Climate and Streamflow Reconstruction on the São Francisco Basin, Brazil, Using Tree-Ring Data

The São Francisco River crosses the most drought-prone region of Brazil and regional economic dynamics are dependent on the water availability in the basin. The seasonally dry forests are widely distributed in the basin, where Cedrela fissilis Vell (cedro) are frequently found. This semi-arid region provides a favorable setting where the deciduous cedro trees form well-defined semi-ring porous annual rings that can be exactly crossdated and used to build climate sensitive chronologies. Therefore, we have developed chronologies of cedro from seasonally dry forest fragments of three sites located in the middle-sector of the São Francisco River basin and south limit of the Brazilian Drought Polygon. The samples were analyzed according to standard procedures: sample preparation, ring count, crossdating and measurement of the tree rings. Dating quality was tested using the computer program COFECHA and ring-width time series where detrended and standardized to produce the final index chronology using the ARSTAN program. The results show that crossdating within and among trees from different sites demonstrate the potential to expand the spatial sampling. The tree-ring chronologies are sensitive with wet season precipitation totals (October - March), and can explain approximately 40% of the variance (1961-2015). Significant correlation was also observed with total annual discharge of the Rio São Francisco River measured at Barra (r=0.48; 1961-2015). However, the correlation disapears after 1993 (r=0.64 for 1961-1993, but r=-0.004 for 1994-2015) and we suspect that the stream gage at Barra has been impacted by human activity. Tree-ring chronologies can provide important information on climate and streamflow variability of São Francisco River, where hydrological records are often short and discontinuous. This chronology is now being extended with 150-yr old trees from the region and may be used to reconstruct climate and streamflow records back to the pre-instrumental period, in order to help understanding the impact and magnitude of hidroclimatic changes over the third largest Brazilian basin. (This research was funded by the Fundação de Amparo à Pesquisa de Minas Gerais - FAPEMIG project number APQ-02541-14 and NSF P2C2 award number AGS-1501321).