Accurate and efficient irrigation has become the key to food production. To achieve effective use of water, we can rely on decision support systems for irrigation. The purpose and goal of the master's thesis is to analyze the accuracy of irrigation forecasts, which were issued by ARSO as part of the Accuracy of Irrigation Forecasting – TriN project (CRP V4-1609) with use of the IRRFIB water balance model. The trials, which data we analyzed, lasted from 1st of October, 2016 until 30th of September, 2018 and were held on 5 locations (Žalec, Jable, Bilje, Dekani and Gačnik) with different crops (hops (Humulus lupulus), potato (Solanum tuberosum), sweet cherry (Prunus avium), olive (Olea europaea) and apple (Malus domestica)) and at least a basic irrigation system present. For each location or crop we compared the daily average amount of soil water with the modeled soil water amount and the predicted and actual amount of water used in irrigation. Based on graphs, we were able to comment on special events, when the soil water content exceeded the field capacity or depleted below wilting point. We calculated the root mean square error (RMSE) and Spearmans correlation coefficient as the measurement of correspondence between modeled and measured soil water amounts. Differences between modeled and measured daily average soil water content ranged from 0 vol. % to 22,3 vol. %, with the largest differences for drip irrigation of olives during the period in 2017. Differences between irrigation amounts analyzed ranged from 0 mm to 16,4 mm. The RMSE value ranged from 1,2 to 13,7 vol.%. Spearmans correlation coefficient ranged from 0,46 to 0,87 (moderate to very strong correlation) or there was no statistically significant correlation.
|
