Spatiotemporal patterns and vegetation forecasting of sunflower hybrids in soil and climatic conditions of the Ukrainian Steppe zone

Long-term


INTRODUCTION
Intensive agriculture has greatly increased crop yields but simplified the production due less diverse farming systems, higher genetic homogeneity, and more uniform farming landscapes in recent decades.It should be noted that the system of relations between natural and man-made causes has changed significantly, which leads to a larger number of abnormal climatic conditions and the risks of their negative effects on the conditions for growing agricultural crops.The main result of the global warming is a decreasing moisture supply.Moisture has become a limiting factor for soils in various climatic zones.Therefore, the preservation of soil moisture, increasing the plant stress resistance to moisture deficit, adaptation of cultivars and hybrids to new climatic conditions, the need to adjust crop rotations and agro-technological operations, and the use of effective multifunctional growth-regulating substances have become an urgent issue.
The Steppe as a physico-geographical area characterized with high temperatures, and scarce, unstable and uneven distribution of atmospheric precipitation (Dudiak et al., 2019;Zhang et al., 2019a), which causes high-risk farming conditions with a possibility of insufficient harvest of agricultural crops (Pichura et al., 2021;Mateo-Sanchis et al., 2021).Since the 2000s, the number of abnormal climatic conditions has tripled, which caused a rise in the average annual temperature by 2.6°C and a higher number of torrential precipitations in the spring-summer period (Pichura et al., 2022).This results in: a decreasing precipitation-productivity (Lisetskii et al., 2020;Oti et al., 2020), an increase in wind erosion (Dudiak et al., 2021), local torrential precipitations and water erosion processes (Özşahin, 2023), as well as higher risks of terrestrial runoff, disruption in transpiration and increased evaporation of moisture in the summer-autumn period, water deficiency for agrocenoses (Assan et al., 2020).Under such extreme climatic conditions, moisture is a limiting factor in the productivity of agricultural crops (Török et al., 2020), therefore the agrotechnological conservation of pre-sowing soil moisture and the effective use of its reserves during the vegetation period of plants are utterly important (Domaratskiy et al., 2020;2022).
One of the promising areas of agricultural science research is the use of the remote sensing data to study the state of the crops based on the normalized difference vegetation index (NDVI) (Ding et al., 2022;Essaadia et al., 2022).Spatiotemporal differentiation of the agrocenoses vegetation index indicates the state of plant development in different phenological stages depending on the natural and climatic conditions, the scope and kind of precipitation, the degree of agrotechnical measures.This makes it possible to define the level of adjustment of the hybrids to specific climatic conditions of certain physical and geographical zones and to evaluate the yield of agricultural crops.
Therefore, the purpose of the study was to forecast the yield to bolster the agricultural profitability, and to increase the level regional and national food security by rationalizing the optimal complex of agrotechnical measures, the choice of adjustable cultivars and hybrids considering the soil and climatic conditions of the Steppe zone.

LITERATURE REVIEW
Sunflower is the main Ukrainian oil crop, which takes about 25-28% in the structure of farm rotations.The seeds of modern zoned cultivars and hybrids contain 50-52% of oil, and the seeds of selective varieties -up to 60% (Shakalij et al., 2019).In comparison to other oil crops, the sunflower provides the highest oil yield per unit area.Sunflower oil makes up for 98% of total oil production in Ukraine.Scientists S.D. Koutroubas et al. (2020), A.U. Jan et al. (2022) explained that ensuring a high level of oil content in crops requires a high level of moisture supply in the cultivation territory.Y. Cheng et al. (2023) found that in arid regions, soil moisture and nitrogen concentration are key indicators of plant development and productivity.During the growing season, the level of water consumption for sunflower changes (Domaratskiy et al., 2020): from the emergence emergence to the inflorescence process, developed specifically for different moisture supply and plant nutrition conditions in order to forecast of the yield of sunflower hybrids according to a particular situation.The results can be used to improve the methodology of researching the vegetation of agricultural crops, to validate crop rotation, to choose the best practical ways for the use of multifunctional growth-regulating substances, to define the climatic adjustment of cultivars and hybrids, to manage resources, to develop adaptive climate technologies in agriculture and crop production, to calculate their efficiency, to forecast the yield and to ensure the profitability of agricultural production in the moisture deficit zone and managing a high-risk farming Keywords: crop production; climate; remote sensing; satellite images; modeling Ukrainian Black Sea Region Agrarian Science, 27 (3), [31][32][33][34][35][36][37][38][39][40][41][42][43][44][45] plants consume about 20% of growing water; during the phenological stages of fruit formation and flowering, the consumption of total vegetation moisture is 60%.In particular, the studies of V. Giannini et al. (2022) established that under the conditions of using standard agrotechnical measures on non-irrigated lands of the Steppe zone, the sunflower yield would not exceed 50% of its biological potential.Scientists I.V. Aksionov et al.Ronchetti et al. (2023) in particular prove that using of satellite images provides relevant and crucial information, which is used to develop the yield forecasting models regardless of the scope of the research area and crop rotation based on NDVI calculation.In modern science, the application of Sentinel satellites has provided a number of new possibilities for yield forecasting due to its spatial resolution and revisit time (Desloires et al., 2023;Zhang et al., 2023).
