Use of Indices Applied to Remote Sensing for Establishing Winter–Spring Cropping Areas in the Republic of Kazakhstan

Abstract

Farmers in Kazakhstan face unreliable water resources. This includes water scarcity in the summer, high fluctuations in precipitation levels, and an increase in extreme weather events such as snow, rain, floods, and droughts. Wheat production is regulated and subsidized by the Kazakh government to strengthen food security. The proper monitoring of crop production is vital to government agencies, as well as insurance and banking structures. These organizations offer subsidies through different levels support. Some farmers already use farmland soil monitoring combined with adaptive combinations of different crops. These include winter–spring plowing crop programs. Winter wheat crops are generally more adaptive and may survive summer droughts. Kazakhstan is a large country with large plots of farmland, which are complicated to monitor. Therefore, it would be reasonable to adapt more efficient technologies and methodologies, such as remote sensing. This research work presents a method for identifying winter wheat crops in the foothills of South Kazakhstan by employing multi-temporal Sentinel-2 data. Here, the researchers adapted and applied a Plowed Land Index, derived from the Brightness Index. The methodology encompasses satellite data processing, the computation of Plowed Land Index values for the swift recognition of plowed fields and the demarcation of winter wheat crop sowing regions, along with a comparative analysis of the acquired data with ground surveys.

1. Introduction

In modern agriculture, remote sensing technologies, including satellite imagery, are becoming increasingly useful for monitoring and managing agricultural crop production. These technologies are especially valuable for large agricultural lands, aiding farmers and researchers in optimizing yields and ensuring sustainable farming practices. The application of remote sensing technologies in various countries has demonstrated significant advantages. For example, in the United States, satellite images are used to monitor corn and soybean production, providing crucial data on crop conditions and facilitating the timely application of fertilizers and pesticides [1,2]. In Europe, remote sensing supports precision farming by offering high-resolution data that help in efficient land management involving the production of diverse crop types [3,4]. Indian researchers have used satellite data for crop area mapping and production estimation, supporting food security and agricultural planning [5,6,7]. China is advanced in satellite data applications for crop monitoring and yield forecasts [8,9,10]. In Brazil and Argentina, remote sensing is utilized to assess pasture conditions and manage rangelands [11,12]. Russia has intensively applied satellite data for analyzing land-use changes and monitoring forest resources [13,14]. In South Africa, remote sensing is used for vegetation monitoring and drought risk assessment [15,16].

Despite the proven benefits of remote sensing in agriculture, its application in Kazakhstan, particularly for monitoring winter–spring plowing crops, remains underexplored. Kazakhstan, with its vast agricultural landscapes, can benefit from integrating satellite technologies into agricultural practices. The Sentinel-2 satellite, with its high resolution and frequent revisit times [17,18,19] represents a valuable resource for agrarian land monitoring and land use change detection. This study focuses on the application of Sentinel-2 imagery for identifying arable lands and monitoring the timing of the sowing and growth of winter–spring plowing crops in the Turkestan region of Kazakhstan. Appropriate cooperation and remote sensing application technologies for farmland are required. Wheat production is one of the main forms of agricultural crop production in Kazakhstan, and it is regulated and subsidized by the Kazakh government. Monoculture crops often become infected with different types of virus. Moreover, Kazakhstan has very unstable water availability: in some periods, large floods cover planning areas, while in other periods, substantial droughts may occur. Appropriate strategies are necessary to adapt to mitigate these issues. Crop monitoring on large plantation sites would help to resolve these issues.

Winter crops constitute the main activity of Kazakh farmers in the foothill zone. The area sown with winter grain crops varies significantly from year to year, depending on the crop rotations carried out on farms and weather conditions. The development of a method for automatically recognizing the sowing mask of winter grain crops based on remote sensing is the basis for solving various problems in space monitoring and directly affects the condition and yield of winter grain crops and the timing and volume of various types of field work on farms.

