How can I assess, using dissolved load data (from surficial waters), if the surficial waters are losing water to ground water (regional acquifer)?
According to Velbel and Price (2007), if you are using solute gauging as the technique to obtain chemical weathering estimates, you need to be sure that the watersheds do not have water depletion to a regional aquifer.
You might look into end member analysis such as Garrett ET al.
https://www.researchgate.net/scientific-contributions/C-Guinn-Garrett-2028286718
If working with experimental catchments, water balance techniques may be helpful, but demanding work to try to assess for deep seepage losses. A geologist may help determine if the aquifer is confined. Karst topography can have seepage losses, or in some instances additions from springs.
If working in Puerto Rico, you can contact the Institute of Tropical Forestry and USGS. There are quite a few stream gauging stations, especially in the Caribbean National Forest. Some streams are flowing on bedrock. Most were igneous geology, but I think Rio Icacos was granitics, and was landslide failure prone. If you review the flow duration curves, you might be able to separate the flashy runoff watersheds with little baseflow from the muted runoff watersheds with elevated baseflow from groundwater. In selecting experimental watersheds or catchments, research hydrologists typically would try to avoid conditions where deep seepage losses are likely. However, in some geology types such as karst research, this may be difficult.