Search for: All records

Lowest events in Lake Titicaca’s water level (LTWL) significantly impact local ecosystems and the drinking water supply in Peru and Bolivia. However, the hydroclimatic mechanisms driving extreme lake-level lowstands remain poorly understood. To investigate these low lake-level events, we analyzed detrended monthly LTWL anomalies, sea surface temperature (SST) datasets covering the period 1921–2023. ERA5 reanalysis covers the period 1940–2023. A multiple linear regression model was developed to compute detrended LTWL anomalies, excluding multidecadal and residual components. Interdecadal Pacific Oscillation (IPO) and Pacific Decadal Oscillation (PDO) indices were also analyzed for the same period. Results indicate that 25% of all LTWL minima events have a short duration of <5 months, while the remaining 75% of all events have a long duration of more than 9 months, respectively. All long-lived LTWL minima events are associated with reduced moisture flow from the Amazon basin toward Lake Titicaca, but the large-scale forcing varies with the phase change of the decadal component in the 11–15 years band of the PDO (PDO_{11–15 years}). Under warm PDO_{11–15 years}phases, LTWL minima are driven by an enhanced South American low-level jet (SALLJ) caused by warm SST anomalies over the eastern Pacific Ocean. Warm SST anomalies over tropical North Atlantic and central Pacific cold events, which reinforce the cold PDO_{11–15 years}phases, driving long-lived LTWL minima through the reduction of SALLJ. Conversely, long-lived LTWL minima events under neutral PDO_{11–15 years}phases are caused by westerly flow anomalies confined to the Peruvian Altiplano. Therefore, PDO and IPO do not drive long-lived LTWL minima events because their relationship does not remain consistent over time. In conclusion, long-lived LTWL minima events exhibit a regional nature and are not driven by the PDO or IPO, as LTWL shows no consistent relationship with these decadal SST modes over time. 

Deep Evidential Regression (DER) places a prior on the original Gaussian likelihood and treats learning as an evidence acquisition process to quantify uncertainty. For the validity of the evidence theory, DER requires specialized activation functions to ensure that the prior parameters remain non-negative. However, such constraints will trigger evidence contraction, causing sub-optimal performance. In this paper, we analyse DER theoretically, revealing the intrinsic limitations for sub-optimal performance: the non-negativity constraints on the Normal Inverse-Gamma (NIG) prior parameter trigger the evidence contraction under the specialized activation function, which hinders the optimization of DER performance. On this basis, we design a Non-saturating Uncertainty Regularization term, which effectively ensures that the performance is further optimized in the right direction. Experiments on real-world datasets show that our proposed approach improves the performance of DER while maintaining the ability to quantify uncertainty. 

Abstract The stable isotope ratio of dissolved inorganic carbon (δ¹³C‐DIC) is a valuable tracer for investigating carbon cycling in aquatic environments. However, its potential remains underutilized due to limited data availability. Fewer than 15% of cruise samples are analyzed forδ¹³C‐DIC, as isotope analysis using isotope ratio mass spectrometry is labor‐intensive and restricted to onshore laboratories. We present over 3500δ¹³C‐DIC measurements from the 2023 Global Ocean Ship‐based Hydrographic Investigations Program A16N cruise in the North Atlantic. Notably, three‐quarters of these measurements were conducted onboard using a CO₂extraction device coupled with cavity ring‐down spectroscopy, a more efficient and cost‐effective method. This extensive dataset providesδ¹³C‐DIC values with spatial resolution comparable to other ocean carbonate chemistry and biogeochemical parameters. This dataset supports improved quantification of anthropogenic CO₂uptake and storage, and may facilitate the development of algorithms to estimateδ¹³C‐DIC in under sampled regions.