Using Sentinel-2 images from September to November 2021 and March to May 2022, along with data from the autumn and spring periods of 2017–2018, the Tasseled Cap Brightness Index [20,21] was processed with the Plowed Land Index, specifically designed to identify arable lands in the foothill zone of South Kazakhstan.

Processed satellite data were used for the winter–spring plowing crop classifications, along with field surveys for the calibration and verification of the processed works. This work mainly emphasizes winter wheat tracking, based on the requests of farmers from the Turkestan region.

Remote sensing technology applications allow the monitoring of large farming areas of soil and crop conditions, in connection with subsidy-support systems provided by state agencies. Through the timely implementation of agricultural measures, such as fertilization and plant protection, the control of land productivity supports the sustainable development of crop yields and soil conditions. Digitized data on the conditions of farm fields allows the more effective planning and management of resources, including water, fertilizers and pesticides. The optimization of winter grain production using remote sensing makes it possible to obtain monitoring data more efficiently. The development of modern agricultural technologies and the need for their maintenance create advancements in remote sensing technology applications.

The purpose of this research is to develop and use remote-sensing-based tools, which makes it possible to efficiently monitor the sowing areas of different crop zones. These efforts will be used to rapidly investigate agricultural production and solve subsequent problems in crop production. The practical applications of these efforts are for the benefit of agricultural producers, as well as land administrations, the Kazakh government, insurance organizations, and banking institutions providing subsidies. The results obtained contribute to optimizing the production of the winter grain crops, increasing the productivity of agricultural land, improving the sustainable development of rural areas and ensuring food security.

2. Materials and Methods

2.1. Study Area

The territory of the Turkestan region boasts diverse landscapes, soils and agro-climatic resources. Significant areas (up to half of the region’s territory) are occupied by plains and sands in the north and west of the region. In the middle of the region, from northwest to southeast, the Karatau Range runs, with most peaks around 1800 m, separating the plains territory from the mountainous regions. Low-mountain and high-mountain ranges of the Western Tien Shan (with elevations ranging from 1500–2000 to 2500–3500 m and more) are in the southeast of the region. The diversity of the region’s relief has resulted in significant heterogeneity of soil-climatic conditions and agricultural specialization in different areas [22]. The cultivation of winter grain crops is concentrated in the foothills of the Turkestan region.

For this study, the Tyulkubas district, located in the foothill zone in the eastern part of the Turkestan region, bordering the Zhambyl region, was selected (Figure 1).

The agricultural lands of the Tyulkubas district cover an area of 148.3 thousand hectares. This includes arable land, perennial plantations, hayfields and pastures. According to statistical data for 2022, arable lands account for 64.5 thousand hectares, of which fodder crops (predominantly alfalfa) occupy 27.6 thousand hectares (43%), winter wheat crops 16.9 thousand hectares (26%), oilseed crops 12.1 thousand hectares (19%), spring wheat crops 5.9 thousand hectares (9%) and vegetables and melons 1.6 thousand hectares (3%). Thus, winter wheat and fodder crops dominate the cropping structure of the district. The cropping areas of agricultural crops vary from year to year, depending on weather conditions and crop rotation.

The Tyulkubas district possesses unique geographical and soil–climatic characteristics that significantly influence agricultural activities and field productivity. Depending on the vertical zonation and decreasing altitude of mountain ranges, a change in soil type occurs from mountain meadow and subalpine soils (the most fertile, with a high humus content of up to 10%) to mountain brown soils (moderately fertile) and chestnut soils (less fertile, with low humus contents of 3–4%).

Land use in the Tyulkubas district is in the dry foothill zone and in the mountainous agroclimatic zones.

The dry foothill zone is characterized by precipitation amounts of 100–250 mm, and the highest moisture during the vegetation period (HTC ranging from 0.3 to 0.5). The vegetation period is characterized by a prolonged period of warm days (up to 160–180 days per year). This zone provides the most favorable agro-climatic conditions for cultivating various crops (grains, oilseeds, fodder, vegetables, fruit and berry crops), primarily under rainfed conditions, sometimes under irrigation.

