Dependence of the Pea Grain Yield on Climatic Factors under Semi-Arid Conditions

Abstract

Field peas are one of the most common crops and are grown in various climatic zones. However, the productivity of this crop can be largely limited by climatic factors. This study investigated the influence of climatic factors on pea grain yield in the semi-arid conditions of the Rostov region of Russia in 2008–2020. To quantify climatic factors, agro-climatic variables were used, such as total temperatures below the minimum temperature, the number of days with temperatures below the minimum temperature, total temperatures above the critical temperature, the number of days with temperatures above the critical temperature, and the Selyaninov hydrothermal coefficient. Agro-climatic variables were calculated using daily climatic variables, such as maximum and minimum temperatures, relative air humidity, and precipitation during pea growing season (April–June). The yield of the pea varied from 90 to 250 kg/ha. In general, the productivity of peas is negatively affected by high temperatures and low humidification level. The yield is negatively correlated with accumulative temperatures above the critical temperature and the number of days with temperatures above the critical temperature and positively correlated with the Selyaninov hydrothermal coefficient and the precipitation in all analyzed areas. The influence of the accumulative temperatures above the critical temperature is the most significant. It explains between 6.6% and 78.9% of the interannual variability of the pea yield. The increase in accumulative temperatures above the critical threshold by every 1 °C will contribute to a decrease in pea grain yield by an average of 0.150 kg/ha. The maximum temperatures in May and June (the period of flowering–grain filling) have the most negative impact on the yield. A 1 °C increase in the average maximum temperature during this period will contribute to a decrease in pea yield by an average of 19.175 kg/ha. The influence of total precipitation during the growing season explains between 12.3% and 50.0% of the variability. The 1 mm decrease in the total precipitation for the growing season will lead to a decrease in pea yields by an average of 0.736 kg/ha. The results of this study can be applied to regional yield forecasting, as well as predicting the impact of climate variability on the grain yield of pea crops in arid areas.

1. Introduction

Pea (Pisum sativum L.) is one of the world’s essential legumes and is currently cultivated in more than 95 countries [1]. According to the FAO, the annual world production of pea grain is approximately 15.7 million tons. Due to their relatively high yields and relatively low production costs, peas are the most widely used alternative protein source, and demand for this crop has only increased in recent decades [2,3]. However, pea production is limited by climatic factors. High temperatures cause reduced productivity in legumes [4,5,6,7], especially in arid and semi-arid regions [8]. Studies show that the negative impact of this factor has already led to a decrease in the gross harvest in Europe [9,10]. For example, in Ukraine, pea yields decreased by an average of 13.7% over the period of 2001–2020 as a result of increased temperature conditions [11]. It is estimated that climate change in southwestern France could reduce grain yields of spring peas by 28% [12]. The mechanisms of the influence of heat stress in different periods of the growing season of pea plants have been studied relatively well [13]. At the same time, it is noted that heat stress in the reproductive phase of pea development is the most significant factor that reduces yield, especially in regions experiencing moisture deficiency [14,15,16]. Dreccer et al. [17] found that field pea yield across all Australian agro-climatic zones was generally consistently negatively associated with maximum temperature and the number of days with extreme temperatures (T_max > 30 °C) during a given developmental phase. Using the drylands of western Canada (Saskatoon) as an example, Bueckert et al. [18] assessed the quantitative dependence of pea yield on the duration of periods with extreme temperatures. The authors showed that an increase in the number of hot days with temperatures above 28 °C, from 14 to 26 days per season, reduced pea productivity in the study region by almost half.

Moisture deficiency during the growing season is another significant factor that reduces the yield of legumes [19]. Studies show that legumes are at low risk of multiple temperature stresses due to their short growing cycle, but drought remains the dominant factor in reducing yields [20]. For example, in most countries of West Africa, due to the increase in the duration of dry periods, the yield of legumes will decrease by 22–24% already in the current decade [21]. In general, in arid regions, field pea productivity generally increases markedly in response to increased precipitation during the growing season [22]. However, according to forecasts, climate changes, expressed primarily in a significant temperature increase, are predicted to result in a higher frequency of adverse climatic events [23]. This fact is of concern since such changes can significantly negatively affect the yield of crops, including peas [24,25].

