Tracer‐Aided Modeling in the Low‐Relief, Wet‐Dry Tropics Suggests Water Ages and DOC Export Are Driven by Seasonal Wetlands and Deep Groundwater

Abstract

Our understanding of how wet-dry tropical catchments process water and solutes remains limited. In this study, we attempt to gain understanding of water and dissolved organic carbon (DOC) transport, storage, and mixing in a 126 km 2 catchment of northern Australia. We developed a coupled, tracer-aided, conceptual rainfall-runoff model (SAVTAM) that simultaneously calculates water, isotope, and DOC-based processes at a daily time step. The semidistributed model can account for the marked hydrological distinction between savanna woodlands and adjacent seasonal wetlands. Using the calibrated model, we tracked the fluxes and derived the age of water in fluxes and storages. Model output matched the seasonal variability, controlled by seasonal rainfall, which switched on and off water and carbon flow pathways from the savanna to seasonal wetlands and ultimately to the perennial river. Such hydrological connectivity is modulated by the karst aquifer system that continuously contributes older waters (decades to century old) to maintain relatively stable and older streamflow during the dry season (average stream water age = 9.7 to 16.2 years). Such older waters occur despite a rapid, monsoon-driven streamflow response. The DOC fluxes were largely sourced from the wetland and riparian forest that transported DOC in the order of 1.9 t C km −2 year −1 to the stream, which was on average 90% of the total simulated DOC exports of 2 t C·km −2 ·year −1 . We conclude that coupled simulation of water and biogeochemistry is necessary to generate a more complete picture of catchment functioning, particularly in the tropics.

Net landscape carbon balance of a tropical savanna: Relative importance of fire and aquatic export in offsetting terrestrial production

“…Our DOC load estimate was also higher than that modelled by Birkel et al. (2020) for the same river system (−5.9 vs. −1.9 to −2.3 Mg C km −2 year −1 ), which we attribute to the lower (daily) frequency used in the previous modelling work which may have underestimated peak fluxes, and to the inclusion of data from a previous, drier, water‐year in the model (Birkel et al., 2020).…”

“…Our load estimates were nearly equivalent (POC) and significantly higher (DOC and DIC) than these global values, a finding that suggests that aquatic C export in the wet-dry tropics may be larger than previously thought, and that more observations are urgently needed in this region. Our DOC load estimate was also higher than that modelled by Birkel et al (2020) for the same river system (−5.9 vs. −1.9 to −2.3 Mg C km −2 year −1 ), which we attribute to the lower (daily) frequency used in the previous modelling work which may have underestimated peak fluxes, and to the inclusion of data from a previous, drier, water-year in the model (Birkel et al, 2020).…”

“…Over 90% of rainfall occurs between November and April, and, at the peak of the wet season, flooded land occupies almost a quarter of the catchment. These wetlands then surficially drain to the river until they dry out around June–July, depending on the magnitude and timing of wet season flooding (Birkel et al., 2020; Duvert et al., 2019). In the dry season, baseflow is sustained by a deep dolomite aquifer (Cook et al., 1998).…”

“…In the dry season, baseflow is sustained by a deep dolomite aquifer (Cook et al., 1998). High fluxes of terrestrially derived DIC and DOC occur during the wet season, which have been related to high connectivity between riparian floodplains, wetlands, shallow groundwater and the stream network (Birkel et al., 2020; Duvert et al., 2020).…”

Tracing Water Sources and Fluxes in a Dynamic Tropical Environment: From Observations to Modeling

“…Recent approaches used stable isotope tracers of carbon in combination with mixing models, TT models and conceptual hydrological models to assess fluvial tropical carbon export. Examples from Australia include among others, Duvert et al (2020) and Birkel et al (2020), and from Costa Rica, Genereux et al (2013) and Sanchez-Murillo et al (2020). Albeit other tracer-aided hydrological modeling approaches exist, such as particle tracking (Davies et al, 2011) and the recent Storage Age Selection (SAS) analytical models (e.g., Benettin et al, 2017), none to the best of our knowledge was yet applied in the tropics.…”