The mountainous agroclimatic zone has fewer thermal resources, a shorter growing season and frost-free period (160 days or less) and increased moisture. The territory is primarily used as pasture and hayfields, and cultivation mainly involves cold-resistant crops (grains, fodder, etc.) under rainfed conditions. Due to limited thermal resources, the region is not suitable for the development of irrigated agriculture.

Precipitation during the autumn period is crucial for seedling emergence and wintering of winter crops. In the foothill and mountainous areas, precipitation during the autumn period can reach 100–200 mm, significantly more than on the plains. By the end of the autumn period, winter wheat crops usually reach the third leaf stage or the beginning of tillering, which, with sufficient snow cover, contributes to good crop wintering.

During the winter period, precipitation in the form of snow falls in the foothill and mountainous areas, reaching thicknesses of 20–30 to 35–40 cm, effectively protecting crops from freezing. Vegetative growth of winter wheat crops typically resumes in the second half of March, with active growth and development observed throughout April and May. Heading and flowering usually occur in the third week of May, followed by maturation in June. Harvesting of winter wheat crops, depending on the ripening dates, usually begins in mid-July and continues until the end of August, sometimes extending into early September. The agro-climatic characteristics of winter wheat cultivation and the timing of field operations were used as a basis for satellite monitoring of cropping areas and the development of a methodology for determining the timing and areas of winter wheat crop sowing in the mountainous region [23].

2.2. Satellite and Ground Data

To address the research objectives, multispectral data from the Sentinel-2 satellite system were utilized, obtained using the Level-1C (L1C) optical multispectral instrument (MSI). These data were downloaded from ESA’s Copernicus Open Access Hub, which provides access to Sentinel-2 satellite data.

The Sentinel-2 satellite mission, comprising two satellites, Sentinel-2A and Sentinel-2B, began on 23 June 2015, with the launch of Sentinel-2A. This mission is characterized by a wide swath, and high resolution, and it is based on optical multispectral visualization. Sentinel-2A and Sentinel-2B orbit the Earth in the same orbit, with a phase difference of 180 degrees, providing high image repeatability, with a 5-day revisit period.

The primary objective of the Sentinel-2 mission is to provide multispectral images for various applications in different domains, such as land use monitoring (e.g., agriculture, vegetation, forests, water bodies), emergency management (e.g., monitoring and managing natural disasters), climate research (e.g., glacier and coastal monitoring) and security provision (e.g., border and maritime surveillance).

The Multispectral Instrument (MSI) on Sentinel-2A satellites comprises 13 spectral bands (VIS, NIR, SWIR) with varying spatial resolutions of 10 m, 20 m and 60 m for different spectral bands. The satellites’ orbit is sun-synchronous, at an average altitude of 786 km, allowing for a clear 5-day revisiting cycle between the two satellites [24,25,26].

For this study, Sentinel-2 images were processed using ArcGIS software, version 8 (provided by ESRI). The list of Sentinel-2 data used, for example, for the years 2021–2022, is provided in

Table 1. Sentinel-2 satellite data used for Tyulkubas District, Turkestan region, for 2021–2022.

Date	The Sentinel-2 Satellite Data Used in 2021–2022
	WN Scene	XN Scene	XM Scene
13.9–18.9	T42TWN_20210913	T42TXN_20210913	T42TXM_20210918
26.9–28.9.	T42TWN_20210926	T42TXN_20210928
11–13.10	T42TWN_20211011	T42TXN_20211013
16–18.10	T42TWN_20211016	T42TXN_20211018	T42TXM_20211018
31.10	T42TWN_20211031
02.11	T42TWN_20211102	T42TXN_20211102
05.11–07.11	T42TWN_20211105	T42TXN_20211107	T42TXM_20211107
27.11	T42TWN_20211127	T42TXN_20211217
04.04–06.04	T42TWN_20220404	T42TXN_20220406
09.04	T42TWN_20220409
14.04	T42TWN_20220414
21.04	T42TWN_20220421
26.04	T42TWN_20220426	T42TXN_20220426	T42TXM_20220426
01.05	T42TWN_20220501	T42TXN_20220501	T42TXM_20220501
16.05	T42TWN_20220516	T42TXN_20220516
24.05–26.05	T42TWN_20220524	T42TXN_20220526
31.05–08.06	T42TWN_20220608	T42TXN_20220531

. The coverage scenes of the Sentinel-2 imagery over the Tyulkubas District are illustrated in Figure 2.

