Onset and demise dates of the rainy season in the South American monsoon region: A cluster analysis result

• Published in: International journal of climatology 2022-03, Vol.42 (3), p.1354-1368
• Main Authors: Rodrigues, Maria A. M., Garcia, Sâmia R., Kayano, Mary T., Calheiros, Alan J. P., Andreoli, Rita V.
• Format: Article
• Language: English
• Subjects: Annual cycles; Annual precipitation; Annual variations; antisymmetric outgoing long‐wave radiation; Climate monitoring; Cluster analysis; Clustering; Confidence intervals; Convection; Equator; Mathematical analysis; Monitoring; Monsoon convection; Monsoons; onset and demise dates; Radiation; Rainy season; Seasons; South American monsoon region; Tropical climate; wet and dry seasons; Wet season; Wind
• ISSN: 0899-8418; 1097-0088
• DOI: 10.1002/joc.7307

This work proposes an adaptation of the method developed in previous papers to determine the onset and demise of the rainy season (ONR and DER) dates in the areas of the South American monsoon system (SAMS) based on the pentad antisymmetric outgoing long‐wave radiation (AOLR). In those papers, the sign change of the mean AOLR in the central Amazon Basin (CAM) and western central Brazil (WCB) from positive to negative defined the ONR, from negative to positive, the DER dates. Since the monsoon convection presents a northwest–southeast oriented progression, the antisymmetric area to the WCB was selected subjectively. Thus, here we propose to use the Ward hierarchical clustering method to select areas in the SAMS and in the northern tropical America (NTA) for the regionalized AOLR calculation. The significant (at the 95% confidence level) negative correlations with the largest magnitude among the clusters in the SAMS and NTA and the outgoing long‐wave radiation (OLR) and precipitation annual cycles in each group define the pairs to calculate the AOLR. Then, the AOLR time series is calculated and subjected to a 5‐pentad running mean filter. This method keeps the climatological features of the convection annual cycle such that the closer (farther) the pair is to the equator the longer (shorter) the rainy season. The ONR and DER dates found with this new method are remarkably close to those found previously. Therefore, the new method proposed here highlights regional aspects of rainy season and can easily be automatized for its routine application at the operational climate monitoring centres, for instance at INPE. This is the most important advantage of the method and might be relevant to the SAMS rainy season monitoring. The Ward hierarchical clustering method is used to select representative areas of the South American monsoon system (SAMS) and of the northern tropical America regions. Using these areas, regionalized antisymmetric outgoing long‐wave radiation (AOLR) indices are constructed to determine the onset and demise dates of the rainy season in the SAMS. The AOLR indices keep the climatological features of the convection annual cycle. Therefore, these indices are useful for the SAMS rainy season monitoring purposes.