Based on the results of the analysis of recent studies, it was established that it is important to develop a model for predicting the yield of agrocenoses, including sunflower, based on the NDVI index by using the data on the distribution of water consumption and plant nutrients on various development stages.

MATERIALS AND METHODS
The study of the development and productivity of sunflower hybrids in the soil and climatic conditions of the Steppe zone of Ukraine was conducted in the period from 2019 to 2021 at the experimental field of the Mykolaiv State Agricultural Research Station of The Institute of Climate-Smart Agriculture of The National Academy of Agrarian Sciences of Ukraine (Fig. 1).Experiments were conducted without irrigation.The total area of experiments: 2019 -1.9 ha, 2020 -8.4 ha, 2021 -0.75 ha.The experimental areas are located on low-humus southern chernozems with a loam that has a high level of dust in its granulometric composition.The content of humus in soils varies from 2.7 to 3.1%, the depth of the humus horizon is 30-40 cm.The reaction of the soil solution is close to neutral (pH 6.5-6.8),hydrolytic acidity is within 2.00-2.52mg equiv.per 100 g of soil.The amount of absorbed bases is 32-35 mg equiv.per 100 g of soil, the degree of saturation with bases is 95.7%.According to the content of mobile nutrients, the soil of the experimental area is characterized by an average content of nitrate-nitrogen in the soil layer 0...20 cm -30.0 mg/kg, 100 mg/kg of mobile phosphorus and a very high content of exchangeable potassium -300.0 mg/kg of soil.The data on actual surface air temperature data (T, °С) and the amount of atmospheric precipitation (P, mm) for the growing season in 2019, 2020 and 2021 was used in the study, while the climatic norms for the research area were based on the data of the Mykolaiv meteorological station for the period of 1970-2020.
Program of scientific research.A field experiment was established in order to study the vegetation process based on the NDVI index and establish the climatic adjustment of high-oleic sunflower hybrids of Ukrainian and French selection.The Ukrainian selection included sunflower hybrids Hektor and Oplot (originator -V.Ya.Yuryev Institute of Crop Science).French selection consisted of DSL403 and P64GЕ133 (produced by Corteva, Brevant) and 8X477KL (produced by Dow Seeds).The experiments were repeated in 2019, 2020 and 2021.In 2019, the sowing date is 04/24, the harvest is 08/26; in 2020: sowing -29/04, harvesting -22/08; in 2021: sowing -10/05, harvesting -12/09.Every year, sunflower hybrids grew in the same sequence within typical soil and climatic conditions, in an area where the winter wheat used to grow before (Fig. 1).The sown area of the first order lot was 168 m 2 , the testing lot was 120 m 2 .Sowing was carried out with a UPS-8 pneumatic precision seeder, the seeding rate was 48.7 thousand units/ha.Accounting and observation were carried out according to the methodology of scientific research in agronomy (Ermantraut et al., 2008 (Papish, 2001).The seed yield was recorded manually, with the subsequent calculation of the yield in tons from 1 hectare of the sown area with 8% seed moisture and 100% seed purity.
Methods of deciphering space images and spatial analysis.Spatial-temporal differentiation of the vegetation of sunflower hybrids was determined on the basis of the calculation of NDVI (Beyer et al., 2023) based on the data of decoded Sentinel 2 space images with a spatial resolution of 10×10 m per pixel.
The NDVI value is calculated by the formula: where NIR -the visible and near-infrared range (Sentinel 2 -Band 8); Red -the red range of the electromagnetic spectrum (Sentinel 2 -Band 4).
