Developing a real-time self-organizing algorithm for irrigation planning of rapeseed cultivation

Rapeseed (Brassica napus L.) is one of the main priorities in the cropping pattern under different climate conditions (Enjalbert et al. 2013). The cultivation area of rapeseed in China is estimated at 258,000 ha, which is equivalent to 3.88% of the total area under cultivation of agricultural products and 38.6% of the total area of cultivation of industrial products (Li & Wang 2022). Rapeseed is the largest oilseed crop in China and accounts for about 20% of world production. For the last 10 years, the production, planting area, and yield of rapeseed have been stable, with improvement of seed quality and especially seed oil content (Hu et al. 2017). The mean yield production values of rapeseed are estimated at 2,100 and 1,700 kg/ha in irrigated and rainfed lands, respectively. One of the distinguishing features of rapeseed among oil-seeds is the ability to be cultivated as a winter crop for rainfed cultivation. Therefore, the cultivation of strategic plants such as rapeseed is one of the main polices to achieve the sustainable development. The effect of drought stress on the yield is a function of genotype, intensity and duration of stress, climatic conditions, and growth stages (Andarziana et al. 2011; Majidi et al. 2015). Adequate fresh-water supply has become an issue of increasing local and international concern (Sedghamiz et al. 2018; Varzi et al. 2019).

In addition to the effect on society, politics, urban planning (Huang et al. 2021), environment (Zhao et al. 2022; Zhang et al. 2023), geology (Liu et al. 2023) and economy, water resources are also known as the most important factor in agriculture. Emphasizing the importance of water management in dry areas, researchers used it as a decision variable in planning systems. Lalehzari et al. (2020) proposed an artificial intelligence-based framework with two economic and productivity objectives to reduce groundwater withdrawal in southern Iran. The pattern of cultivation of different crops was developed using the simulation of production functions and optimization by genetic algorithm. Changing the duration and depth of irrigation was one of the important results to increase productivity. Tian et al. (2021) investigated the potentials of growing rapeseed across the Yangtze River Basin to serve the goal of boosting bioenergy production and improving edible oil security in winter fallow fields. The assessments take into consideration of climate change adaptations on sowing dates and on the choice of varieties with suitable growth cycle length. Results showed that a 60% realization of the production potential would increase total rapeseed supply by 9.1 million tons.

The challenges evaluated in different researches indicate the existence of three main components in the development of agricultural water management decision systems, including the required information, the simulation process, and the efficiency of the optimization model. The selection of each of these components is based on the needs of the plant, climate, soil and decision variables (Nyakundi et al. 2022). The innovation of this research compared to previous researches was in the assessment of drought stress. Deficit irrigation scenarios were programmed to evaluate the role of each irrigation on biomass production in two stages for each experiment. The previous research works have applied deficit irrigation to all irrigation events based on a specific level (Safavi & Falsafioun 2017; Sun et al. 2017; Li et al. 2018). In simulation process, the yield production and soil moisture balance were developed for rapeseed cultivation based on daily information of growth, temperature, rainfall, moisture and stresses. Lalehzari et al. (2016) and Sedghamiz et al. (2018) used seasonal or multi-day time steps. Moreover, the simulation model follows an unchanged computational process during the optimization process (Varade & Patel 2018; Lalehzari & Kerachian 2021). But in the present study, the time and amount of complementary irrigation is reorganized to improve biomass in each iteration according to the output of the PSO model. Crop growth modeling and soil moisture balance were provided using real-time programming in MATLAB 2019b software. Furthermore, due the lack of systematic programs for irrigation scheduling in most of agricultural activities, many beneficial opportunities may stay unknown during the growing season.

The most important climatic information used as input data is the maximum and minimum daily temperature, potential evapotranspiration, CO₂ concentration, and rainfall (Raes et al. 2009). For each study area, a rapeseed field was selected and soil information, irrigation, canopy cover (CC) data, and cultivating dates were collected. Climatic information of the cropping season is summarized in Figure 1.

