	Weights
Tmax > 35 °C	1.000	3.956	0.798	1.000	5.471	0.845	0.822
Tmax > 40 °C	0.253	1.000	0.202	0.183	1.000	0.155	0.177
	Consistency Ratio (CR) = 0.000			Consistency Ratio (CR) = 0.000

,

Table 8. AHP pairwise comparison matrix for the water stress sub-criteria.

Subgroup	Experts					Farmers					Mean
Criteria	Very Light	Light	Medium	Heavy	Weights	Very Light	Light	Medium	Heavy	Weights	Weights
Very light	1.000	0.186	0.126	0.156	0.044	1.000	0.259	0.135	0.123	0.042	0.043
Light	5.373	1.000	0.223	0.410	0.151	3.866	1.000	0.163	0.145	0.089	0.116
Medium	7.950	4.478	1.000	1.165	0.458	7.384	6.128	1.000	1.516	0.468	0.463
Heavy	6.423	2.440	0.859	1.000	0.346	8.106	6.903	0.660	1.000	0.401	0.372
	Consistency Ratio (CR) = 0.046					Consistency Ratio (CR) = 0.079

and

Table 9. AHP pairwise comparison matrix for the main climate suitability criteria.

Subgroup	Experts					Farmers					Mean
Criteria	Maturity	Frost Stress	Heat Stress	Water Stress	Weights	Maturity	Frost Stress	Heat Stress	Water Stress	Weights	Weights
Maturity	1.000	2.460	4.004	0.678	0.3111	1.000	1.450	1.467	0.644	0.2092	0.2551
Frost stress	0.407	1.000	2.569	0.286	0.1431	0.690	1.000	1.217	0.138	0.1128	0.1271
Heat stress	0.250	0.389	1.000	0.136	0.0657	0.681	0.822	1.000	0.143	0.1029	0.0822
Water stress	1.476	3.497	7.331	1.000	0.4801	1.552	7.237	6.971	1.000	0.5751	0.5254
	Consistency Ratio (CR) = 0.007					Consistency Ratio (CR) = 0.064

shows the mean judgement of two sub-groups (experts and farmers) calculated based on geometric mean (AIJ). Then, the final weights were calculated based on AIP to aggregate two sub-groups.

Table 6, Table 7, Table 8 and Table 9 show consistency ratio (CR) values for main, maturity and water stress criteria, which were below the threshold value of 0.1, suggesting a reasonable level of consistency in the experts’ and farmers’ judgment. The judgement for heat stress was acceptable as there were only two sub-criteria compared (RI is 0). The analysed four climate factors that we indicated were as follows: probability of maturity, frost stress risk, heat stress risk and water stress risk. Water stress risk (53%) and probability of maturity (26%) were the main factors contributing to suitable areas’ evaluation for maize cultivation in Poland according to respondents’ opinion (Figure 9). The other two analysed climate factors, frost risk (13%) and heat stress risk (8%), were found to be less important but are still considered by experts. They most often decided to use the medium-early varieties (67%) rather than early varieties (23%) and medium-late varieties (10%). The heat stress indicators, with an event when the maximum temperature exceeds 35 °C for 5 days or more, were the most important (83%) in relation to an event when the temperature is more than 40 °C for at least one day (17%). In order to avoid water stress, the respondents were considering cultivating maize on medium (47%) and heavy soils (37%). However they were also considering cultivating maize on light (12%) and very light soils (4%), although they see climate risks. It is important to note that the calculated weights of the sub-criteria and main criteria differ between the two sub-groups (experts and farmers). Therefore, if we do not aggregate them and produce suitability maps for different groups, the results may vary.

3.4. Climate Suitability for Maize in the Baseline and Future Scenarios

The presented suitability maps of the main climatic criteria (maturity, frost stress, heat stress and water stress) as well as the final climatic suitability maps were classified into four suitability classes: high (S1), moderate (S2), marginal (S3) and not suitable (N).