Space photographs with no overcast over the experimental field were used in the research.The frequency of image processing was 10-16 days, which made it possible to determine the NDVI value for the main phenological stages of the development of sunflower hybrids, namely: seedlings (BBCH 00-09), the first pair of true leaves (BBCH 10-12), inflorescence formation (BBCH 14-59), flowering (BBCH 61-69), ripening (BBCH 71-99).The calculation of NDVI at all phenological stages made it possible to study the development process of sunflower hybrid crops and establish changes in the terms of the phenological stages of plants in regards to the climatic conditions of the year: dry, moderately humid and wet year.Space images were processed considering the sowing and harvesting dates in 2019, 2020, 2021.
In order to improve the quality of the visualization of maps of the spatiotemporal distribution of NDVI values and to increase both the accuracy of the interpretation of the vegetation index within individual plots and the characteristics of the heterogeneity of the vegetation of sunflower hybrids, the interpolation of the values obtained on the basis of the decoding of the Sentinel 2 space images was carried out.The interpolation was carried out using the method of geostatistical analysis with a radial basis function (Chen et al., 2023;Pichura et al., 2023).This method allows to establish an accurate interpolation of surface changes in NDVI values while preserving the input raster data.
Ukrainian Black Sea Region Agrarian Science, 27 (3), [31][32][33][34][35][36][37][38][39][40][41][42][43][44][45] The raster calculator of the ArcGis 10.6 software was used to calculate the generalized map of NDVI year values of sunflower hybrids based on the NDVI raster surfaces of the main phenological stages of vegetation and the formation of plant productivity.The correlation and regression analysis (Grover & Kaur, 2021) was used to develop functions for predicting the yield of sunflower hybrids depending on the spatial-temporal differentiation of the values of the vegetation index.
Space images processing, cartograms constructing, spatiotemporal and correlation and regression analysis were carried out with the licensed software product ArcGis 10.6 and Microsoft Excel 2010.The conducted research complies with all ethical norms according to The Convention on Biological Diversity (2022).

RESULTS AND DISCUSSION
Analysis of climatic research conditions.Climatic conditions of the studied area are characterized as medium-arid.The average statistical value of the normal (in period of 1970-2020) air temperature during the growing season was 18.0°C (Fig. 2a), the standard deviation was 4.9°C, and the level of variation was 27.3%.
In particular, the average value of the air temperature during the growing season in the dry year (2020) was 19.0°C, the standard deviation was 6.3°C, and the level of variation was 33.3%.In a moderately humid year (2019), the average air temperature during the growing season was 20.4°C, the standard deviation was 5.5°C, and the level of variation was 27.0%.In the wet year (2021), the average air temperature during the growing season was 18.4°C, the standard deviation was 6.6°C, and the level of variation was 35.8%.
It has been established that over the past 10-15 years, the frequency of abnormal climatic conditions of a torrential nature has increased (Pichura et al., 2022).In July 2020 (Fig. 2b), a drastic rise in the vegetation index was observed during the flowering period (BBCH 61-69), but the torrential nature of atmospheric precipitation had no positive outcome or prolonged effect on the formation of the productivity of sunflower hybrids.In particular, in the dry year (2020), a high value of the standard error (84.1 mm) and the level of variation of seasonal changes in atmospheric precipitation (139.7%) were recorded, which confirms their anomaly during the growing season.

Figure 2. Climatic conditions during the sunflower growing season for 2019-2021: a -average monthly air temperature (°С); b -amount of precipitation (mm) Source: compiled by authors, based on the data of the Mykolaiv meteorological station
The average monthly value of atmospheric precipitation during the growing season for 50 years (1970-2020) had been 43.6 mm (Fig. 2b), with the standard deviation 12.6 mm, and the level of variation 28.9%.The year 2019 was close to typical (normal) climatic conditions, the average monthly precipitation during the growing season was 47.0 mm, the standard deviation was 13.3 mm, and the level of variation was 28.3%.A typical wet year of 2021 in the Steppe zone is characterized by an average monthly precipitation of 72.8 mm, a standard deviation of 32.4 mm, and a high level of variation of 44.5%.
Well-developed sunflower crops consume from 500 mm to 600 mm of water during the growing season (Kohan, 2021) (the minimum water requirement is met with 300-400 mm of atmospheric precipitation).During the growing season of 2019, the amount of atmospheric precipitation was 235 mm, in 2020 -295 mm (in July, an unusually abnormal amount of precipitation was estimated to nake 70.5% of the share of the growing season, being an unproductive precipitation of a torrential nature), in 2021 -364 mm.
