3.3. D-excess analysis
Spatial variations of reconstructed D-excess are explained by theoretical considerations such as the effects of (i) distance or proximity to main moisture sources to the study areas (Guan et al., 2013); (ii) influence of oceanic or terrestrial sources (Aemisegger et al., 2014); (iii) depletion of the heavy isotopes from coastal regions toward the continental interiors (Gat & Matsui,1991; Rozanski et al., 1993). Particularly, for the Andean region, the effects of oceanic or terrestrial sources and the depletion of heavy isotopes with distance to coastal regions are most evident (Upper left inserts in Fig. 7). This can be explained by the predominance of terrestrial sources depleted in δ18O and δ2H and the moisture transport route (e.g. from the Atlantic Ocean is larger than the Caribbean region causing major fractionations). For the Caribbean region, the proximity effect of main moisture sources to the target region is the most visible, given that, for this region, D-excess reconstruction shows predominant values close to the Caribbean Sea and the Atlantic Ocean, which agrees with previous reports from Terzer et al., (2021).
The seasonal variation of D-excess ranged from 7 to 15 ‰ VSMOW, reaching its maximum in December/January (Fig. 7a, Fig 7b) and its minimum in June/July (Fig. 7g, Fig 7h). These changes are explained by air-sea conditions at distinct water vapor sources with specific D-excess values (Gat & Gonfiantini, 1981) and modified by the mixing of inland water vapor from evapotranspiration with the air-mass (Gat, Joel, Bowser, & Kendall, 1994) and by secondary evaporation (Frits, Drimmie, Frape, & Oshea, 1987). In continental areas, the D-excess value is strongly controlled by temperature effects (δ18O temperature coefficient increases to about 0.5 ‰/°C) (Frits et al., 1987). This effect is evident in our interpolation with lower D-excess in the Caribbean region, and with higher temperatures and higher D-excess compared to the Andean region. In addition to the temperature effect, and considering the Rayleigh distillation phenomenon, δ18O values become lower when water vapor rains out at lower temperatures than oceanic sources (Dansgaard, 1964). There is an inverse relationship between δ18O and D-excess. As expected, our data in the Andean region shows this pattern, as its predominant moisture source is terrestrial, and is located away from the ocean. In the Andean region, the months that showed the highest values of D-excess correspond to the season with the low values of rainfall, with December and January being the months with the highest values of D-excess (Fig. 7a, Fig 7b), reaching up to 15‰ VSMOW. D-excess reduces in June and July (Fig. 7g, Fig 7h), due to the effect of secondary evaporation which can be especially large with small amounts of precipitation during hot dry months. April and May (the rainy season) report the lowest concentrations of deuterium excess, with a range variation between 8‰ VSMOW and 10‰ VSMOW.