As one of the leaders among the agricultural regions in Russia, the Rostov region makes a significant contribution to ensuring the country’s food security. The gross harvest of pea in the region in recent years has been approximately 253 thousand tons [26]. In recent decades, there has been a marked increase in pea acreage in the South of Russia. In the Rostov region, the pea growth area annually increases by about 3.125 thousand hectares [27]. However, the negative impact of climatic conditions on pea production is also typical for this region. For example, Lysenko [28] reports an increase in pea yield with an increase in moisture in the Azov agro-climatic zone. At the same time, the growing season temperature regime negatively affects the yield. Independent studies have confirmed these observations [29]. Pea production in the Rostov region is predominantly rainfed, and the whole region’s moisture regime is characterized as arid (dry) and semi-arid (the multi-year average HTC indicator is 0.71 [30]). At the same time, the humidification regime of the region is also characterized by both significant spatial heterogeneity and sharp interannual contrasts [31]. In addition, studies of climate trends show that the regions of the south of Russia (including the Rostov region) are characterized by a significant increase in heat supply and the number of days with stressful temperatures for plants. However, no significant increase in moisture supply is observed [32]. Based on the greenhouse gas concentrations model, a further increase in annual minimums by 4–6 °C is predicted by 2041–2060, accompanied by an overall decrease in precipitation (http://neacc.meteoinfo.ru/research/91-change-climat21-eng (accessed on 22 September 2023)). In this regard, agricultural producers may face an even more significant reduction in pea harvests caused by the influence of climatic factors.

In this study, using the semi-arid conditions of the Rostov region as an example, we analyze the dynamics of field pea grain yield and evaluate its relationship with climatic factors such as temperature and moisture availability. We focus on assessing the degree of influence, as well as establishing the quantitative correlation between pea grain yield and a specific climatic factor. The findings may be useful for decision making by field pea producers in other semi-arid and arid regions.

2. Materials and Methods

2.1. Study Area, Yield, and Meteorological Data

Rostov region is located in the southern part of the East European Plain and partly in Ciscaucasia (45°58′–50°13′ N and 38°11′–44°20′ E). The Rostov region’s climatic conditions are generally suitable for cultivating most crops. However, the region’s climatic conditions, both in terms of temperature and precipitation, are heterogeneous (Figure 1).

Under such conditions, it is evident that the degree of influence of climatic factors on pea yield varies from region to region. With this in mind, we considered data on pea grain yields for all administrative districts of the Rostov region in 2008–2020, provided by the Federal State Statistics Service (Rosstat) (https://rosstat.gov.ru/ (accessed on 22 September 2023)). At the same time, data on Aksai (AK), Millerovsky (MI), Neklinovsky (NE), Salsky (SA), and Tsimlyansky (TS) districts were selected as control locations. The main selection criteria were localization in different parts of the region, sufficient completeness of data on pea grain yield in 2008–2020, as well as the presence or proximity of the location to a meteorological station (Figure 1). Thus, weather data from the meteorological station Rostov-on-Don for AK, meteorological station Chertkovo for MI, meteorological station Taganrog for NE, meteorological station Gigant for SA, and meteorological station Tsimlyansk for TS were used to characterize climatic factors. For the marked meteorological stations, weather daily data of maximum temperature (T_max), minimum temperature (T_min), and precipitation (Pre), as well as relative air humidity (Rh) were obtained from the World Data Center of the Russian Research Institute for Hydrometeorological Information (RRIHI-WDS) and were used [33].

