Pragmatic national fire outbreak intervention planning requires intelligent data generative models capable of integrating the varied benefits of decentralization underlying the data generation processes. This paper focuses on developing realistic national fire outbreak data generative models based on regional data generation processes with region-specific spatial data issues controls. The concept of compound count random variables is adopted to activate the natural assumption of empirical mixtures with non-parametric weights such that the resulting data generative model is robust to data issues at the region-level. This allows data challenges to be treated at the observation level within regions so that regional resources are factored in the analysis with avoidance of bias. The resulting data is a consolidated national fire data that avoids large observations which may yield low probabilities of occurrence of event whiles indeed the event can rapidly occur. Results based on real fire data example provides evidence on the effectiveness of the methods in generating better national data with unique calibration of regional contribution towards national intervention planning.