The images named in Table 1 are the standard identifiers for the Sentinel-2 satellite datasets. Each identifier consists of several parts that provide information about the location and date of acquisition. For example, in T42TWN_20210913, T42TXN_20210928, T42TXM_20210918 images, the first letter, T, indicates that this is a level 1C image, atmospherically corrected and ready for use.

The next number, 42, is related to the UTM (Universal Transverse Mercator) zone code.

The TWN, TXN and TXM are related to the different areas within the UTM zone, including the letter T, the latitude band strip within the zone 42. The two-letter codes WN, XN and XM are the territory squares within the UTM zone 42, designed by the MGRS (Military Grid Reference System). Each MGRS square (WN, XN, XM) is 100 × 100 km within UTM zone 42. The left part of the combination of letters W, X and X represents the square image’s position on the longitude axis. The right part of the combination of letters N, N and M represents the square image’s position on the latitude axis.

WN: W—indicates a specific longitude position (east–west) within the zone. N—indicates a latitude position (north-south) within the zone.

XN: X—another longitude square adjacent to square “W”. N—the same latitude position as square “WN”.

XM: X—the same longitude square as “XN”. M—a square located further south in latitude compared to “N”.

The square codes WN, XN and XM represent different areas within the same UTM zone 42. These squares are necessary to indicate exactly which part of the Earth’s surface was captured by the Sentinel-2 satellite. Such designations allow precise identification of the territory covered by a specific image.

The final numbers represent the image collected in the format year–month–day, YYYYMMDD, such as 20210913 (year 2021, September 13), 20210928 (year 2021, September 09) and 20210918 (year 2021, September 18).

Within each territory, square scenes (WN, XN, XM) have differences in cloudiness. The scenes without cloudiness were used for the research applications. The WN scenes were without cloudiness. Some XN and XM scenes had high cloudiness and were excluded from the data processing, which is why some columns within the XN and XM scenes are empty.

The use of Sentinel-2 mission data provided an effective tool for analyzing changes on the land surface in the Tyulkubas district, allowing us to obtain valuable information about plowed lands, sown areas and the condition of winter grain crops to achieve our research objectives.

For the analysis, calibration and verification of the processed satellite data results, the ground farmlands survey data sets were used, covering various agricultural plots, including carefully selected test GPS (global positioning system) points. The map of field survey routes and test farms in the Tyulkubas district of the Turkestan region during the period from 2 June to 7 June 2022, is presented in Figure 3. The data were collected from the farms “Bratya”, “Makulbek” and “Kyzylzhar”. The farmers of these lands provided the related field support and access to their lands. Farmers provided important information on land use, cultivation of agricultural crops, sowing and harvesting dates and explanations of their important agricultural activities, as well as problems in their fields.

The route surveys were conducted from 2 June to 7 June 2022 (160 GPS points) and during April–May 2018 (over 450 GPS points, 188 GPS points for winter and spring grain crops). The survey results included determining the crop type in the field, GPS observation coordinates, photography, developmental stage and condition of winter and spring grain crops, with fall and spring plowing activities. Visual differences in the spring conditions of fields of the main cultivated agricultural crops during the April 2018 survey in the foothill zone of the region are illustrated, for example, in Figure 4 and Figure 5. During the spring period (early April), the fields of winter grain crops stand out well due to their possessing the most developed green biomass and plant height among the general array of fields, compared to other spring-planted crops. In contrast to winter grain crops, fields of perennial alfalfa also have sufficiently developed biomass, but their tonality is lighter greenish, and there are no pronounced plant rows in the fields in spring. In fields of spring grain crops, only seedlings emerge, so these fields have a weak greenish tonality. Mass field work, as well as plowing and sowing of safflower are actively underway in safflower fields during this period, and in some early fields, the first seedlings of safflower appear (Figure 5d).

