Modeling tropical Fresh Submarine Groundwater Discharge (FSGD) and its associated nitrogen fluxes at regional scale
Fresh submarine groundwater discharge (FSGD) is a poorly studied flux and water pathway that can discharge a fraction of terrestrial nitrogen (N) surplus into the ocean. It is assumed that FSGD plays a major role in driving water quality in the coastal ocean and marginal seas, especially in the tropics. Rapidly changing land use in subtropical and tropical catchments to farmland results with an increased fertilizer application in more N losses to aquatic ecosystems. Relatively high tropical monsoon rain amounts suggest a high potential for FSGD draining groundwater to the ocean. Furthermore, coastal urbanization is found to function as an additional source of coastal N especially in the tropics, were most coastal mega cities are found. While oceanic N inputs through rivers are well studied, because the integrated lumped catchment N loss is drained through one easily accessible point at the surface, diffusive groundwater discharge occurring at the interface between ocean and aquifer at long coastal stretches are essentially invisible and submarine, which makes them difficult to quantify. As a result, directly measuring FSGD requires great effort and is currently only feasible at local scale, which motivated me to develop the regional transient lumped water balance model CoCa-RFSGD to quantify daily FSGD at long stretches of coastlines at the whole island of Java, Indonesia. The model calculates water balances for and fluxes between top soil, sub soil and aquifer at areas, which are not part of a larger river catchment, discharging groundwater largely to the ocean. If top soil is saturated surface runoff occurs forming seasonal rivers, which drain additional precipitation into the ocean. To estimate N recharge into the aquifer the soil modules were extended by a mass balance model including N transformations hydrolysis, volatilization, nitrification, denitrification and sorption processes. This model was specifically adapted each for two research questions: (1) Is catchment N surplus derived from creek water quality measurement related to land use types in sub-tropical basins in NSW, Australia and (2) does fast-changing coastal land use patterns in tropical Kerala, India, from largely traditional home gardens to urbanization explain strongly raised nitrate and ammonium concentrations in coastal springs? Java generates a total FSGD of annualy 15.27 km3. Especially areas with high top soil waterholding capacity and precipitation surplus generate large FSGD locating Indonesia, Vietnam, Nigeria, Cameroon, Equatorial Guinea and Columbia as potential FSGD hotspots. The model can be applied anywhere in the world to locate regional FSGD hotspots without obtaining in situ data, because it can run solely with global data sets, which makes it especially feasible for the locations mentioned. Furthermore, CoCa-RFSGD extended by N mass balances revealed for sub-tropical basins in Australia that a horticulture proportion of 3% drove a 3.5-fold increase of N losses to water ways and a 6.7-fold increase of N losses to atmosphere compared to catchments without horticulture. The blueberry horticultural land use lost 92 kg-N/ha of which 85% evaded to the atmosphere and 15% discharged through surface waters. In Kerala, India, paddy fields showed the highest contribution of N to groundwater with annually 56 kg-N/ha (standard deviation (SD)=13 kg-N/ha), rubber plantations and home gardens contributing 7 kgN/ha (SD=2 kg-N/ha) and 14 kg-N/ha (SD=4 kg-N/ha), respectively, while urbanization resulted in an increased amount of N stored in pit latrines, depending on pH within the pit latrine. Analyzing N groundwater recharge and historical land use changes from 2002 to 2015 of that region suggests that while fertilizer application rates doubled, tourism increased 13-fold, coastal spring N discharge increased 10-fold. Large N inputs to groundwater must originate from pit latrines, explaining high ammonium concentrations just after the rainy seasons, which could not be reproduced modeling fertilizer N recharge. Overall, my lumped model provides a simple but effective tool to upscale FSGD point measurements to a full year and to upscale local aquatic water quality measurements to catchment scale. This allows me to assess the importance of a changing land use on aquatic nitrogen loads. Tropical catchments face large groundwater recharge, hence large FSGD, and at the same time, due to saturated soil and high soil temperatures, high atmospheric losses of N.