In the arid conditions of the Ukrainian Steppe, the level of soil moisture is a factor that determines the a b productivity of agrocenoses.It was estimated that 60-70% of the total sunflower water consumption during the growing season are provided by atmospheric precipitation, while 30-40% -by the soil moisture supply (Domaratskiy et al., 2020).Using the thermostatic-weight method, it was established that the pre-sowing moisture reserves in a meter-long soil layer of the experimental fields in the dry year of 2020 amounted to 41 mm, in the moderately humid year of 2019 -69 mm, and in the wet year of 2021 -89 mm.
The second half of the growing season in 2019 was characterized by a moisture deficit, being 23.9% less than normal atmospheric humidity, as well as an increase in temperature by 11.5%, which mainly caused stress in plants with a subsequent decrease in productivity.In a dry year of 2020 sunflower hybrids were constantly under stress, caused by a significant deficit of soil and atmospheric moisture in combination with anomalous unproductive rainfall of a torrential nature in July as a result of high air temperature.Peculiarly, in the wet year of 2021, there were no signs of stressful weather conditions and a productive precipitation was recorded during the budding and flowering period (BBCH 51-69), which is the defining phenological stage of plant development in crop formation.Therefore, in the conditions of extreme agriculture of the Steppe zone, agrotechnological measures are aimed at preserving moisture.
Studying the state of sunflower hybrid crops.A change in the activity of the plant's photosynthetic processes and the production of chlorophyll content at a certain macrostage and phenological phase are indicators of plant development.The study of changes in the photosynthetic activity of sunflower hybrids was carried out based on the analysis of NDVI values, which is common in forecasting the productivity of agrocenoses.
The value of the NDVI index within the experimental fields, during 2019, 2020 and 2021, was calculated on the basis of data from satellite images from the Sentinel 2 spacecraft.Using the images, the state of the green mass of plants, which absorbs electromagnetic waves in the visible red range and reflects them in the near infrared was determined.In particular, the maximum absorption of solar radiation by chlorophyll occurs in the red zone of the spectrum (central wavelength of Sentinel 2 -665 nm), and in the near infrared zone (central wavelength of Sentinel 2 -842 nm) -the maximum reflection of energy by the cellular structure of the leaf.
The moderately humid year (2019).By interpreting a series of satellite images in a moderately humid year at the beginning of the vegetation process (May 5, 12 days from the sowing date), young seedlings were recorded in sunflower hybrid crops with an average value of the NDVI index of 0.26 ± 0.03 and insignificant level of spatial variation -8.1%.After foliar feeding of the hybrids, a heterogeneous reaction to multifunctional growth-regulating substances was observed, which was later confirmed by satellite images on May 30 (37 days from the sowing date).It should be noted that the growth of Oplot hybrid plants increased and the NDVI value varied within 0.54-0.77, the Hector hybrid NDVI value varied within 0.54-0.80 and the DSL403 hybridwithin 0.51-0.78.A downward reaction to re-regulating substances was observed in hybrid P64GE133 with an NDVI value of 0.41-0.67 and hybrid 8X477KL with an NDVI value of 0.43-0.62.At the end of the phenological phase of "inflorescence formation" on June 14 (52 days from the date of sowing) and the beginning of the "flowering" phase on June 19 (57 days from the date of sowing), good (0.55-0.70) and thriving state of vegetation was recorded of all sunflower hybrids (> 0.70).During this period, the average value of NDVI was 0.72 ± 0.06, the level of spatial heterogeneity was 8.2%.
The second half of the growing season of sunflower hybrids in 2019 includes the second half of the flowering phase (BBCH 67-69) and the macrostages of "fruit formation" (BBCH 71-79), "fruit and seed ripening" (BBCH 80-89) and "senescence" (BBCH 92-99), which are components of the phenological stage of ripening (BBCH 71-99).It should be noted that the second period of plant vegetation was characterized by stressful conditions caused by the lack of moisture and high temperatures.This led to the sharp deterioration of photosynthetic processes and the shortening of the "fruit formation" period.On July 4 (72 days from the date of sowing), the average value of NDVI was recorded -0.63 ± 0.09 with noticeable manifestations of heterogeneity in the formation of productivity of sunflower hybrids, the level of spatial variation -14.1%.During the vegetation period on the "fruit and seed ripening" macrostage, on July 24 and August 13, rapid seed maturation of P64GE133 and 8X477KL hybrids was recorded.In the macro stage of "senescence" from the period of full ripeness (seed moisture about 10%, BBCH 92) to harvest (August 23), the average NDVI value was 0.37, on August 26 it was 0.30.