Table 10. Percentages of different climate suitability levels for maize cultivation in Poland according to maturity, frost stress, heat stress, water stress for the baseline (1981–2010) and projected climate according to low emissions (LE, RCP4.5) and high emissions (HE, RCP8.5) scenarios for the 2050s (2041–2070) and 2080s (2071–2100).

RCP					LE								HE
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
Maturity	43	27	19	11	94	3	1	2	96	1	2	1	100	0	0	0	100	0	0	0
Frost stress	24	65	11	0	46	53	1	0	70	30	0	0	51	47	2	0	51	46	3	0
Heat stress	100	0	0	0	100	0	0	0	100	0	0	0	100	0	0	0	26	71	3	0
Water stress	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.

and Figure 10 show the percentages of the suitability classes of Poland’s area.

From the reclassification results (Figure 11), taking into consideration maturity, it is shown that around 43% of the study area was highly suitable, whilst 27% was moderately suitable and 19% was marginally suitable for maize cultivation. However, about 11% of Poland was classified as unsuitable. Furthermore, as a result of the future increase in the accumulated temperature sum in both scenarios, the highly suitable areas will increase and it is projected that they will cover more than 90% of Poland, allowing cultivation of later maturing varieties than in the current climate. The obtained results show that frost stress limits climate suitability for maize production in Poland; for the baseline 24% of country area is highly suitable, 65% moderately suitable, while 11% is marginally suitable for maize cultivation. In the future, frost stress is predicted to decrease, and more than 50% of Poland is going to be highly suitable with the largest share of 70% of the area in the 2080s for RCP4.5. Considering heat stress, the entire analysed area was highly suitable under the current climatic conditions, and it will remain so in the future except for the HE scenario, in which heat stress will occur more frequently in the 2080s than in the past, thus reducing the highly suitable areas to moderately and marginally suitable in 71 and 3% of Poland, respectively. Due to the occurrence of water deficiency in the current climate, only central Poland is limited by water stress and has a moderate or marginal climate suitability (19%) for maize production. In the future, depending on the period and scenario analysed, these conditions are projected to extend to 30–45% of Poland. The increase in water stress will continue into the far future, and by the end of the century, according to the HE scenario, 21% of the area of Poland will be marginally suitable for maize production, and 11% will be unsuitable.

Finally, the climate suitability analysis (Figure 10 and Figure 11) using the AHP weights shows that 62% of Poland was highly suitable and 38% was moderately suitable under current climate conditions. Furthermore, in the future climate conditions the highly suitable area will increase to 69–73%, except in the 2080s according to the HE scenario where this area will decrease to 47%. The rest of Poland is projected to be moderately suitable according to the selected climate factors.

4. Discussion

The climate in Poland and Europe is changing, with both average air temperatures and the frequency of extreme weather events increasing [68]. Observed climate changes may have both positive and negative effects on plant development. Projected warming may lead to a decrease in wheat yield, while it may increase in maize yields [69].

With future climate change and projected temperature increase, an acceleration of the beginning and delay of the end dates for the maize growing season are observed in the future reaching 15–25 days by the 2080s for the RCP8.5 scenario. These alterations will allow for the prolongation of the growing season, which will, in turn, allow the introduction of longer-duration cultivars [70]. A longer growing season can be associated with an increase in productivity and therefore an increase in nutrient requirements [70]. Furthermore, in both current and future climate conditions, the geographical diversity in the specific dates of the beginning and end of the growing season should be noted. The delayed beginning date and an accelerated end of the growing season in the northern part of the country was observed compared to the other parts of the country. This is also reflected in the study conducted by Nieróbca et al. [9], where the south-eastern part of Poland has more suitable climatic conditions for maize cultivation than northern regions [9]. Another benefit of climate change is the increase in thermal resources, allowing the cultivation of even medium-late varieties in a large part of Poland. However, even under future climate conditions, in a small area of the country in regions with a lower probability of maturity (less than 80%), the risk of a cold year should still be taken into account when choosing a variety [45].

