Rice farmers in Northern Ghana are susceptible to climate variability and change with its effects in the form of drought, water scarcity, erratic rainfall and high temperatures. In response, farmers resort to weather and seasonal forecast to manage uncertainties in decision-making. However, there is limited empirical research on how forecast lead time and probabilities influence farmer decision-making. In this study, we posed the overall question: how do rice farmers respond to forecast information with different probabilities and lead times? We purposively engaged 36 rice farmers (12 rainfed, 12 irrigated and 12 practising both) in Visually Facilitated Scenario Mapping Workshops (VFSMW) to explore how probabilities and lead times inform their decision-making. Results of the VFSMW showed rainfed rice farmers are most sensitive to forecast probabilities because of their over-reliance on rainfall. Also, an increase in forecast probability does not necessarily mean farmers will act. The decision to act based on forecast probability is dependent on the stage of the farming cycle. Also, seasonal forecast information provided at a 1 month lead time significantly informed farmer decision-making compared to a lead time of 2 or 3 months. Also, weather forecast provided at a lead time of 1 week is more useful for decision-making than at a 3 day or 1 day lead time. We conclude that communicating forecast information with their probabilities and at an appropriate lead time has the potential to help farmers manage risks and improve decision-making. We propose that climate services in Northern Ghana should aim at communicating weather and seasonal climate forecast information at 1 week and 1 month lead times respectively. Farmers should also adapt their decisions to the timing and probabilities of the forecast provided.

中文翻译：

---

加纳北部水稻种植系统的预测概率、提前期和农民决策


加纳北部的稻农很容易受到气候变化的影响，其影响包括干旱、缺水、降雨不稳定和高温。为此，农民依靠天气和季节预报来管理决策中的不确定性。然而，关于预测提前期和概率如何影响农民决策的实证研究有限。在这项研究中，我们提出了总体问题：稻农如何应对不同概率和提前期的预测信息？我们特意邀请 36 名稻农（其中 12 名采用雨养稻农，12 名采用灌溉稻农，12 名采用两种稻农）参加视觉辅助场景绘图研讨会 (VFSMW)，以探索概率和交付周期如何影响他们的决策。 VFSMW 的结果显示，雨养稻农由于过度依赖降雨而对预测概率最为敏感。此外，预测概率的增加并不一定意味着农民会采取行动。根据预测概率采取行动的决定取决于农业周期的阶段。此外，与 2 或 3 个月的提前期相比，提前 1 个月提供的季节性预报信息可以显着为农民决策提供信息。此外，提前 1 周提供的天气预报比提前 3 天或 1 天提供的天气预报对决策更有用。我们的结论是，在适当的提前时间内传达预测信息及其概率有可能帮助农民管理风险并改进决策。我们建议加纳北部的气候服务应致力于分别提前 1 周和 1 个月传达天气和季节性气候预报信息。 农民还应该根据所提供的预测的时间和概率来调整他们的决定。