A combination of indices based on satellite data was used to identify winter crops.

2.3. Methods

To analyze changes in brightness on the Earth’s surface and extract key spectral components, the Tasseled Cap transformation was applied to multi-zone Sentinel-2 satellite data in the study. The Tasseled Cap transformation is a mathematical algorithm developed by R. J. Kauth and G. S. Thomas [27], which allows for the extraction of primary spectral components from multi-zone satellite images such as Sentinel-2 data. The Tasseled Cap transformation involves a set of coefficients applied to each spectral channel of the image to derive new components characterizing Earth’s surface features. As a result of the Tasseled Cap transformation, three key indices are obtained: the Brightness Index, the Greenness Index and the Wetness Index [28,29,30,31].

The Brightness Index derived from the Tasseled Cap transformation is a linear combination of values from various spectral bands of the image. The linear combination works as follows: the values of each spectral band are multiplied by their respective weighting coefficient, and then the products are summed together. The formula of the Brightness Index (BI) from the Tasseled Cap transformation for Sentinel-2 data is as follows [21]:

BI = 0.3037 × b2 + 0.2793 × b3 + 0.4743 × b4 + 0.5585 × b8 + 0.5082 × b11 − 0.1863 × b12,

where

• b2—is the blue band of the Sentinel-2 sensor;
• b3—is the green band;
• b4—is the red band;
• b8—is the near-infrared band;
• b11—is the shortwave infrared band 1;
• b12—is the shortwave infrared band 2.

The Plowed Land Index (PLI) is derived from the Brightness Index of the Tasseled Cap transformation. The formula for the Plowed Land Index represents an original linear combination of values from six Sentinel-2 spectral bands: B2, B3, B4, B8, B11 and B12. In the process of forming this index, each spectral band was carefully weighted and transformed using specific weighting coefficients. The signs in the formula were chosen to optimize the detection of plowed lands in the foothill zone of the Turkestan region. As a result, a new linear combination with calculated weighting coefficients for detecting plowed lands in the studied region was obtained, as follows:

PLI = −0.0037 × b2 − 0.1793 × b3 + 0.5403 × b4 − 0.5585 × b8 − 0.1082 × b11 + 0.3013 × b12.

Plowed lands are known to be characterized by low positive values. For instance, NDVI values for plowed fields typically range from 0.00 to 0.10 [32,33]. In the created PLI, its minimum positive values (0.00–0.01) also indicate the presence of plowed fields, while higher index values (0.10–0.50) indicate fires [34,35]. For graphical analysis, the following land surface types were selected: plowed fields, bare soil, dense and sparse vegetation and fires, with 5 fields of each type analyzed using 30 pixel points within each field, resulting in a total sample size of 750 points. The graphical representation of the pixel analysis of the PLI for different types of objects varied from −0.180 to 0.260 (Figure 6). The graph shows the lowest negative values of the Plowed Land Index (from −0.180 to −0.130), indicating dense vegetation, values from −0.080 to −0.040, indicating sparse vegetation, values from −0.020 to −0.006, indicating bare soil, values from 0.004 to 0.030, indicating plowed soil and values from 0.050 to 0.260, indicating fires. Thus, small positive values of the PLI allow for a clear distinction of plowed territories from the general set of analyzed objects.

Comparative results of identification of plowed fields using the Plowed Land Index and NDVI for 11 November 2021 are presented in Figure 7. The analysis of these indices showed that the most effective tool for delineating fields with autumn plowing is the Plowed Land Index (0.00–0.010), which clearly defines the boundaries of plowed lands and accurately determines their quantity and area. Determining plowed fields based on NDVI data (0.00–0.10) does not always coincide with actual plowing, as it less distinctly defines the boundaries of plowed fields. Some fields with NDVI values (0.00–0.10), as shown in Figure 7, were found to be without plowing, either in a post-harvest state or with sparse vegetation of perennial grasses or weeds. The delineation of plowed fields using the Plowed Land Index in autumn and spring most clearly defines the boundaries of plowed fields (Figure 8).

