Soil background and vegetation cover were other factors studied in this research, which also often have a high impact on the quality of image analysis , mainly in traditional pixel-based classifications. On the one hand, the OBIA4tillage procedure was insensitive to soil hue, reporting similar good results for both clear and dark soils . On the other hand, the OBIA4tillage procedure reported the best results in areas with extreme values of ground vegetation coverage, i.e., either in zones of dense vegetation or in bare soil , while it had some imprecision in estimating tillage direction in zones with medium vegetation density , as the tillage marks were more difficult to identify in these latter scenarios due to the impact of irregular vegetation patches on ground structure. Previous investigations in similar olive grove scenarios have shown that UAV-based multi-spectral images achieved better estimates of ground vegetation density in cover interval ranges of 10–15% at both high and low values of vegetation cover and in ranges of 30% at medium vegetation cover values , showing that remote sensing analysis depends on the relative weights of vegetation and soil components in complex agricultural scenarios . The OBIA4tillage procedure was valid to identify the signs of tillage operations with high accuracy in the images collected by both sensors, although the results obtained in the conventional-colour images of the RGB sensor was slightly more precise than the results obtained in the colour-infrared images of the multispectral sensor , which can be attributed to the higher spatial resolution of the former with respect to the latter . The OBIA4tillage procedure did not base its decisions on individual pixels but on groups of adjacent pixels with similar spectral responses that served to define the typical tillage objects in the image, regardless of the image type used . By comparing the spatial and spectral resolutions of both sensors used, the critical aspect in tillage identification was to have remote images with GSD of a few centimetres, rather than having images with more spectral bands or infrared information, which led to the OBIA4tillage procedure to obtain better results with the RGB sensor because its spatial resolution exceeds that of the multi-spectral sensor.
The OBIA4tillage outputs were three georeferenced raster files in GeoTIFF format arranged in a grid of 0.5 m2 cells containing a value of tillage direction, slope aspect and contour farming ,hydroponic gutter which can also be exported as a vector file in shape file format and as a tab file in ASCII format for further assessment at intraparcel and interparcel levels . A complete analysis of these files would allow the user to know the plowing patterns applied by each farmer, to identify the predominant and/or unusual tillage directions and to map the areas where contour farming has or has not been applied according to the land slope, which opens up various topics for discussion. For example, in the study region, a homogeneous pattern with a predominant tillage direction was observed in most parcels, which in many cases did not coincide with the longest side of the parcel, thus contradicting the baseline assumption adopted in previous methods for mapping tillage direction . Tillage patterns were different for each parcel due mainly to the personal decision of each farmer, which was based more on economic than environmental criteria. The application of contour farming in the study region was moderate in location 1 but relevant in location 2 , which revealed an uneven involvement of the local farmers in the challenge of controlling soil erosion risks. According to a farm survey in the study locations, the primary criterion of the local farmers was to reduce the number of tractor maneuvers, which in turn depended on a combination of factors such as the size and shape of the parcels, their topographic characteristics, planting framework and physical obstacles. The shifts in the tillage direction compared with the prevalent route occurred only in very localized areas and mainly in the turns at the end of each pass, when bordering trees and avoiding obstacles or singular elements inside the parcels. This valuable information can be integrated into existing models for tillage translocation that are limited in complex scenarios with numerous trees or obstacles and can feed the algorithms that predict soil redistribution arising from different patterns of tillage in a given parcel .
Overall, the OBIA4tillage procedure represents a methodological improvement for the study of tillage lands compared with previous research , and this procedure was proven to be an effective tool to support studies on soil erosion risk assessment, as noted by Rawat et al. , because the tillage direction maps generated with this procedure could serve to refine the estimations of the tillage transport coefficient and the supporting practice factor, and thus contribute to the achievement of an ideal model for estimating soil loss from arable land as argued by Panagos et al. ; enable effective and timely control of tillage events through the adoption of a remote sensing system that reports the occurrence of tillage at various spatial and temporal scales ; and provide governments and public agricultural agencies with innovative tools to monitor tillage areas and verify the implementation of GAEC measures related to efficient soil management, such as ensuring compliance with contour farming strategies as stipulated in agricultural regulations such as the European Common Agricultural Policy . Land and soil are essential resources for the survival and development of human societies as they provide people with necessary goods and services . Approximately 38% of Earth’s terrestrial surface is used for agriculture and has historically been affected by soil degradation . Soil erosion can lead to land and soil degradation , low and declining agricultural productivity, poverty, food shortages, and nutrition insecurity , which has greatly hampered the sustainable development of agriculture and rural areas in developing countries and has become a global concern . Approximately 2.71 106 km2 of the land in China was affected by soil erosion, accounting for 29.8% of the total territory according to the 2019 Monitoring Bulletin of China’s Soil Erosion . Areas that have undergone severe soil erosion in China are concentrated in poverty-stricken areas because soil erosion is closely associated with the loss of soil nutrients, decreased productivity of cultivated land, food insecurity, and frequent natural disasters . The intense soil erosion that has occurred in the Loess Plateau is world-renowned . The subregion known as the Loess Hills accounts for approximately 60.7% of the plateau area and suffers from the most severe soil erosion in China . Over the past four decades, as the foundation of national economy, China’s agriculture has led to remarkable progress due to the opening of the country to the global market, institutional innovation, increase in agricultural input, and adoption of new technologies . However, the development of agriculture is still hampered by soil erosion. In particular, soil erosion in the Loess Hills not only impacted the development of local agriculture but also led to ecological and environmental problems in the Yellow River Basin. As a result, the Chinese government has launched a series of projects in an attempt to control soil erosion and conserve the upriver environment.