The dry year (2020).In 2020, the recorded conditions were extremely dry for the cultivation of sunflower hybrids, which caused a shorter growing season and the terms of individual phenological phases of plants.In particular, the beginning of the growing season in 2020 was characterized by a low level of soil moisture and a small amount of precipitation.This resulted in weaker plant seedlings and a critically low level of photosynthetic processes at the beginning of the phenological phase of inflorescence formation.After the cultivation of crops when the 6-8 true leaves were formed, the delayed reactions of all hybrids to the action of multifunctional re-regulating substances were recorded, as a result of stressful climatic conditions.According to the satellite image decoding data for May 19 (21 days after sowing), a low level of NDVI value was calculated -0.23±0.02with an insignificant level of variation -8.2%.The lack of precipitation caused further suppression of plant development, which was confirmed by the re-sults of interpreting the satellite image from June 8 (41 days after sowing), the NDVI value was 0.36±0.04 with a significant level of variation -10.3%.June 2020 was characterized by heavy rainfall, which improved photosynthetic processes in sunflower hybrids.At the beginning of the flowering phase, on June 23 (56 days after sowing), the NDVI value was 0.70±0.03with a slight variation of 4.9% (Fig. 3; Fig. 4).

Figure 4. Seasonal distribution of NDVI of sunflower hybrids on the experimental field (2020) Source: compiled by authors
The end of the flowering phase in July 3 (56 days after sowing), was also characterized by high NDVI values of 0.69 ± 0.03 with a variation level of 9.8%.The lack of atmospheric and soil moisture supply in the second half of the plant growing season caused a sharp decrease in the photosynthetic activity of sunflower hybrids and, with a corresponding reduction in the term of the "fruit formation" (BBCH 71-79), stimulated the acceleration of "fruit and seed ripening" (BBCH 80-89) and "senescence" (BBCH 92-99) of plants, on July 23 (86 days after sowing), the NDVI value was 0.41 ± 0.04 with a variation level of 9.8%.As of August 7 (101 days after sowing), the NDVI value was 0.30 ± 0.04 with a high level of variation -12.2%.
On August 17-18 (112 days after sowing), during the "senescence" macrostage, the NDVI value was 0.25 ± 0.03 with a high level of spatial variation -11.6%.High spatial variation was caused by significant spatial heterogeneity of plants as a result of stress caused by climatic conditions.It was found that Hektor and DSL403 sunflower hybrids matured faster in dry periods.The lack of moisture caused the deterioration of photosynthetic processes, a significant decrease in the content of chlorophyll in plants, a shortening of the terms of important phenological phases and the vegetation period of sunflower hybrids as a whole.
The wet year (2021).The beginning of the growing season in 2021 was characterized by favorable climatic conditions in the pre-sowing period, which ensured a high level of soil moisture during the sowing period.This determined the high energy and uniformity of seedlings, recorded on May 14 (5 days from the sowing date), the value of the NDVI index -0.25 ± 0.03, the level of spatial variation -6.2% (Fig. 5).

Figure 5. Seasonal distribution of NDVI of sunflower hybrids on the experimental field (2021) Source: compiled by authors
On June 8 (21 days from the date of sowing), after processing the crops, a high heterogeneity of the response of the hybrids to multifunctional re-regulating substances was observed, this is due to the redistribution of moisture in the field and the hybrids' adjustment to the climatic conditions of the Steppe.The value of NDVI was 0.42 ± 0.04 with a high level of spatial variation of 14.0%.During this period, the high photosynthetic capacity of the Oplot hybrid was recorded, the NDVI value reached the level of 0.56.Hybrids DSL403 (NDVI -0.39) and P64GE133 (NDVI -0.40) had a relatively low level of photosynthesis.
On June 23 (45 days after sowing), during the flowering phenological stage, all the sunflower hybrids were characterized by a high level of photosynthetic processes, the NDVI value was 0.75 ± 0.06 with a spatial variation of 8.5%.Systematic and productive atmospheric precipitation and high soil moisture supply in the first part of the vegetation period led to the prolongation of the phenological stage of flowering, which contributed to the increase in plant productivity.In 2021 the flowering phase had lasted for 33 days, which is 2.3 times longer than in the previous two years.The maximum value of NDVI during the flowering period is 0.89-0.93.A high level of NDVI was recorded in the phenological stage of the "fruit formation" macrostage -0.74 (76 days from the date of sowing) and the macrostage of "fruit and seed ripening" -0.54 (95 days from the date of sowing).