Due to climate change and temperature increase, the number of frost days is projected to decrease in Poland [1,30]. Żarski et al. [71] assessed changes in climate risk indicators, such as the occurrence of late spring frost for grain maize in the Bydgoszcz region. They showed that late spring ground frost in May and June occurred 83% of the time in this region during the period analysed (between 1985 and 2014). However ground temperature below −2 °C occurred 30% of the time [71]. However, when the near-future was analysed (2026–2050), an increase in frost days compared to the historical period was predicted for central Poland [30]. Regarding frost risk that increased in the recent years, the phenological strategies (sowing later with respect to thermal requirements) that helped trees tolerate past frost frequencies will thus be increasingly mismatched with future conditions [55].

Despite the fact that maize is a thermophilic plant, the increased risk of heat that has caused plants injuries is reported by Ioannis Charalampopoulos [72]. For BL, the maximum temperature did not exceed 35 °C for 5 consecutive days and exceeded 40 °C for one day in just one year for a small area in Poland. However, the number of years with these extremes will increase in the future, and according to RCP8.5, heat stress will limit the suitability for maize cultivation in the 2080s. Previous studies have also assessed that the number of extremely hot days with Tmax above 35 °C will increase in the future in Poland [30]. In cooler zones of Europe (including Poland) lower heat stress is observed for maize than for wheat and barley whereas it is higher in warmer zones compared to current conditions [70]. Olesen et al. [70] noted that the occurrence of heat stress in the summer could be reduced by adopting earlier sowing dates.

Our study indicates that the most important element in assessing the climate suitability of areas for maize cultivation in Poland is water stress. A useful indicator for measuring this stress is climatic water balance, i.e., the difference between total precipitation and total evaporation. At present, evapotranspiration exceeds summer precipitation in almost all parts of Poland, resulting in an increase in water deficit [27]. A negative trend in availability of soil water during summer months and possible important effects on yield in central and eastern Europe is indicated by Pinke et al. [73], with warning about grain production stability and needs for intervention including implementation of water conservation agricultural practices. Olesen et al. [70] observed that water-conserving tillage practices can be a crucial adaptation measure in both warm and dry zones. In the preceding study [26], the impact of irrigation on water stress in maize cultivation was evaluated, and it was found that this practice is an appropriate method for adapting the increase in water stress related to climate change. According to work by Gobin et al. [74], the observed risk of drought for grain maize shifts earlier in the season. This is due to the fact that spring drought affects earlier vegetative stages, while summer drought poses a risk to anthesis. The development of smart irrigation systems and water management strategies is required, however relatively little research effort in this direction was noted by Lopes [75].

Our overall assessment of the consequences of climate change shows an improvement in growing conditions in northern Poland and a worsening in the central part of the country. However, the projected worsening of some climate conditions does not imply a decline in maize yields, at least according to the 2050 perspective, as pointed out by Parent et al. [76]. This trend may be positive or unchanged compared to the baseline if farms adopt upgraded practices regarding choosing varieties with optimal sowing dates and growing cycle lengths, and using the best of genetic variability suitable to the local environment. Ramirez-Cabral et al. [4] found crucial the importance of evaluating the regional climatic suitability for maize production variation under climate change. The ecoclimatic index (EI) was evaluated using a CLIMEX distribution model at the global level, assuming an A2 emissions scenario. In the current climate, the majority of Poland was identified as having medium climate suitability, with a small area in the centre indicated as having marginal suitability. In future climate conditions, the CSIRO model predicts that by 2050, the majority of Poland will be suitable for maize cultivation, with marginal climate suitability in central Poland that will increase in 2100. In contrast, the MIROC model indicates that the climate suitability for maize cultivation will be enhanced by climate change with the majority of Poland becoming optimal for this purpose by 2100 [4].