2.2. Quantifying the Influence of Climatic Factors on Pea Yield

In this study, the climatic factors are represented by quantitative agro-climatic indicators and climatic variables characterizing the temperature regime and the regime of providing moisture for the growing pea season in the Rostov region. Applying agro-climatic indicators of the entire growing season makes it possible to fully assess the impact of climatic factors on crop yields since their calculations often take into account the biological characteristics of a particular crop [17]. In this study, we used an approach in which the agro-climatic indicators of the temperature regime were calculated as the accumulated daily temperatures during the sowing–harvesting period, which do not correspond to the optimal temperature range for the development of peas [19]. These indices included the total minimum daily temperatures when the minimum temperature reached the T_min threshold (TLB, °C), the number of days when the minimum temperatures were below T_min (DLB, days), the total maximum daily temperatures when the maximum temperature exceeded the T_critical threshold (TAC, °C), and the number of days with maximum temperatures above T_critical (DAC, days). Obviously, the optimal temperature range boundaries for peas vary due to several other factors (genetics, soil, humidity, etc.). Nevertheless, many studies show that even when the minimum temperatures drop below 5 °C or the maximum temperatures exceed 26 °C, degradation of the growth and development of pea is noted, and, as a result, the yield decreases [18,34,35,36,37,38]. Therefore, we took the values for T_min = 5 °C and T_critical = 26 °C as threshold temperature values for calculating agro-climatic indices. Therefore, agro-climatic temperature indices were calculated using the following equations:

$$\mathrm{T}\mathrm{L}\mathrm{B}={\sum }_{\mathrm{S}\mathrm{o}\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{g}}^{\mathrm{H}\mathrm{a}\mathrm{r}\mathrm{v}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}}{\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}\le 5 °\mathrm{C}$$

(1)

$$\mathrm{D}\mathrm{L}\mathrm{B}={\sum }_{\mathrm{S}\mathrm{o}\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{g}}^{\mathrm{H}\mathrm{a}\mathrm{r}\mathrm{v}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}}{\mathrm{d}\mathrm{a}\mathrm{y}}_{{\mathrm{T}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}\le 5 °\mathrm{C}$$

(2)

$$\mathrm{T}\mathrm{A}\mathrm{C}={\sum }_{\mathrm{S}\mathrm{o}\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{g}}^{\mathrm{H}\mathrm{a}\mathrm{r}\mathrm{v}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}}{\mathrm{T}}_{\mathrm{m}\mathrm{a}\mathrm{x}}\ge 26 °\mathrm{C}$$

(3)

$$\mathrm{D}\mathrm{A}\mathrm{C}={\sum }_{\mathrm{S}\mathrm{o}\mathrm{w}\mathrm{i}\mathrm{n}\mathrm{g}}^{\mathrm{H}\mathrm{a}\mathrm{r}\mathrm{v}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{i}\mathrm{n}\mathrm{g}}{\mathrm{d}\mathrm{a}\mathrm{y}}_{{\mathrm{T}}_{\mathrm{m}\mathrm{a}\mathrm{x}}}\ge 26 °\mathrm{C}$$

(4)

where T_max and T_min are maximum and minimum temperatures (°C). Planting and harvesting dates of peas in the Rostov region fall in the period April–June. Cultivation of crop occurs in rainfed conditions. To assess the moisture regime during the planting–harvesting period, the generally accepted agro-climatic indicator Selyaninov’s hydrothermal coefficient (HTC) was used, which was calculated using the following equation:

$$\mathrm{H}\mathrm{T}\mathrm{C}=\frac{10\ast {\mathrm{P}\mathrm{r}\mathrm{e}}_{>10}}{\sum {\mathrm{T}}_{>10}}$$

(5)

where Pre_>10 is the annual precipitation for the period with average daily temperatures above 10 °C (mm) and ΣT_>10 is the cumulative average daily temperatures for the same period (°C). Average daily temperatures are calculated as the average between T_max and T_min. To determine the periods of the most significant influence, the correlation between pea yield and climatic variables for each month (April, May, and June) and the entire growing season (April–June) was also evaluated. Monthly averages of T_max and T_min represent climate variables, and Rh and monthly Pre. Thus, using agro-climatic indices and climatic variables will make it possible to assess the contribution of climatic factors to pea grain yield.

