A seasonal outlook for crop diseases and pests is most valuable when it provides timely and accurate forecasts that not only farmers and/or extensions, but also governments can use. While farmers use the information in their crop management decisions, the government can take advantage of the seasonal outlook for the implementation of national-level disaster preparedness plans. In particular, the seasonal outlook can be utilized for collaborative disease and pest controls in Korea, because most collaborative controls are done in a preventive manner before the disease and pest infestation actually takes place. The decisions based on the seasonal outlook include which diseases or pests to focus on during the control measures, amount of agrochemical stocks needed, and what ratio of pesticide and fungicide to mix for the spray.
At the APEC Climate Center (APCC), we aim to develop a seasonal disease and pest outlook (SDPO) using the APCC multi-model ensemble (MME) seasonal climate forecast (SCF). To realize the SDPO, we focused on two studies this year. First, we developed and utilized spatial and temporal downscaling methods of the APCC MME SCF to generate SDPOs. The SDPOs were generated using two agricultural models: the EPIRICE model for rice leaf blast and sheath blight and the temperature-dependent development model for small brown planthopper. Based on the predictability of each outlook with different lead times, the potential applicability of the seasonal outlooks was evaluated. The predictability of the SDPO can be defined by the sum of errors from the SCF, each downscaling procedure, and the agricultural model. In order to quantify the relative contribution for the final uncertainty that the resulting SDPO carries, we examined each component for their uncertainty levels through a comparison with observed weather data in each downscaled station level. However, this study did not consider the uncertainty carried by the agricultural model. Overall, the uncertainties from the SCF and downscaling procedures resulted in relatively low predictability of the resulting SDPOs in the least forecasting lead time required for its practical use. Nevertheless, SDPO showed a range of predictability based on locations, resulting from different performances of the downscaling skills for each station.
Secondly, we tried to elucidate the probable relationship between climatic phenomena and disease and pest epidemics over the South Korean region using statistical methods, such as the Moving Window Regression (MWR) and the Climate Index Regression (CIR). Brown planthopper was selected for this analysis because of its migration characteristics. The brown planthopper cannot overwinter in Korea. Instead, an infestation is initiated by windborn summer migrants from China. The infestation begins with a northward migration that occurs in late March from the Indo-China Peninsula, which fits well with the concept of our statistical modelling – utilizing a long-term, multi-regional influence of certain climatic phenomena and/or indices. The MWR resulted in a promising statistical model that showed a relatively high correlation between the national infestation trends of brown planthopper in Korea and some tempo-spatial climatic variables near its annual migration path. The CIR statistical model resulted in a relatively low correlation compared to the MWR model. However, the climate indices selected from the model will provide a starting point for further analysis of the potential relationship between climactic phenomena and pest epidemics, which governs the brown planthopper infestation in the rice growing countries in Asia.
In conclusion, the SDPOs generated by downscaled APCC MME SCFs and selected agricultural models were shown to have a limited applicability due to the relatively low predictability and short lead time. However, a complete SDPO generation system starting from SCF to tempo-spatial downscaling and subsequent agricultural model application was developed in the study. This system could be adopted for other geographical areas where the SCF predictability is high enough to be utilized to generate an applicable SDPO. The statistical models developed in this study showed a promising predictability for rice brown planthopper infestation over South Korea, although the dynamical relationships between the infestation and climatic phenomena selected need to be further elucidated. In fact, these statistical models should be more applicable to the least developed countries near the equator, where the SCF predictability is relatively higher and the disease/pest infestation is mostly affected by climatic factors other than human interventions such as excessive agrochemical spray, sophisticated crop management and infrastructure, and frequent updates with resistant cultivars. Nevertheless, multi-disciplinary and multi-directional approaches from both climate and agricultural sciences should be tried in order to connect the SCF with the applicable agricultural information, which will certainly reduce the uncertainties of SDPOs. Finally, it should be noted that successful production of information, irrespective of its potential usefulness for the decision-makings of agricultural stakeholders, doesn’t guarantee its successful application without direct and consistent interaction with relevant key stakeholders.
Keyboards: Seasonal Climate Forecast (SCF), Multi-Model Ensemble (MME), Seasonal Disease and Pest Outlook (SDPO), Tempo-Spatial Downscaling, Moving Window, Climate Index