A high level of moisture supply, the extension of the duration of the flowering phase provided favorable conditions for the formation of fruits and ripening of sunflower seeds.On September 6 (119 days after sowing), at the end of the "ripening fruit and seed" macro-stage, the NDVI value was 0.39; on September 12, in the "senescence" and harvest macro-phase, the NDVI value was 0.32.Spatial variation during the growing season of plants was caused by spatial differentiation of soil moisture and different levels of plasticity of hybrids to the climatic conditions of the Steppe zone.
Forecasting the yield of sunflower hybrids.Sunflower has a deep root system and the ability to consume water from a depth of 2 to 4 meters in dry years, however it still needs moisture supply.In different periods of the phenological stages of development, the sunflower consumes moisture unevenly, most actively during the phase of intensive growth, flowering and seeding.( Spatial differentiation of NDVI year values, which are functionally correlated with spatial variation of yield of sunflower hybrids, was calculated using a raster calculator (Fig. 6).It was determined that in the moderately humid year of 2019, the value of NDVI year varied within 0.57-0.69, in the dry year of 2020 -0.48-0.59, in the wet year of 2021 -0.49-0.71.It was established that the values of the vegetation index reflect photosynthetic processes and the production of chlorophyll content, which depend on the conditions of moisture supply of the year and the amount of mineral nutrients, but do not identify the overall features and hybrids' adjustment to the soil and climatic conditions of physiographic zones.Therefore, in order to clarify the correlation between NDVI year values to the productivity of individual hybrids, the generalization of NDVI year values was carried out by mathematically relating the raster values of the vegetation index to its average value of the corresponding year.With the calculations, raster surfaces were created, where the value "1" corresponds to the average value of NDVI year for each particular year.Thus, the average value of NDVI year of a separate lot corresponds to the average yield value of the respective sunflower hybrid.Next, the average yield of each sunflower hybrid was calculated: Oplot in 2019 -3.0 t/ha, in 2020 -2.01 t/ha, in 2021 -3.04 t/ ha; Hector in 2019 -2.05 t/ha, in 2020 -1.65 t/ha, in 2021 -2.16 t/ha; DSL 403 in 2019 -2.53 t/ha, in 2020 -1.88 t/ha, in 2021 -2.77 t/ha; P64GE133 in 2019 -2.83 t/ha, in 2020 -1.96 t/ha, in 2021 -3.02 t/ha; 8X477KL in 2019 -2.32 t/ha, in 2020 -1.71 t/ha, in 2021 -2.82 t/ha.
Based on the aforementioned data, cartograms of the spatial differentiation of yield of sunflower hybrids were constructed (Fig. 7), correlating to the distribution of generalized NDVI year values according to the formula: where NDVI iyear -the value of the vegetation index within the experimental lot of a separate cultivar or hybrid of a culture; Aver (NDVI iyear ) -the average value of the vegetation index within the experimental plot of a separate variety or hybrid of a culture; Aver(СY i ) -the average value of the productivity of a separate variety or hybrid of a culture within the experimental area.

