زهرا مرزبان قره خانلو

Agricultural water allocation is dependent on forecasting and estimating various hydrological, agricultural, and economic factors. However, the prediction of those factors is uncertain, so this research is aimed to provide an irrigation scheduling model taking into account uncertain hydrological (rainfall, evapotranspiration, and soil moisture) parameters. Additionally, this inquiry can distinguish between five roots of hydrological uncertainty including initial condition, input data, parameters (soil porosity, effective root zone, relative soil moisture level under which the plant’s stomata begin to close), model structure (soil moisture simulation module), and observation (soil moisture). A novel integrated simulation-data assimilation approach is presented that could identify the system (calibration-validation) dynamically. Soil moisture dynamics are calculated through the simulation part, and assimilation of the (artificial) observation is performed by using Dual state–parameter estimation using the ensemble Kalman filter. The combination of simulation and data assimilation shows the capability of this model to estimate parameters and state variables correctly under different conditions. This research indicates the effects of ensemble sampling size, irrigation technologies, and time step interval of the data assimilation on the model. The optimization is also based on the particle swarm optimization (PSO) algorithm and taking into account the goal of maximizing the net profit from irrigated agriculture. In this research, artificial experiments are used to eva‎luate the proposed performance in two stages. The first step is simply to eva‎luate the correct performance estimates and model models taking into account the uncertainty of the five hydrological roots. In the second stage, the proposed plan in estimating the change of state-parameter as well as achieving the optimal decision-making decision is the goal in the conditions of climatic, agronomic and economic uncertainty. The results show that with the change of the target quality in irrigation supply and crop yield, the difference between the target value and its value becomes less than 5% and for neutral recipients, this difference is negligible. However, with increasing decision-making conservatism and selective agro-economic change, there is a 7 to 10 percent difference between the results. It is worth noting that in all cases the system detection is done correctly. It is also shown that if the decision maker is neutral or a significant amount of water supply increases by a maximum of 30%, the area to the irrigation area increases by 76% and the average amount of irrigation decreases by 18%.

شبيه‌سازي رطوبت خاك , الگوريتم بهينه ساز PSO , جذب داده , فيلتر آنسامبل كالمن , تخمين دوگان , خطر پذيري تصميم گيرنده

Soil moisture simulation , Particle Swarm Optimization , Data assimilation , Ensemble Kalman filter , Dual estimation , Decision-maker risk-taking

Zahra Marzban Gharehkhanloo

Hosein Alizadeh