INTEGRATING AI METHODOLOGIES IN FORECASTING MODELS FOR CLIMATE CHANGE PREDICTIONS

Abstract

Background: The bringing in of Artificial Intelligence (AI) in the climate change forecasting models would help in producing more accurate forecast results and better the measures that are taken for mitigating it. However, the use of AI in this field has failed to meet certain technical and systemic barriers.

Objective: The objectives of this research will be to ascertain quantitatively the level of preparedness of the professionals towards the use of AI in deriving climate change forecasts, the level of resistance that professionals will exhibit in incorporating AI into their modeling, and how willing they are to use it in the same process.

Methods: An online and self-completion survey with a structured format was administered to 250 respondents of the four target populations of ML/AI users, climate scientists, and environmental policymakers. To analyze the data, basic descriptive statistics and inferential statistics were applied: Lilliefor tests to check for normal distribution, Cronbach’s Alpha coefficient of reliability, correlation, and regression analysis to check the relation between AI familiarity & confidence levels.

Results: The analysis of the data unveiled rather considerable fluctuations in the perceived efficiency of AI with the help of the Lilliefors test that pointed to the non-normality of the distribution. Cronbach’s alpha of 0. The reliability analysis of the AI-indexed perception questions showed low internal consistency in 046. Hypothesis three was not supported as statistical test results revealed that there is no medium to perfect positive correlation between the degrees of familiarity with AI on the one hand and confidence in the effectiveness of the same on the other hand. Some of the challenges to the integration of AI revealed from the survey include high costs and lack of support from the government, however many of the respondents indicated interest in adopting AI for sustainability initiatives.

Conclusion: The study also shows that despite the entice for the use of AI in climate change predictions, there a challenges such as lack of funds and poor support from institutions in its application. Furthermore, raising the awareness of AI alone implies that people’s confidence in the impact of this technique will not necessarily rise either. Combating all these barriers through financial investments, policy support, and well-documented AI applications might lead to better implementation of AI in climate science.