Figure 7. Cartograms of the distribution of the yield of sunflower hybrids in 2019-2021 Source: compiled by authors
Yield cartograms make it possible to establish the spatiotemporal heterogeneity of the productivity for an individual sunflower hybrid depending on the climatic conditions in 2019, 2020 and 2021.In the average humid year (2019), the yield of sunflower hybrids ranged from 1.86 to 3.18 t/ha.The minimum yield values were recorded for Hector hybrids were 1.86-2.15t/ha, for 8X477KL -2.10-2.42t/ha, average yield levels for DSL 403 -2.44-2.60 t/ha, while other have reached their maximum yield: Oplot -2.70-3.18t/ha and P64GE133 -2.65-2.92t/ha.In the dry year (2020), hybrids Hektor -1.53-1.76t/ha and 8X477KL -1.58-1.81t/ha had the minimum yield values, hybrids DSL 403 -1.75-2.00t/ha, P64GE133 -1.80-2.06t/ha and Oplot -1.85-2.16t/ha.In the wet year (2021), the Hektor hybrid had the minimum yield value -1.82-2.56t/ha, the average yield level was recorded in the hybrids DSL 403 -2.65-2.93t/ha and 8X477KL -2.50-2.98 t/ha, the maximum level was achieved by hybrids Oplot -2.60-3.37 t/ha and P64GE133 -2.80-3.17t/ha.It was established that each sunflower hybrid has individual genetic features of plant adjustment to the soil and climatic conditions of the Steppe, which determines the yields.Therefore, a model of the yield prediction function was developed for each hybrid (Table 1).Functions with a high level of approximation (r 2 ) describe the spatiotemporal processes of vegetation formation of yield for each individual sunflower hybrids according to the level of moisture supply during the year.Function models were created on the basis of three raster periods of water supply and mineral nutrition of plants.The NDVI 1 value contains the average value or the raster surface of the spatial distribution of the vegetation index of sunflower hybrids in the macrophase of the formation of the embryo inflorescences of the BBCH 16-19 plant, the NDVI 2 value contains the average value or the raster surface of the spatial distribution of the vegetation index of sunflower hybrids in the flowering macrophase of BBCH 61-67, the NDVI 3 value contains the average value or raster surface of the spatial distribution of the vegetation index of sunflower hybrids at the end of the macrophase of fruit formation and the beginning of seed ripening (BBCH 79-80).These functions provide a level of approximation of actual data with 97.2-99.9%level of data approximation, which confirms the high accuracy of predicting the yield of sunflower hybrids.
It should be mentioned that the scientific research in the field of agriculture and crop production S.M.   2020) are approximated and do not take into account the causality of the spatiotemporal differences in the vegetation of varieties and hybrids of crops in regards to the unified BBCH scale.Therefore, the researchsuggests to take into account the peculiarities of the vegetative development of different sunflower hydrides according to the unified BBCH scale.This made it possible to establish regularities in the formation of the productivity of agrocenoses depending on their development in the main phenological phases and to determine the genetic features of the climatic plasticity of each sunflower hybrid to the conditions of their cultivation in the Steppe zone of Ukraine.
In the studies by I.V. Aksionov et al. (2021) and P. Kamath et al. (2021) forecasting the yield of agricultural crops was carried out on the basis of correlation and regression analysis, which involves determining the influence of individual natural and agrotechnical factors and further developing regression models of correlation between the actual yield and production factors.Such models are based on field, experimental and official data of institutions monitoring the natural and climatic conditions of the research region and do not take into account the spatiotemporal patterns of plant development in separate phenological phases of their development.Therefore, the author's approach to the development of regression models for predicting sunflower yield based on the new NDVI according to the unified BBCH scale is proposed.This would allow to quickly react to the changes on the plant vegetation state and to make adjustments to agrotechnical processes in order to increase the productivity of agrocenoses.
In the scientific works of G. Ronchetti et al. (2023) and Desloires et al. (2023) prove the relevance of using satellite images for forecasting and improving the mechanism of managing the agrocenoses development within a separate field.Thus, taking into account the existing developments, the approaches of using a series of satellite images for detailed research and establishment of spatiotemporal regularities in vegetation and the formation of agrocenoses productivity, considering their genetic features, zonal phytopotential, soil, local nature and climate conditions and elements of varietal agricultural technology, has developed and improved significantly.
The previous scientific works and the original research confirm the scientific and practical value of using modern technologies of remote sensing to improve the monitoring of crops, manage and determine the efficiency of agrotechnical measures, increase the reliability of forecasting the agrocenoses productivity within a specific field or the area of individual agricultural producers.The presented results are particularly aimed at determining the sunflower cultivars and hybrids, which would be the most adjustable to the soil and climatic conditions of the Steppe zone, and defining the specific features and effective ways to use chemical and biological multifunctional re-regulating substances to increase the plants' stress resistance to climate changes and moisture deficit.This enables the land users to adjust the system of growing agricultural crops in the zone of moisture deficit and to manage high-risk farming in order to ensure the profitability of production.