The first step in the simulation model is to estimate the CC at different times of the growing season. This parameter is the main factor in calculating the transpiration and the separation of the transpiration values from the evapotranspiration. Table 2 shows the average CC level in different months of the growing season from March to October. The development of vegetative growth of the plant will increase the transpiration of the plant compared to evaporation and increase the final biomass. In the research of Lalehzari et al. (2016) in Baghmalek Plain in Iran, the CC level reached 72% in June, which reduced the yield of rapeseed by 6% compared to optimal conditions. Tavakoli et al. (2014) showed that a longer vegetative growth period before the onset of the cold season increases the CC area. Therefore, in a region with a severe cold season where the temperature is less than 0 °C in a 3-month period, the suitable planting time has caused sufficient growth of the rapeseed plant and reduced vulnerability in winter.

Yield production and water productivity are two other criteria for assessing the impact of changing the date of cultivation on final biomass, as shown in the figure. Under irrigated cultivation conditions, it cannot be expected that crop yields will change significantly due to changes in planting dates because only the effect of air temperature in this regard can be investigated. The difference in temperature due to the delay in planting dates has led to a slight reduction in crop yields at 32 kg/ha, respectively. In the study area, the reduction in planting date has resulted in a 12 kg increase in yield per hectare.

The previous studies have shown that delays in rapeseed cultivation reduce the growth of vegetative growth plants, reduce yields of biomass plants, and reduce yields due to increased temperatures during the reproductive stage (Javidfar et al. 2001; Faraji et al. 2009; Honar et al. 2012). Selecting the planting period earlier increases water and nutrient uptake during the fall season, resulting in increased plant growth. Danaie et al. (2014) showed that planting date is one of the most important components of climate diversity in rapeseed cultivation, which has the greatest impact on yield production compared to other managerial scenarios. As a result, by choosing the optimum planting date, the most adaptation can be made between the growth process of the plant and the climatic conditions (Javidfar et al. 2001; Ozer 2003; Sinaki et al. 2007; Faraji et al. 2009; He et al. 2017; Chen et al. 2019).

Water productivity as an indicator of performance interference and water consumption can show an acceptable assessment of the impact of changing planting dates. According to this index, the 10-day decrease in planting date due to the decrease in the volume of water used and the optimal use of precipitation to complete the vegetative growth period has increased production per unit of allocated water. Furthermore, the increase in planting date has reduced water productivity. Tayyab et al. (2022) considered that increasing water productivity requires a paradigm shift in water policy and management. Therefore, two technical and political approaches should be considered simultaneously for sustainable development. Each of these approaches separately cannot provide the needs of farmers and consumers in optimal conditions.

Using optimization methods to develop a sustainable strategy of water distribution during the growing season is necessary for achieving the best performance in reducing water consumption. The results of optimal irrigation planning showed that the growth of rapeseed in the climatic zone of Henan province without water stress requires 9,390 m³/ha of irrigation water, which with sustainable planning leads to the production of 1,408 kg/ha and water productivity of 0.15 kg/m³. Planting date is one of the effective management parameters for increasing water productivity and effective use of rainfall. Therefore, the results of modeling estimate an average decrease of about 10% in the crop water requirement and an increase of 1% in water productivity due to a 10-day decrease in planting date compared to experiments. Rapeseed was usually incorporated in rotation with crops such as rice and cereals. Therefore, attention to the crop rotation and the possibility of changing and managing the time of planting and harvesting the first crop should be considered. Effective use of rainfall and deficit irrigation can be considered in daily planning to increase water productivity.

This work was supported by the High-tech Key Laboratory of Agricultural Equipment and Intelligence of Jiangsu Province (MAET202104), University-level scientific research projects of Yibin Vocational and Technical College (ZRKY21ZDXM-02, ybzysc20bk01, ybzy20cxtd02), Grant SCITLAB (SCITLAB-1013) of Intelligent Terminal Key Laboratory of Sichuan Province.

This work was also supported by the 2021 key project of Guangxi Social Sciences think tank (self-selected project), no. ZKZXKT202285, project title: Research on the cultivation of new professional farmers in Guangxi under the rural revitalization strategy, July 25, 2022.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.