2.3. Statistical Analysis and Interpolation

For each control study area in 2008–2020, the relationship between pea grain yield and weather indices were estimated using the Pearson correlation coefficient at a statistical significance level of p < 0.05. Also, correlation between pea yield and the value of each index is represented by simple linear regression. The parameters of the linear regression can be estimated using the least squares method, and the estimated model is

$$\mathrm{p}\mathrm{e}\mathrm{a}\mathrm{y}\mathrm{i}\mathrm{e}\mathrm{l}\mathrm{d}={\mathsf{\alpha }}_0+{\mathsf{\alpha }}_1\ast \mathrm{c}\mathrm{l}\mathrm{i}\mathrm{m}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{x}$$

(6)

where α₀ and α₁ are intercept and regression coefficient, respectively. ${\mathsf{\alpha }}_1$ is the regression coefficient (or the slope of the regression line) that captures the magnitude of the impact of changes in the weather predictor index on pea yield. To assess the impact of each weather index on pea yield across different periods, a coefficient of determination (R²) analysis was used. Spatial representation and analysis of pea yield and weather data were carried out using the Spatial Analyst module in ArcGIS 10.4. The interpolation analysis was carried out using an inverse distance weighted (IDW) technique. IDW is one of the most popular and frequently used forms of deterministic spatial interpolation of both climatic and other data [19]. The interpolation of pea grain yield was made based on the data on the yield of the administrative districts of the Rostov region; the completeness of data for the period 2008–2020 was at least 60%.

3. Results

3.1. Spatial Distribution and Dependence of Pea Grain Yield on Agro-Climatic Indicators

Between 2008 and 2020, a trend toward an increase in pea grain yield was recorded for all studied areas (Figure 2). The distribution of yields and agro-climatic indices in the Rostov region has high spatial variability (Figure 3). High spatial variability of yields is caused by a combination of factors: climate, soil, genetics, etc.

Nevertheless, when the distribution of yields and the variability of the values of agro-climatic indices are compared, it can be seen that the minimum yield in the range of 90–110 kg/ha was observed in the eastern part of the region, which is characterized by the most extreme climatic conditions. The value of HTC from April to June in this region is minimal and does not exceed 0.97.

At the same time, TAC and DAC are at least 1163.7 °C and 37.9 days, respectively. The maximum yield is typical for the southern and southwestern parts of the region (230–250 kg/ha), where the values of TAC, DAC, TLB, and DLB mainly do not exceed the values of 1086.9 °C, 35.7 days, 29.7 °C, and 14.6 days, respectively, and the value of HTC ranges from 0.97 to 1.10 (Figure 3). As we move to the northern part of the region, where the pea yield also decreases to 110–130 kg/ha, the TAC and DAC values are minimal. They do not exceed 971.8 °C and 32.5 days, respectively, and the HTC value increases to a maximum of 1.23. This moisture level in the northern part of the region, in contrast to other parts, against the background of the minimum values of TAC and DAC, is probably generally sufficient for pea vegetation.

This conclusion is also confirmed by the results of a correlation analysis of the dependence of pea yield on agro-climatic indices for the control districts of the region from 2008 to 2020 (Figure 4). For example, in AK, NE, SA, and TS, a significant (p < 0.05) positive relationship (r = 0.59, r = 0.66, r = 0.65, r = 0.62, respectively) was found between pea grain yield and HTC. In MI, this relationship is also positive (r = 0.26), but it is not significant.

A negative effect of accumulative maximum temperatures on pea yields was also found (Figure 4). Thus, in all the studied areas, a negative relationship was noted between pea yield and TAC indices (r = −0.67, r = −0.70, r = −0.69, r = −0.89, and r = −0.26 for AK, MI, NE, SA, and TS, respectively) and DAC (r = −0.65, r = −0.70, r = −0.69, r = −0.89, and r = −0.27 for AK, MI, NE, SA, and TS, respectively). At the same time, this relationship is significant in AK, MI, NE, and SA (p < 0.05).

