SYNOPTIC CLIMATOLOGY AND CHANGES IN PRECIPITATION ASSOCIATED WITH THE SOUTH ATLANTIC CONVERGENCE ZONE UTILISING A CLOUD BAND IDENTIFICATION TECHNIQUE


 <p>Austral wet season precipitation (October through March) in the subtropical parts of Brazil is related to the strength and position of the South American Convergence Zone (SACZ), one of the main features of the South American Monsoon System. The SACZ can be defined as the aggregation of individual tropical-extratropical (TE) cloud bands. Such TE cloud bands have deep convection and heavy rainfall linking the tropical convection over the Amazon rain forest to the mid-latitude weather systems in the Southern Ocean. Utilising a cloud band identification technique, which consists of an object-based algorithm that identifies TE interactions, we detected individual weather systems and explored their associated precipitation characteristics and changes since 1980. Each event is characterised by the total precipitation within the contour of the low-value OLR. For this, we considered three different datasets: observed precipitation from various weather stations over Brazil, gridded to a 0.25&#176; lat/lon resolution; satellite-based rainfall from TRMM (version 3B42); and reanalysis-based precipitation from ERA5. Here we explore the spatial characteristics and associated precipitation statistics of the SACZ events identified through the proposed technique. The monthly spatial signature of the selected events is similar among the three data sources and corresponds to the SACZ location. The selected events account for 25% to 50% of the total monthly precipitation during the wet season, with the largest percentages occurring in December and January. Over South-eastern Brazil, we identified a reduction in the number of events and in total precipitation during these events, resulting in a reduction of their contribution to the total precipitation climatology during the last decade. The drying trends occur mostly in December; in January, the areas with reduced precipitation migrate northward and precipitation increases over Southern Brazil, suggesting that the poleward migration of the SACZ is more pronounced during these months. These results demonstrate the relationship between synoptic systems and the changes in the location of the SACZ described in recent studies. In the next steps, we will diagnose the reanalysed and climate-simulated circulations associated with these events, identifying possible mechanisms responsible for the poleward shift of the SACZ.</p>



CHANGES IN PRECIPITATION ALONG THE SACZ
al 2016) (Zilli, et al 2019) Observed changes in seasonal (DJF) precipitation [Zilli et al 2016]: ▪ Reduction (increase) in precipitation over eastern (southern) Brazil (left) ▪ Reduction (increase) in the number of rainy days over southeastern (southern) Brazil (right).
Changes in circulation over recent decades [Zilli et al 2019]: ▪ Weakening of the poleward winds along the northern margin of the SACZ ▪ Southwestward shift of the SACZ, reducing (increasing) precipitation along its equatorward (poleward) margin.
➢ What are the main synoptic characteristics associated with individual Tropical-Extratropical (TE) cloud band events, such as the SACZ? ➢ How are TE events changing over recent decades? What are their contribution to observed changes in seasonal precipitation?

Number of days with events
persisting 4 or more days (contour) and their contribution to climatology. Values within the signature of TE cloud bands. → ▪ Algorithm identified 67% of the days with observed SACZ ▪ Algorithm identified more (fewer) events than observed over tropical (subtropical) Brazil ▪ Discrepancies: algorithm focused on identifying TE cloud bands, independent of persistence and forcing. ▪ Decrease in number of days with TE events during entire rainy season, except in Jan. ▪ Delay of the peak at the end of the rainy season.

EFFICIENCY OF THE ALGORITHM
▪ Increase in total number of TE events in the beginning (Nov) and end (Feb and Mar) of the rainy season (inset). ▪ Peak in median number of days with TE events anticipated to Dec, during which the variability also increased.

CHANGES IN PRECIPITATION DURING TE EVENTS -INTENSITY
Changes in average precipitation rate (left, in mm.day -1 ) and in their contribution to the monthly total precipitation (right, %) considering TE events persisting 1-3 days (top) and 4 or more days (bottom). Changes are significant in months with filled symbol. Values spatially averaged within the signature of the TE cloud bands.
▪ Dec: precipitation rate is increasing (decreasing) during TE events persisting 1-3 (more than 4) days (left). ▪ Changes affect the contribution of individual days with TE events to the total precipitation only when considering TE events persisting 4 or more days (right).
▪ Precipitation rate during TE events is increasing in Jan-Mar, independently of the event's persistence (left).

SPATIAL CHANGES IN TE EVENTS OVER RECENT DECADES
Changes in number of days with TE events (contour) and total precipitation during these events (shades), considering events persisting 1-3 days (left) and 4 or more days (right). Significant differences (p < 0.1) are stippled. Dataset: gridded observation (left column) and ERA5 (right column). Precipitation values within the spatial signature of TE cloud bands.
▪ No significant changes in total precipitation during TE events persisting 1-3 days (left), except for an increase over southern Brazil in Jan ▪ Changes in the number of TE events persisting 4 or more days, and consequently on their total precipitation, suggests a poleward displacement of the SACZ, specially in Dec and Jan.

CONTRIBUTION TO CHANGES IN CLIMATOLOGY
Left: total precipitation (contour) and its changes (shades). Right: Changes in number of days with TE events (contour) and their total precipitation (shades), considering events persisting 4 or more days. Precipitation only within the spatial signature of TE cloud bands. For both panels: significant differences (p<0.1) are stippled; dataset: gridded observation (left column) and ERA5 (right column).

1-3 days ≥4 days
Composite of the vertical profile of zonal winds zonally averaged over South Atlantic (-40°W to -20°W) during Jan with large number of TE cloud band events persisting 1-3 days (left) and 4 or more days (right). Contour: monthly average; shades: monthly anomalies.
Left: median (line) and interquartile range (envelope) of the number of days with TE events identified using NOAA CDR (blue) and HadGEM3 n216 (hatching) OLR data. Right: Spatial representation of the number of days with TE events (contour) in NOAA CDR (blue) and HadGEM3 n216 (red) and the difference between the sources (shades, with red shades indicating fewer events in the model).