ABSTRACT
To generate information about the monsoon onset and withdrawal we have to choose a monsoon definition and apply it to data. One problem that arises is that false monsoon onsets can hamper our analysis, which is often alleviated by smoothing the data in time or space. Another problem is that local communities or stakeholder groups may define the monsoon differently. We therefore aim to develop a technique that reduces false onsets for high-resolution gridded data, while also being flexible for different requirements that can be tailored to particular end-users. In this study, we explain how we developed our technique and demonstrate how it successfully reduces false onsets and withdrawals. The presented results yield improved information about the monsoon length and its interannual variability. Due to this improvement, we are able to extract information from higher resolution data sets. This implies that we can potentially get a more detailed picture of local climate variations that can be used in more local climate application projects such as community-based adaptations.