The relationship between pea grain yield and TLB and DLB indices is positive in most cases, but it is not significant. It is worth noting that minimum temperatures below five °C are mainly characteristic of the early stages of pea development (germination and sprouts appearance). They lead to delayed germination, poor crop formation, reduced competitiveness with weeds, and reduced grain yield. In our study, TLB and DLB indices of all districts were observed mainly in April and extremely rarely in May.

3.2. Correlation between Pea Grain Yield and Monthly Climatic Indicators

The distribution of monthly climatic variables in the Rostov region from 2008 to 2020 also showed significant spatial variability (Figure 5).

At the same time, the distribution of average monthly maximum and minimum temperatures does not undergo significant spatial changes. Their value increases evenly from the beginning to the last month of the growing season, while the amount of precipitation and humidity during the growing season show significant spatial variability from month to month.

Analysis of the dependence of pea yield on climatic variables during the study period generally showed a negative effect on the temperature factor and a positive effect on moisture factors (Figure 6). In April, when seedlings are actively formed in the plant, the effect of maximum and minimum temperatures on yield in the control areas of this study is multidirectional and not significant. However, the effect of precipitation and humidity in April is generally predominantly positive for all parts, which may indicate a lack of moisture for the entire region during this period. At the same time, there is a significant (p < 0.05) relationship between yield and precipitation in April in the southwestern part of the region in AK and NE (r = 0.69 and r = 0.59, respectively) and a noticeable (r = 0.48) but not significant relationship in the southern part in SA. There was also a noticeable but not significant relationship between yield and moisture in AK (r = 0.49) and significant (p < 0.05) in NE (r = 0.57).

However, the most noticeable negative impact on the yield of peas was exerted by the temperature regime in May, when the plant passed the flowering phase and seed filling began. During this period, for all control areas, a negative relationship between yield and minimum and maximum temperatures was noted (Figure 6).

It is also noteworthy that a significant relationship (p < 0.05) with the maximum temperature was noted for the AK, NE, and SA regions (r = −0.59, r = −0.58, and r = −0.85, respectively) located in the southern and southwestern parts. In the northern part of the region, the flowering and seed-filling phase occurs later and falls approximately in mid-May and early June. Therefore, MI also showed a significant negative relationship (p < 0.05) between pea yield and maximum temperature in June (r = −0.60). Correlation analysis showed a positive but not significant relationship between yield and precipitation in May for all five control areas, which may indicate a generally acceptable level of moisture in this period (Figure 6). At the same time, the relationship between yield and relative air humidity in May is also positive for all regions, and in the southern (SA) and northeastern (TS) parts of the region, this relationship is significant (p < 0.05) (r = 0.67 and r = 0.61, respectively). In June, during the grain filling and peas ripening period, a noticeable positive but not significant relationship between yield and relative air humidity was observed for each region.

4. Discussion

Our results generally agree with previous studies in the Rostov region. For example, Korobova et al. [39] and Lysenko et al. [40] also showed a positive dependence of pea yield on moisture level during the growing season in the Azov agro-climatic zone of the region. Insufficient moisture levels caused by limited rainfall are the main cause of poor and unsynchronized emergence, affecting plant density’s uniformity and negatively affecting yields [41,42]. At the same time, it is also noted that the current temperature regime in the region harms pea productivity. It is noteworthy that a similar influence of climatic factors was noted in other regions of Russia [43]. In the conditions of the Ural region (Republic of Bashkortostan), Kuznetsov et al. [6] found that the highest yield of field peas was observed in years when the temperature during the growing season was 0.81–1.10. Similar patterns for legumes have also been identified in other regions. In East Azerbaijan province (Iran), the highest chickpea yield (~680 kg/ha) was observed in the southeastern part of the region, where the sum of critical temperatures (T_max ≥ 29 °C) and the number of days with such a temperature in the season did not exceed 27.5 °C and 16 days [19]. At the same time, the authors also emphasize that increased precipitation has a positive effect, while temperatures have a negative effect on the yield of the crops studied.

