GSA Annual Meeting in Seattle, Washington, USA - 2017

Paper No. 189-11
Presentation Time: 10:50 AM

PRELIMINARY CHARACTERIZATION OF BINARY KARST AQUIFERS WITH TRACER TESTS AND TIME SERIES ANALYSIS


FERRARI, José Antonio1, CALUX, Allan2, HIRUMA, Silvio Takashi3, ARMANI, Gustavo3 and KARMANN, Ivo2, (1)Instituto Geológico, SMA, Rua Joaquim Távora, 822,, São Paulo - SP, 04015-001, Brazil, (2)Instituto de Geociências, USP, Rua do Lago, 562, São Paulo - SP, 05508-080, Brazil, (3)INSTITUTO GEOLOGICO - SMA, Rua Joaquim Távora, 822,, São Paulo - SP, 04015-001, Brazil, ferrari.ig2@gmail.com

The studied site is a polygonal karst developed in a synclinal structure in the Atlantic Rainforest, southeastern Brazil. The carbonatic surface (10.4 km2) receives allogenic recharge from drainage basins (13.9 km2) formed in psammitic rocks. Two main springs drains the karst on the opposite flanks of the synclinal: Alambari (AL) and Ouro Grosso (OG). The karst is inserted in a conservation unit and the hydrological investigation supports its management.

Qualitative dye tracer tests were performed to identify recharge areas of the two springs. Monitoring stations at springs measured the water discharge (Q) and the specific conductance, (SC) every hour. The rainfall (R) was measured by a pluviometer connected to an event logger. The time series (2014 to 2016) were analyzed with autocorrelation (ACF) and cross-correlation functions (CCF) to compare the flow dynamics of both systems.

Tracer tests indicate that AL spring drains most of the area. Field observations show that the main volume of perennial sink waters is related with this spring. The average values of the parameters from the hydrologic monitoring are: AL – Q= 0.6 m3/s, SC = 137.7 µS cm-1; OG – Q= 0.1 m3/s, SC=158.2 µS cm-1. The mean annual rainfall in the region is 1250 mm. The global analysis of Q (daily average) with ACF shows that memory effect in OG is 3 times higher than the obtained for AL. The same analysis for SC shows that the memory is 1.5 times higher in AL. The CCF was used to analyze the relations between R, Q and SC time series (in hour basis). When analyzing CCF for R x Q, the maximum value occurs after 4 h for AL (r= 0.31) and after 3 h for OG (r= 0.25). Contrasting results were observed when CCF was applied for R x SC. The CCF for AL shows the usual behavior with a “negative peak” (after 13 h) that represents the pulse of fresh infiltrated rainwater, whereas OG shows a “negative peak” (after 2 h), followed by a 50 h peak (peaks identified with 99% of confidence intervals).

The results indicate that AL system is more efficient to drain, and OG system more competent to store. The slow drain in OG causes a fast increase in water mineralization, which also may explain the higher average SC in this system. The contrasting SC dynamics revealed by CCF is related with the size of the systems and with the provenance and distribution of concentrated and diffuse recharge.