These projects aim to simultaneously improve farmers’ livelihoods and alleviate poverty. Several key projects nationwide include the Grain for Green Program , small watershed management projects , and other soil conservation projects that built terraces on sloping arable lands and check-dams in the valleys . These projects have made substantial progress in terms of environmental conservation . In particular the GGP, also known as the Sloping Land Conversion Program the largest arable land retirement program worldwide, has substantially contributed to land cover increases in participating areas and has made great progress toward achieving its initial goal of reducing soil erosion . Some aspects of soil conservation are advantageous for agriculture, examples include the retention of soil moisture, reduction in production costs, increased in the production of crops, fruits, and timber in India and western Honduras , the improvement in soil properties and crop productivity of the treated watersheds in Ethiopia , and the reduction in runoff and sediment yield while simultaneously boosting corn yield of participating areas in Kenya . Evidence from China indicates that the land use structures, agricultural structure, output value, and livestock farming in the participating areas of the GGP have improved significantly during the program . However, several studies have suggested that the GGP has, to some extent, restricted livestock farming involved in these projects. Due to differences in the technology level and the scope of application, projects that addressing the conservation of soil and water have had both positive and negative effects on local agriculture and the livelihoods of farmers in developing countries . Several studies have also indicated that soil and water conservation may reduce crop yields depending on the conditions of a specific site, economic development, agricultural regime, and public policy . Overall, the previously mentioned studies have primarily focused on the unidirectional impacts of soil conservation activities on agricultural development or local livelihoods, while ignoring the interactions among agricultural factors, such as labor force and financial capital, in response to soil conservation. Furthermore,hydroponic nft channel information on whether such interactions may influence the relationship between environmental projects and agricultural development is lacking. The Ansai District of Yan’an Prefecture in northern Shaanxi Province and the Anding District of Dingxi Prefecture in Gansu Province were selected for this study, as both are representative of dryland farming in the Loess Hills, have different erosive environments and agricultural conditions, and are areas in which distinct farming activities are practiced. The main objectives of this study were to examine the spatio-temporal changes in the soil erosion rate following the implementation of soil and water conservation projects and their effects on local agriculture; identify the controlling factors that affect local agriculture and reveal how these factors interact with each other to indirectly influence agricultural effectiveness; and determine how such interactions and influences differ in the dryland farming areas over time. This study is expected to provide new insights into the relationship between the protection of erosive environments and agricultural development in the Loess Hills.
The soil erosion intensity modelled in Ansai district mainly included moderate, strong, and extremely strong erosion, accounting for 21.81, 17.34, and 17.45%, respectively, of the total erosion area that occurred in 2000 . Moderate and severe erosion affected 24.15% of the farmland and 39.35% of the grassland in the area, indicating that approximately 63.5% of the soil erosion with a rate of more than 25 t ha 1 yr 1 was associated with farmland and grassland. While extremely strong erosion was dominant in the grasslands in 2000 , moderate erosion was prevalent in farmlands. Farmlands and grasslands with moderate or severe erosion were predominantly distributed on the gentle slopes in the southern part of Ansai . By 2017, in terms of the results of CSLE, mild and moderate erosion dominated the farmland and grassland of Ansai, affecting 25.48 and 28.92% of the total area of farmland and grassland, respectively, whereas strong or extremely strong erosion had been successfully controlled in these areas. Strong, extremely strong, and severe erosion occurred on 13.09, 4.62, and 0.31% of the land in this district, respectively, and predominantly occurred on the steep slope wastelands that were scattered northwest of Ansai . Compared to Ansai, the intensity of soil erosion modelled in Anding gradually decreased from north to south in 2000 and mainly included moderate and strong erosion, accounting for 41.67 and 29.17% of the total erosion, respectively . Moderate or severe erosion of farmland accounted for 25.05%, while 52.81% of the erosion occurred in grasslands, indicating that approximately 77.87% of the soil erosion, or more than 25 t ha 1 yr 1 , occurred in farmlands and grasslands. Moderate erosion affected 28.45% of the grassland and 11.46% of the farmland in 2000. By 2017, severe soil erosion had been completely alleviated , and mild and moderate erosion dominated in both farmland and woodland in Anding, accounting for 27.28 and 20.68%, respectively.Strong, extremely strong, and severe erosion accounted for 11.45, 2.77, and 0.43% ofthe total erosion, respectively, and predominantly occurred in the northeast of Anding .