Testing δ^18 O and δ^2 H isoscapes in the hyperarid Atacama Desert

Spatial models of variations in the stable isotopic composition (isoscapes) of water resources have become a key tool to portray factors influencing the hydrological cycle across the landscape. Although isoscapes could contribute significantly to the water management of water resources in different environmental contexts, the performance of interpolation methods should be addressed in the light of the data structure, but also of topographic complexities. Here, we evaluated the influence of low-density data and abrupt and rugged relief on the performance of geostatistical (ordinary kriging - OK) and mechanistic (Inverse Distance Weighted -IDW) interpolation methods in estimating spatial variations in and values of meteoric waters across the western slope of the Andes (17.5º-24.5ºS). Specifically, we focused on the Atacama Desert -the world´s driest desert-. That is, we generated a regional database that gathers pre-existing and novel stable oxygen and hydrogen isotope measurements of regional meteoric waters. The irregular spatial distribution of these data led us to test the performance of interpolators by generating independent oxygen and hydrogen isoscapes for two distinct zones within our study area: Zone I (18°S-21°S) and Zone II (22º-24ºS). These zones were defined according to the density of oxygen and hydrogen stable isotopic data. Our results point to a higher performance of the IDW interpolator compared to OK approaches. We indeed evince that the IDW isoscapes capture quite well the predictable elevation effect on the distribution of oxygen and hydrogen stable isotope values across the southern portion of the Zone I, but also distinct isotopic enrichments over Altiplano basins as well as the coastal area of both hydrological zones. This method, however, fails in predicting the elevational distribution of 18O/16O and 2H/1H ratios over the northern Zone I and the entire Zone II due to the irregular spatial distribution of data. Although such limitations exist, the IDW interpolator appears as the best tool for generating oxygen and hydrogen isoscapes in our study area. This is consistent with other studies demonstrating that IDW is a superior interpolator when dealing with spatially uneven grids. In this sense, we challenge the prevailing notion that the geostatistical OK approach is invariably the most effective technique for spatial analyses of stable isotopes, and in turn the default method to implement isoscapes.水资源稳定同位素组成的空间分布模型（同位素景观（isoscapes））已成为刻画景观尺度上影响水文循环各因素的核心工具。尽管同位素景观可在不同环境背景下为水资源管理提供重要支撑，但插值方法的性能不仅需结合数据结构进行评估，还需考虑地形复杂性。本研究针对安第斯山脉西坡（17.5°S-24.5°S）区域，评估了低密度数据以及突变崎岖的地形对两种插值方法性能的影响：地统计方法（普通克里金（ordinary kriging，OK））与机理方法（反距离加权（Inverse Distance Weighted，IDW）），这两种方法分别用于估算区域大气降水水的δ¹⁸O与δ²H值的空间变异特征。本研究聚焦于世界最干旱的沙漠——阿塔卡马沙漠。具体而言，我们构建了一套区域数据库，整合了已有的与新测得的区域大气降水稳定氧、氢同位素实测数据。由于这些数据的空间分布不均，我们针对研究区内两个独立分区生成了各自的氧同位素景观与氢同位素景观，以此检验插值器的性能：分区I（18°S-21°S）与分区II（22°S-24°S），分区划分依据为稳定氧、氢同位素数据的采样密度。研究结果显示，相较于OK方法，IDW插值器的表现更优。我们证实，IDW生成的同位素景观能够较好地捕捉分区I南部区域内高程对稳定氧、氢同位素值分布的可预测影响，同时也能反映阿尔蒂普拉诺盆地以及两个水文分区沿海区域的显著同位素富集现象。然而，由于数据空间分布不均，IDW方法无法准确预测分区I北部以及整个分区II内¹⁸O/¹⁶O与²H/¹H比值的高程分布特征。尽管存在上述局限，IDW插值器仍是本研究区域内生成氧、氢同位素景观的最优工具，这与其他研究结论一致——即当处理空间分布不均匀的网格数据时，IDW是更优异的插值方法。据此，我们对当前主流观点提出了挑战：以往认为地统计OK方法始终是稳定同位素空间分析的最有效技术，同时也是构建同位素景观的默认方法。