Indeed, the analysis of surface brightness changes using the Sentinel-2 images’ Plowed Land Index enables accurate identification of plowed areas in fields and their boundaries in foothill and mountainous regions.

2.4. Methodology

The methodology for recognizing the final mask of winter wheat crop sowing is based on the utilization of Sentinel-2 imagery, along with analysis of the dynamics of calculated values of the Plowed Land Index and NDVI in the research area section, covered by the collected Satellite datasets (see Figure 9). The approach involves identifying spectral changes during autumn and spring associated with soil cultivation and crop sowing, and considering the timing of field operations in the studied region.

In the foothill regions of southern Kazakhstan, autumn soil cultivation and sowing of winter wheat crops typically commence in the second half of September and continue until November. The pace and quality of autumn fieldwork are largely determined by soil moisture and precipitation [36]. However, on some fields, only soil cultivation for spring sowing of spring crops may occur in autumn. Therefore, not all plowed fields in autumn may be sown with winter wheat crops [37]. This factor has been considered in the development of the methodology for automated recognition of the mask of winter wheat crops based on satellite information.

Figure 10 illustrates the visualization of spectral changes occurring in the array of fields in the Tyulkubas district during the autumn and spring periods using Sentinel-2 imagery. A combination of channels, including SWIR2 (short-wave infrared 2), NIR (near infrared) and the green channel, was utilized for the analysis. The results of autumn plowing of fields in Figure 10a are presented as a vector layer, where plowed fields are depicted in shades of brown. Fields with shades of green characterize the absence of autumn soil cultivation or sparse vegetation.

In the spring period (late March–early April), vegetation resumes for overwintering agricultural crops (winter wheats, perennial grasses), while in other fields, spring plowing is carried out for sowing spring crops. Visualization of the spring changes in the same field array using Sentinel-2 images is presented in Figure 10b. Here, spring plowing of the fields is depicted in shades of pink, while fields with shades of green indicate spring vegetation of winter wheats and perennial grasses. Additionally, fields of winter wheat crops in spring have a darker green tone compared to fields of perennial grasses (alfalfa), which were plowed in autumn. In contrast to winter wheat crops, plowing is not observed in autumn for fields of perennial seeded grasses. These autumn and spring agronomic features of field conditions allow for their effective use in recognizing fields of winter wheat crops.

Sentinel-2 satellite data for the autumn (September–November) and spring (April–May) periods provide multi-zone images of medium resolution with high repeatability. Atmospheric correction was applied to all satellite images to eliminate the influence of atmospheric conditions on the images. Subsequently, the agricultural land field masks were digitized based on the Sentinel-2 images.

According to the methodology, the first step involves calculating the Plowed Land Index, which is an effective tool for assessing and analyzing changes in agricultural land. Subsequently, based on the identified ranges of the Plowed Land Index, the detection of autumn plowing fields was carried out in the territory of the Tyulkubas district (Figure 11).

The next stage of the methodology involves determining the spring plowing of fields for sowing spring crops based on the Plowed Land Index and excluding such fields from the mask with autumn plowing. The remaining satellite mask of fields with only autumn tillage will characterize the mask of sown winter wheats.

The methodology involves additional correction of the satellite mask for winter wheat crops based on the analysis of the increase in NDVI values in the spring period (Figure 12). To confirm the vegetation of winter wheat crops in April-May, the Normalized Difference Vegetation Index (NDVI) was calculated. Fields with winter crops in the spring period were characterized by a noticeable increase in NDVI values, while for other fields with spring sowing, NDVI values remained low. Fields plowed in autumn with very low NDVI values in spring were excluded from the winter wheat crop mask. Therefore, the active growth of NDVI values in the spring period for fields plowed in autumn served as an additional tool confirming that winter wheat crop fields were correctly identified.