Based on the calculated linear regression equations and the value of the coefficient of determination (R²), it can be concluded that high temperatures, generally, are the most significant negative factor determining the yield of peas in the Rostov region (Figure 7 and

Table 1. Regression equations for the correlation of field pea grain yield and climatic variables of maximum (T_max, °C) and minimum (T_min, °C) temperatures, precipitation (Pre, mm), and relative humidity (Rh, %) for each month and the entire growing season (April–June) in five control locations of the Rostov region in 2008–2020.

Control District	Climatic Variable	April	May	June	April–June
AK	Tmax	y = −8.148x + 303.64 (R2 = 0.054)	y = −18.330x + 597.51 (R2 = 0.351)	y = −24.397x + 870.37 (R2 = 0.285)	y = −29.313x + 845.43 (R2 = 0.365)
	Tmin	y = −7.139x + 208.28 (R2 = 0.040)	y = −26.208x + 492.15 (R2 = 0.417)	y = −18.653x + 484.19 (R2 = 0.121)	y = −32.173x + 548.85 (R2 = 0.300)
	Pre	y = 1.779x + 99.18 (R2 = 0.478)	y = 0.066x + 159.54 (R2 = 0.002)	y = 0.325x + 150.12 (R2 = 0.022)	y = 0.361x + 110.66 (R2 = 0.133)
	Rh	y = 2.503x + 10.60 (R2 = 0.063)	y = 1.349x + 78.09 (R2 = 0.020)	y = 3.629x − 36.99 (R2 = 0.183)	y = 4.603x − 112.73 (R2 = 0.139)
MI	Tmax	y = 4.921x + 45.32 (R2 = 0.018)	y = −11.857x + 403.92 (R2 = 0.170)	y = −22.775x + 765.18 (R2 = 0.361)	y = −24.712x + 684.27 (R2 = 0.220)
	Tmin	y = 8.557x + 89.096 (R2 = 0.052)	y = −16.288x + 297.70 (R2 = 0.152)	y = −15.670x + 355.4 (R2 = 0.106)	y = −11.73x + 240.69 (R2 = 0.039)
	Pre	y = 0.440x + 106.99 (R2 = 0.039)	y = 0.595x + 93.81 (R2 = 0.093)	y = −0.459x + 144.81 (R2 = 0.019)	y = 0.602x + 42.62 (R2 = 0.123)
	Rh	y = 4.492x − 171.48 (R2 = 0.177)	y = 1.175x + 49.72 (R2 = 0.014)	y = 3.128x − 51.23 (R2 = 0.112)	y = 4.455x − 152.46 (R2 = 0.129)
NE	Tmax	y = −9.569x + 381.80 (R2 = 0.055)	y = −18.638x + 662.33 (R2 = 0.337)	y = −24.222x + 925.72 (R2 = 0.241)	y = −27.933x + 859.59 (R2 = 0.315)
	Tmin	y = 3.800x + 203.12 (R2 = 0.006)	y = −21.902x + 547.12 (R2 = 0.298)	y = −11.515x + 450.43 (R2 = 0.034)	y = −25.601x + 581.06 (R2 = 0.128)
	Pre	y = 1.500x + 173.63 (R2 = 0.345)	y = 1.114x + 174.41 (R2 = 0.255)	y = 0.213x + 224.10 (R2 = 0.006)	y = 1.030x + 101.61 (R2 = 0.050)
	Rh	y = 6.112x − 172.06 (R2 = 0.321)	y = 3.984x − 25.33 (R2 = 0.153)	y = 4.498x − 22.16 (R2 = 0.227)	y = 7.410x − 230.28 (R2 = 0.349)
SA	Tmax	y = −10.551x + 384.11 (R2 = 0.136)	y = −22.877x + 742.22 (R2 = 0.714)	y = −18.211x + 738.67 (R2 = 0.277)	y = −30.334x + 914.87 (R2 = 0.599)
	Tmin	y = −8.881x + 250.65 (R2 = 0.096)	y = −32.398x + 592.17 (R2 = 0.568)	y = −13.406x + 423.08 (R2 = 0.068)	y = −32.763x + 575.77 (R2 = 0.365)
	Pre	y = 2.021x + 140.17 (R2 = 0.231)	y = 0.701x + 158.12 (R2 = 0.153)	y = 0.529x + 177.24 (R2 = 0.056)	y = 0.727x + 103.17 (R2 = 0.318)
	Rh	y = 0.612x + 162.52 (R2 = 0.004)	y = 6.175x − 201.09 (R2 = 0.454)	y = 3.843x − 10.73 (R2 = 0.250)	y = 6.587x − 198.61 (R2 = 0.317)
TS	Tmax	y = −4.607x + 234.61 (R2 = 0.016)	y = −13.527x + 472.92 (R2 = 0.174)	y = 4.162x + 40.93 (R2 = 0.006)	y = −13.005x + 454.97 (R2 = 0.064)
	Tmin	y = −7.215x + 202.85 (R2 = 0.040)	y = −15.343x + 361.44 (R2 = 0.145)	y = 14.137x − 93.57 (R2 = 0.079)	y = −11.114x + 297.38 (R2 = 0.035)
	Pre	y = 1.176x + 130.31 (R2 = 0.054)	y = 0.742x + 110.63 (R2 = 0.208)	y = 1.025x + 123.81 (R2 = 0.126)	y = 0.959x + 36.86 (R2 = 0.430)
	Rh	y = −0.054x + 163.95 (R2 = 0.000)	y = 5.008x − 161.50 (R2 = 0.262)	y = 2.584x + 16.86 (R2 = 0.036)	y = 5.627x − 181.24 (R2 = 0.123)

