Precision crop production research activities were started during the mid-‘90s at the Institute of Biosystems Engineering, Faculty of Agricultural and Food Sciences, University of West Hungary. On the basis of the experiences with DSSAT (Decision Support System for Agrotechnology Transfer) the impact of climate change on maize yield (three soil types) was investigated until 2100. DSSAT crop growth model is used worldwide.
The coupled model intercomparison project global climate model ensemble (Diffenbaugh and Field, 2013, Science, pp. 486-491.) was integrated into the DSSAT model. Six validated regional climate models (RT3 program) provided data for the yield prediction.
These models (C4I-HadCM3, DMI-ARPEGE, KNMI-ECHAM5, ETZH-HadCM3Q, MPI-ECHAM5, SMHI-BCM) have daily parameters in 25 km2 resolution. A1B (this scenario assumes balance between the use of fossil fuels and renewable energy sources) scenario database was used for all six models: daily maximum and minimum temperatures, wind speed, precipitation amounts, relative humidity, potential evaporation, sunshine duration and surface radiation. Furthermore about fifty additional soil and agrotechnological parameters were taken into consideration. Matlab (Ver.: 7.14) with Mapping (Ver.: 3.5) and Statistic (Ver.: 8.0) Toolboxes were used to open and process the raw data downloaded from the project website (http://www.ensembles-eu.org/). The predicted climatic data were spatially investigated: coordinates of the experimental field were taken into consideration.
By the year 2100 four out of the six models have predicted annual precipitation between 400-750 mm. In the investigated location the average annual precipitation in the last ten years was 590 mm, which is significantly lower than the predicted. At the same time in the critical vegetation period of corn one of the models (SMHI-BCM) has predicted a notable increase (400 mm) compared to the average ten-year data (250 mm). The other models have predicted roughly 200 mm precipitation, which is rather dry. The most extreme prediction is provided by the ETZH-HadCM3Q model, which shows 120 mm.
Our experiences and calculations were carried out in the 15.3 ha research field belonging to University of West Hungary, which contains loam, sandy loam and silt loam soil types, therefore these soil types were used for the DSSAT calculations.
One of the models (C4I-HadCM3) offered prediction until 2075 the other five models (DMI-ARPEGE, KNMI-ECHAM5, ETZH-HadCM3Q, MPI-ECHAM5, SMHI-BCM) gave a prediction until 2100. It important to note, that two models have predicted significant difference for the vegetation period. The SMHI-BCM predicted 40% increase in precipitation by 2100 compared to the 250 mm actual data, at the same time ETZH-HadCM3Q model predicted 60% decrease. The earlier model calculates with 400 mm precipitation in the vegetation period, at the same time the latter calculates with 100 mm only.
One of the models (C4I-HadCM3) does not predict any changes until 2075 (we have predictions up to date), and three of them expect 20% (50-60mm) decrease in precipitation.
It is important to note that while SMHI-BCM model predicts wet conditions in the vegetation period for the investigated years, DMI-ARPEGE for the same years calculates with dry vegetation period.
It is also conspicuous that the differences in yield in the investigated loam – sandy loam –silt loam soils remain the same. Differences in the yield predictions are notable: the largest difference is predicted by ETZH-HadCM3Q model, which predicts a 50% decrease in the loam soil and a 60% decrease in case of sandy loam and silt loam. SMHI-BCM model calculates with a 6-67% decrease in all three soil types.
All models predicted decreasing yields in the loam à sandy loam à silt loam direction, which had already been prove by our experiences (Nyéki et al., 2013, In: Proceedings. 9th European Conference on Precision Agriculture. pp. 443-451).
Out of the five regional climate models only ETZH-HadCM3Q has predicted a significant difference in the year 2100 for all three soil types. Not taking the ETZH-HadCM3Q data into consideration, the average prediction for 2100 is 1.7 t/ha less in the case of loam and 1.8 t/ha less in the case of silt loam soil compared to the 2013 predictions.
Should the ETZH-HadCM3Q scenario occur, it is highly likely that this area will not be suitable for corn production.
Precision irrigation will be needed in order to keep the present stable corn yields that are currently achieved with natural rainfall.
In the case of the SMHI-BCM model scenario occurring the current situation will not change noticeably, however, should the other four models occur, precision site-specific irrigation technology has to be applied.