The results of determining the mask of winter wheat crops using satellite data involved verifying the satellite data against ground-based survey materials and actual information from individual farms. In Figure 10, all fields identified based on satellite data are confirmed by ground-based survey data. Thus, the developed methodology allows for the successful and efficient recognition of the final mask of winter wheat crop sowing based on satellite information in real-time operation.

3. Results

Based on the developed methodology, maps of the distribution of plowed lands in the Tyulkubas district of the Turkestan region were constructed for the autumn period of 2021 and the spring period of 2022, with separate highlighting of fields where autumn and spring plowing was observed (Figure 13), as well as a map of the location of the final areas of sown winter wheat crops in 2022 (Figure 14).

The advantage of using satellite data for monitoring tasks is the ability to obtain standardized operational reports in the form of maps, tables and graphs, reflecting the dynamics of the main agricultural activities on fields at various administrative levels. As an example,

Table 2. Area of autumn plowing of fields in the Tyulkubas district in 2021 based on satellite imagery.

Observation Period	The Area of Autumn Tillage of Fields in 2021 (ha):			The Total Area across 3 Scenes (ha):	The Total Area across 3 Scenes (%):
	TWN Scene	TXN Scene	TXM Scene
05.09–13.09.2021	6159.52	272.5		6432.02	29
14.09–26.09.2021	2768	298.7		3066.7	14
27.09–13.10.2021	2374	260.1		2634.1	12
14.10–18.10.2021	1821.2	95.2		1916.4	9
19.10–31.10.2021	2475.6			2475.6	11
01.11–10.11.2021	1804.5	775.3	1571.43	4151.23	19
11.11–27.11.2021	923.9			923.9	4
01.12–17.12.2021		508.3		508.3	2
Total for the season (ha)	18,326.7	272.5	1571.43	22,108.3	100

illustrates the determination of the areas and progress of field plowing based on calculated values of the Plowed Land Index in the Tyulkubas district for the autumn period of 2021. The graphical data in Figure 15 show that during the period from September to October, the main plowing of fields was carried out on 75% of the area, in November on 23% of the area, and in the first half of December on a negligible area of 2%.

The investigated remote sensing datasets included the Autumn 2021 and Spring 2022 periods. The Autumn period of 2021 was the main investigated period of time, shown in Table 2 and Figure 15, since this was when the main agrotechnical works were provided for the winter crops. Spring 2022 was used to exclude other spring crops, which is why they are now included in Table 2 and Figure 15. The main research goal was dedicated to the Autumn 2021 period’s winter crops. Figure 15 is a bar chart illustrating the percentage change in plowing during the Autumn period of 2021. The chart displays the data by period to highlight the key stages and intensity of plowing activities: from September 5 to September 13, 29% of the total plowing work activities took place, from September 14 to September 26, 14% of the total plowing work occurred, and so on for the following periods. This approach is important in terms of the period of time, days and how long the plowing works lasted. In southern Kazakhstan, it is critical to complete the autumn plowing by the end of October, with some flexibility in terms of days, depending on weather conditions, soil moisture and temperature. This is necessary to ensure that winter crops can germinate before the onset of winter. Successful germination ensures the deep rooting of plants, which contributes to their good overwintering and, subsequently, has a positive effect on crop yields. If plowing and sowing are carried out outside the optimal period, this leads to poor wintering and, as a consequence, a decrease in yield in the spring. Thus, the data presented in Figure 15 allow us to monitor and analyze the periods of the main plowing activity and their compliance with the agronomic requirements for the optimal growth of winter crops in South Kazakhstan.

4. Discussion

The calibration and verification of the processed satellite data results were carried out using the field survey data from individual farms in the Tyulkubas district (“Bratya”, “Makulbek”, and “Kyzylzhar’’ farms). According to the farm data, 68 fields were involved in the verification process, of which 22 were fields of winter grain crops and 46 were fields of other agricultural crops. All the fields of winter grain crops identified by the satellite data were confirmed by agricultural formations (Figure 16). The verification also involved data from route surveys and the results of expert interpretations of Sentinel-2 satellite imagery, as well as official statistical data.