). Analysis in the control areas showed that from 6.6% (in TS) to 78.9% (in SA) of the interannual variability in pea grain yield in the region can be explained by interannual fluctuations in the sum of temperatures above the critical threshold (Figure 7). A study by Lamichaney et al. [14] also showed that high temperatures during the growing season could reduce pea yields by up to 60%. Similar results were obtained by Zhelyazkova et al. [22] in South-Central Bulgaria, where temperature as a factor determined up to 71.5% of field pea yield. An equal contribution to the yield is made by the total number of days with such temperatures.

As noted earlier, considering current climate trends in the region, a further increase in temperature saturation and a decrease in moisture levels are most likely in the next decade [31,32]. Under such conditions, according to the linear regression equations, a further increase in accumulative temperatures above the critical threshold (TAC) by every 1 °C will contribute to a decrease in pea grain yield in the Rostov region by an average of 0.150 kg/ha. Moreover, each additional day with such temperatures will reduce the average yield for the region by 4.481 kg/ha. During the growing season, fluctuations in maximum temperatures explained the interannual variability in pea grain yield from 17.0% (in MI) to 71.4% (in SA) in May and from 0.6% (in TS) to 36.1% (in MI) in June, respectively (Table 1). At the same time, a further increase in the average maximum temperature by 1 °C in May will contribute to a decrease in pea yield from 11.857 kg/ha (in MI) to 22.877 kg/ha (in SA). In the reproductive phase of the development of peas (as well as other leguminous crops), the effect of high temperatures is negative [44,45]. High temperatures negatively affect pollen development, viability, and seed setting [46]. However, both male and female reproductive organs are adversely affected by high temperatures, and floral organs are also vulnerable to heat stress before pollination and in the post-pollination stage [47]. Adequate moisture levels are also required during this period, as pea sensitivity to water deficiency increases during flowering and seed filling due to high evapotranspiration [48,49].

