Estimation of rapeseed yield using Geo-data and remote sensing (Case study: Selected Lands cultivated in Qazvin plain irrigation network)

Among the applications of remote sensing in agriculture, we can mention the estimation of crop yield, the preparation of the cultivation map, the factors affecting the crop yield. The models presented to estimate the crop yield are generally based on the calculation of vegetation indicators, which are used to estimate the amount of production using these indicators and with a specific algorithm. Researchers have used other methods (in addition to the direct use of vegetation indices) to estimate crop yield. In this regard, we can refer to Bastianssen and Ali's research (Bastianssen and Ali, 2003). This model (Bastiansen model) is a combination of the Monteith model to calculate the absorbed photosynthetic active radiation, the Stanford model to determine the absorbed energy efficiency, and the SEBAL model to describe the spatial-temporal changes of evapotranspiration.

This research was carried out in rapeseed fields in the cultivated lands of Qazvin plain irrigation network. In this research, the fields were selected to cover soil texture, soil salinity, different crop management, irrigation water salinity and different irrigation methods. In order to be able to analyze the leaf area index in the process of crop performance modeling, rapeseed cultivars were the same in all the selected fields. In this research, a hybrid model was used to estimate crop yield, including the Monteith model to calculate the absorbed photosynthetically active radiation (APAR), the Stanford model to determine the light consumption efficiency (LUE), and the surface energy balance algorithm (SEBAL). In order to evaluate the crop yield prediction model, Pearson's correlation coefficient was used between the data to analyze the correlation of yield and leaf area index in different stages of growth.

The analysis of the leaf area index in the studied fields showed that the date of cultivation was one of the most important factors influencing the process of plant phenological growth and consequently the difference in crop yield in the fields. Considering that the potential yield of rapeseed in the Qazvin Plain irrigation network is estimated at 4000 kg/ha, none of the farms have reached the maximum leaf area index, and considering the direct effect of the leaf area index in the flowering stage on the crop yield, the maximum yield potential in the selected farms is not available. Therefore, the leaf area index in the flowering stage is considered a suitable criterion for estimating the yield reduction of rapeseed. The results of Pearson's correlation coefficient analysis showed that crop yield had a significant correlation with leaf area index in development and middle stages of rapeseed growth, and the highest correlation was related to the middle stage of crop growth.The results of Pearson's correlation coefficient analysis showed that there was a significant correlation at the 1% probability level between the field recorded data and yield estimation values. Also, the values of explanation coefficient (R2), root mean square error (RMSE), mean bias error (MBE), mean absolute error (MAE) were equal to 0.91, 444.06, 41.23, 433.03 kg/ha respectively. Is. Also, the results of the correlation coefficient analysis of yield values and calculated evapotranspiration based on the SEBAL method showed that there was no significant correlation.

Several factors are effective in product performance, but modeling by simplifying the relationships related to a phenomenon, justifies the mutual relationships between independent and dependent variables by spending the least amount of time and money. The results of the research regarding yield estimation using vegetation indices, evapotranspiration algorithms and hybrid models show that it is possible to make an acceptable estimate of crop yield by using Remote Sensing techniques. For example, the results of present research showed that by preparing the selected image of Landsat 8 satellite (OLI and TIRS) related to the beginning of rapeseed flowering period in the following years and extracting the leaf area index in the middle period of growth, the yield of field can be predicted with reasonable accuracy. Also, the leaf area index in the rapeseed flowering stage is a suitable measure to estimate the yield gap of the rapeseed crop. The important point is that the accuracy of predicting crop performance by satellite images is still reported was average. The accuracy of field measurements, the low spatial resolution of satellite images, as well as the presence of clouds, fog, gas, and suspended particles, along with the complexities related to plant growth modeling, have an effect on reducing the accuracy of yield prediction and the validity of models. Although these researches are expected to improve and expand with the variety of satellite images and the entry of cloud computing into the field of complex computing.