For satellite processed data output verification, individual field arrays from route surveys were used (Figure 17).

For these purposes, 60 fields were selected, out of which fifty-six fields with winter wheat (more than 93%) showed matching results between the satellite and ground data, with only four fields having discrepancies, mostly due to small areas. The verification of the satellite-derived winter wheat crop mask based on the Plowed Land Index included a comparison with results from expert interpretation of the images. For this purpose, the areas of autumn and spring plowing fields and the winter wheat crop mask were calculated using ArcGIS 10.8 software. As a result, the area of autumn plowing fields according to the developed methodology in the Tyulkubas district in 2022 amounted to 22,108.3 hectares. For some fields, re-plowing was conducted in spring, covering an area of 2556.4 hectares, and for a small number of fields, covering 90.0 hectares during the April–May period, there was no increase in NDVI or a very weak increase, which is atypical for winter crops. Consequently, according to the methodology, the total area of fields to be subtracted from the mask of fields with autumn plowing amounted to 2646.4 hectares. Taking this into account, the final area of the winter wheat crop mask, according to the developed technology, was 19,461.9 hectares. Additionally, visual expert interpretation of the satellite images was conducted for verification. The area of winter wheat according to the expert interpretation was 19,020.8 hectares. The difference between the areas according to the developed technology and expert interpretation was 441.1 hectares (2.3%). The official statistical data indicate that the area under winter grain crops in the Tyulkubas district in 2022 amounted to 16,885.5 hectares. The satellite results exceed the statistical data by 2576.4 hectares (15.3%). As is known, actual crop areas often differ slightly from official statistical data, a discrepancy that was confirmed by our results in this case. Farmers from the Tyulkubas district are also reviewing to develop more groundwater use programs using groundwater wells and climate change precipitation prediction analyses, which will require additional research programs, similar to those our research group’s colleagues developed in the neighboring Zhambul [38,39] and Almaty [40] regions.

5. Conclusions

The developed methodology, based on the RS applications, analysis of the Plowed Land Index (PLI) for various fields in the autumn and spring periods and ground information data, can be a supportive tool for the rapid automatic monitoring of crop production, including winter grain crops. The efficient approach involving RS data processing for the farming areas and timing of sowing, as well as the generation of maps and tabular data in the context of regions, farms and individual fields, allows researchers to identify the depression areas in fields and carry out timely agro-technical measures to improve the condition of crops. Wheat production is one of the main forms of agricultural crop production in Kazakhstan, and it is regulated and subsidized by the Kazakh government. The determination of the areas for cultivating fields and sowing crops on farms is based on taking into account the field area according to their land use map, which has not changed on farms for many years. Data from the space monitoring of land cultivation show that there is a change in the area of individual fields due to the development of new fields or changes in the configuration of fields (adding or reducing areas). Space data can be used to record all the changes occurring in fields in real time. The determination of the actual sowing mask for winter grain crops in a particular year can serve as the basis for solving further problems in the space monitoring of grain crops (assessing the condition, yield and volume of agricultural work in fields). Satellite data will allow the real-time assessment of the actual areas and timings of the sowing of winter grain crops, which will later be used to assess the conditions and carry out agricultural work in depressed areas of fields. These efforts will also help farmers in communications and ensure efficiency in the provision of information to government agencies, as well as insurance and banking institutions, in order to receive the proper level of subsidy support. The obtained results of the verification of the Sentinel-2 satellite data confirm the accuracy and reliability of using the arable land index (PLI) in identifying winter wheat planting areas. The developed methodology for using satellite data for crop monitoring can be expanded for larger regions in Kazakhstan, which is a major agricultural crop producer.

Farmland insurance institutions, banking companies and land developers have shown the most interest in expanding this methodology to Kazakhstan’s other regions.