It is also noteworthy that the climatic conditions in May and June in the Rostov region are also significant factors in determining the yield of other crops. Licker et al. [50] also showed that the temperature factor in May and June explained up to 49% and 16% of the year-to-year variability in winter wheat yields, respectively. For maize yields in the southwestern region, the temperature factor in June explained up to 58.7%, and the moisture factor in May in the eastern part up to 41.7% of the interannual variability [51].

The contribution of the sums of minimum temperatures below the threshold is insignificant and does not exceed 9.6% of the interannual yield variability (Figure 7). It should be noted that at rather high values of temperature indices in TS, a low degree of influence of the temperature factor on the yield of peas was observed. At the same time, an increase in temperatures in this part, according to our calculations, will have a minimal negative impact on productivity. This is probably due to the fact that in this part of the region, the value of relative air humidity during the growing season of peas is quite high (Figure 5), which somewhat levels out the negative impact of high temperatures.

The importance of the moisture factor for the productivity of peas in the region is also noticeable. For example, the level of moisture supply regime during the growing season, quantified by the HTC index in our study, explained from 5.8% (in MI) to 43.8% (in NE) of interannual pea yield variability (Figure 7). At the same time, the moisture supply regime of the growing season, expressed as total precipitation, explained from 12.3% (in MI) to 50.0% (in NE) of interannual yield variability (Table 1). Zhelyazkova et al. [22] also estimated the contribution of quantity to the yield of field peas in certain periods of vegetation. According to their findings, precipitation in June, as a factor, explains up to 81.9% of yields in South-Central Bulgaria. A 1 mm decrease in the total precipitation for the growing season will lead to a decrease in pea yields on average in the Rostov region by 0.736 kg/ha. At the same time, in the southwestern part of the region (in AK), where the amount of precipitation is maximum, the decrease in yield will be minimal (0.361 kg/ha). Notably, the contribution of relative air humidity during the growing season in the NE and SA regions, where the maximum yields were recorded, was 34.9% and 31.7%, respectively. During this period, an increase in relative air humidity by 1% contributes to pea yields of 7.410 and 6.587 kg/ha in NE and SA, respectively.

Declines in the productivity of legume crops, including field peas, in the context of climate change are also common in other regions. Model data showed that the yield of field peas in northern Italy will decrease by 12.6% by 2040 under current conditions [52]. A comparable decrease in pea yield is expected in the El Nubaria zone (Egypt) [53]. Generally, field pea production in the Mediterranean region decreased by 600 kg/ha as a result of every 1 °C increase in average temperature during flowering [54]. Moreover, in the United States in the last decade, soybean yields decreased by an average of 2.4% for every increase in growing season temperature by 1 °C [55]. Experiments in various agro-climatic regions of Australia have shown that field peas are most sensitive to climate change, and in the future, yield reductions can reach an average of 45% [56]. Taken together, climate change can significantly reduce gross crop yields in many regions of the world, which will negatively impact global food security [57,58].

5. Conclusions

As a result of this study, the influence of climatic factors on the pea grain yield in semi-arid conditions of the Rostov region was estimated. The productivity of pea in the region is limited not only by the lack of moisture during the growing season, but also to a greater extent by the influence of high temperatures, especially during the flowering period and grain filling. At the same time, low temperatures did not have any significant effect on yields. According to the regression equations obtained, climatic changes, all other things being equal, will lead to an even greater decrease in the yield of pea grain in the Rostov region. This will be especially acute in the eastern part of the region. Similar problems are likely to be common in other regions. To offset such losses, a number of adaptation measures are required. When growing field peas, producers need to focus on drought-resistant and heat-resistant varieties of field peas with a short growing season, and shift the sowing time of the crop to earlier dates, which will avoid the negative impact of high temperatures. To preserve moisture and reduce wind load, it is advisable to use methods of non-fall tillage. This study will be useful for developing adaptation measures, adjusting pea breeding programs, and manipulating agricultural practices. In addition, the results of this study can be applied to regional yield forecasting, as well as predicting the impact of climate variability on the grain yield of pea crops in arid areas.