CONCLUSIONS
By interpreting the Sentinel 2 satellite images and analyzing of seasonal changes in NDVI values, the spatiotemporal patterns of the sunflower hybrids vegetation and their adjustment to the soil and climatic conditions of Ukrainian Steppe zone were established.A series of maps of with the differentiation of NDVI values at every phenological stage was developed, which made it possible to study the development process of sunflower hybrid crops and to establish changes in the length of the phenological stages of plants according to the climatic conditions of the respective year: the dry, the average humid and the wet year.In the average humid (2019), favorable conditions for plant development were recorded in the first half of the growing season, while the second half of the growing season was characterized by a decrease in soil moisture and a shortening in the flowering stage of sunflower hybrids.It was proved that in the dry year (2020) there was a reduction in the length of the flowering stage of sunflower hybrids, a low level of the NDVI vegetation index was recorded in the phase of inflorescence formation (0.22-0.40) and the ripping stage (0.30-0.40).In the wet year (2021), a prolongation of the flowering stage and a high value of the vegetation index during all the phenological stages of plant development were recorded.It was established that each sunflower hybrid has individual genetic features of plant adjustment to the Steppe soil and climatic conditions, which determines the yield of the crop.A high level of adjustment was observed in: sunflower hybrid Oplot with an average annual yield of 2.01 to 3.04 t/ha and hybrid P64GE133 with a yield in the range of 1.96-3.02t/ha.The hybrid DSL 403 had an average level of plasticity, its yield was 1.88-2.77t/ha.A low level of plasticity was recorded in the 8X477KL hybrid with a yield of 1.71-2.82t/ha and the Hector hybrid with a yield of 1.65-2.16t/ha.The water consumption of sunflower hybrids for the formation of a yield unit (t/ha) was calculated for each year of the study: the dry year -927 ± 80 m 3 /ha, the average humid year -1106 ± 163 m 3 /ha, the wet year -1540 ± 232 m 3 /ha.A yield prediction function model was developed for each hybrid.Functions with a high level of approximation describe the spatiotemporal processes of yield vegetative formation for individual sunflower hybrids according to the level of moisture supply of the year.The function models were based on the raster models of the spatial distribution of NDVI values in the most active periods of intensive growth, flowering and seeding.For situational forecasting of the yield of sunflower hybrids, it is recommended to use a system of function models developed for different conditions of the crop's moisture supply and mineral nutrition.The proposed functions provide the 97.2-99.9%level of data approximation, which confirms the high accuracy of predicting the yield of sunflower hybrids.The research results are a relative for improving the methodology of vegetation research of agricultural crops, validating crop rotation, determining the effectiveness of agrotechnical measures, correct selection of varietal and hybrid composition, forecasting the yield of sunflower hybrids in the soil and climatic conditions of the Steppe zone.

Figure 1 .
Figure 1.The location of the experimental fields and the arrangement of sunflower hybrid sowing in the period of 2019-2021 Notes: 1 -Oplot; 2 -Hector; 3 -DSL403; 4 -P64GE133; 5 -8Х477KL Source: compiled by authors From the seedlings emergence (BBCH 09) to the inflorescence formation (BBCH 15), plants consume about 20% of vegetation moisture; during the phenological phases of inflorescence formation (BBCH 16-59) and flowering (BBCH 61-69), the consumption of total vegetation moisture is 60%, this period is crucial in the formation of productivity; on the other hand, 20% of vegetation moisture is consumed during the fruit formation and at the beginning of ripening (BBCH 71-80).Defining the NDVI index during the main stages of sunflowers' moisture supply and mineral nutrition provided the possibility of creating a general cartogram of the spatial distribution of the vegetation index values in 2019, 2020 and 2021.Generalized cartograms of NDVI year values are the sum of raster surfaces of differentiation in NDVI values in the main periods of plant vegetation.The first raster surface in particular identifies the NDVI values at the beginning of the formation of embryo baskets in the macrophase BBCH 16-19, the second raster surface contains the NDVI values of the characteristics of the macrophases BBCH 61-67, the third raster surface contains the distribution of NDVI values of the macrophase BBCH 79-80.The corresponding NDVI raster is assigned a weighting factor of the cumulative influence of the values of vegetation moisture consumption and mineral nutrition elements on the formation of crop yield.Thus, the values of the raster surface NDVI BBCH 16-19 are assigned a weighting factor of 0.2, the values of NDVI BBCH 61-67 are assigned a weighting factor of 0.6 and the values of the raster surface NDVI BBCH 79-80 are assigned a factor of 0.2.The generalized map of NDVI values is created by the formula: NDVI year = 0.2NDVI BBCH 16-19 + + 0.6NDVI BBCH 61-67 + 0.2NDVI BBCH 79-80 .

Figure 6 .
Figure 6.Spatial differentiation of generalized NDVI year values of sunflower hybrids in 2019-2021 Source: compiled by authors