It should be noted that in our assessment we did not consider the spread of pests, weeds, or diseases among crops, which are also related to new climatic conditions, so it is relatively unknown how these major pathways of crop failure may amplify each other in the future [77]. Furthermore, our analysis did not consider non-climatic factors (e.g., topography, soil conditions and land use) or the potential genetic advancement of the species. Suitability may be significantly influenced by soil characteristics, particularly organic matter [78]. It would be beneficial for future studies to include such an analysis [4]. In our study we adapted the sowing date of maize to the predicted climatic conditions, which was not a fixed day of the year, therefore the predicted conditions for maize cultivation in relation to frost stress will improve in the future. In future studies it would be advisable to see how the sowing date affects these conditions. Karapetsas et al. [79] evaluated the effectiveness of SICSs (Soil-Improving Cropping Systems) as a long-term mitigation measure against the impacts of climate change on land suitability for maize, in addition to assessing current suitability for maize. There are some methods that can be used in combination in MCDA to provide more comprehensive assessments in agricultural land evaluation. Agrawal et al. [80] proposed a machine learning (ML)-based approach as a valuable method for assessing the suitability of agricultural land. Rangzan et al. [81] used an integration of Fuzzy logic, the Analytical Hierarchy Process (AHP) and satellite images in identifying optimal areas for wheat cultivation. AHP is a multi-criteria method based on pairwise comparison. It relies on expert judgement to determine priority scales, based on the principle that people's experience and knowledge are at least as valuable in decision making as the data they use [82]. More complexity and uncertainty can be brought to the process with large group of decision makers [83]. As the values of the criteria weights have an impact on the final suitability assessment, the choice of group and number of decision experts is challenging and varies between studies. Han et al. [23] noted that consultation with several experts is required to assign weights to selected criteria using an AHP method. Rodcha et al. [84] invited seven experts on crop plantations to determine the weights. Morales and de Vries [85] established the criteria weights with 20 experts. Some authors evaluated the weights based on previous studies, the literature and expert judgment [25,44]. Therefore, the group of experts could be expanded in future studies. A sensitivity analysis of the influence of the weights given by different respondents or different groups could be carried out in the future for a detailed analysis of the results [86].

5. Conclusions

The analysis of climate suitability for maize is important for increasing the capacity to cope with the impacts of climate change on maize production. In this study, climate suitability for maize was assessed for the baseline period and two future periods (2050s and 2080s). The climate risk indicators were assessed with possible sowing and harvesting dates set based on air temperature in order to adapt the crop to climate conditions and provide more realistic results.

A deeper understanding of the pros and cons of climate change impacts on maize development was achieved through the use of multi-criteria evaluation. The criteria chosen reflected the positive effects of climate change, i.e., the increase in thermal resources, and the negative effects related to the frequency of extreme weather events (frost, high temperatures, water stress). The use of a mean ensemble from six GCM-RCM climate simulations reduced the uncertainty in projections of further climate. Additionally, the participation of crop experts and farmers was important for incorporating their knowledge in the evaluation of criteria weights, having in mind that this evaluation is also subjective. Our main findings indicated that areas of high climate suitability for maize production in Poland are expected to increase due to increase in maturity achievement and a reduction in frost stress, with the exception of the 2080s forecast under scenario RCP8.5. In this scenario, a reduction in climate suitability is projected due to an increase in heat and water stress.

The results and methodology used can be useful in planning adaptation measures for climate change conditions. These changes in climatic conditions will require the implementation of adaptation agricultural techniques. Heat and water stress may limit the suitability of maize in the future, but these negative effects of climate change can be attenuated with new varieties that are resistant to increasingly frequent heat and drought stress. On the other hand, greater thermal resources as a result of climate change provide opportunities to grow maize varieties with higher heat requirements with less risk of not reaching maturity. Projected higher air temperatures will result in the acceleration of sowing dates and the delay of harvesting dates, affecting the potential scheduling of field operations, the possibility of introducing longer-duration cultivars, as well as increased productivity.