Journal Paper Digests 2026 #1
- Upscaling Models for the Large-Scale Assessment of Soil Functions
- Autistic voices are an overlooked minority in geosciences
- Large contribution of antecedent climate to ecosystem productivity anomalies during extreme events
- Hydrologic connectivity as a predictor of degradation thresholds across semiarid sites with different vegetation patterns
- Spectra-based predictive mapping of soil erodibility and analysis of its influence mechanism: A typical case study for Northeast China
- Assessment and intercomparison of 23 global satellite and model-based soil moisture products using cosmic ray neutron sensing observations over Europe
- Global warming intensifies extreme day-to-day temperature changes in mid–low latitudes
- A Novel Hybrid Predictive Model Based on Mixture Density Networks With Weighted Conformal Inference Strategy for Runoff Interval Prediction Across Australia
A Novel Hybrid Predictive Model Based on Mixture Density Networks With Weighted Conformal Inference Strategy for Runoff Interval Prediction Across Australia
Abstract Accurate runoff forecasting helps mitigate flooding and drought risks and ensure water security under changing conditions. Compared to deterministic prediction models, interval prediction can more effectively quantify uncertainty, enhancing practical applicability. However, the Mixture Density Network (MDN) model—a state-of-the-art probabilistic modeling approach in hydrology—is susceptible to bias from distributional misspecification, and its prediction intervals are often overly wide, reducing practical utility. We therefore innovatively incorporated the Weighted Conformal Inference (WCI) strategy, which accounts for distributional shifts in runoff sequences, and integrated it with MDN to develop the WCI-MDN model for runoff interval prediction. To validate the effectiveness of the WCI strategy, we constructed six models in total: MDNs and WCI-MDNs under three distributions—Gaussian Mixture (GMM), Laplace Mixture (LMM), and Countable Mixtures of Asymmetric Laplacians (CMAL)—and evaluated their accuracy and robustness using data from 222 basins in the CAMELS-AUS data set. Results indicated that among the three MDN models, the LMM distribution achieved the best interval prediction performance, followed by the CMAL and GMM distributions. After introducing the WCI strategy, the coverage width-based criterion (CWC) for GMM, LMM, and CMAL distributions decreased by approximately 61.1%, 48.7%, and 54.3%, respectively, across all basins, demonstrating that the WCI-MDNs achieved higher prediction reliability. Furthermore, compared to the MDNs, the standard deviation of the CWC for the WCI-MDNs was reduced by 66.7%–81.8%, indicating higher robustness. Thus, the study improved the existing MDNs, providing a promising new approach for runoff interval prediction.
Global warming intensifies extreme day-to-day temperature changes in mid–low latitudes
Global warming is increasing the number and intensity of many extreme weather and climate events. Here we argue that extreme day-to-day temperature changes, exceeding the 90th percentile threshold of historical records, are an independent, but largely ignored, aspect of extreme weather events. Such extreme temperature changes have a stronger impact on human health in many locations than do diurnal temperature variations. Global observations show that such events have become more frequent since the 1960s in low and mid-latitudes but decreased at high latitudes, primarily due to GHG forcing. Climate models project a further amplification of extreme day-to-day temperature changes under warming, with frequency, amplitude and total intensity rising by ~17%, ~3% and ~20%, respectively, by 2100 in regions covering 80% of global population. Increased extreme day-to-day temperature changes are associated with drier soil and increased variability in pressure and soil moisture, posing substantial risks to societal and ecosystem resilience and adaptation.
Assessment and intercomparison of 23 global satellite and model-based soil moisture products using cosmic ray neutron sensing observations over Europe
Comprehensive evaluation of satellite and model-based soil moisture (SM) products is essential for their further development and application. With the advent of Cosmic Ray Neutron Sensing (CRNS), which has an observation radius of 130–240 m, the spatial representativeness mismatch between these grid-based SM products and ground single-point observations during the evaluation process can be feasibly relieved. In this study, we systematically evaluated 23 gridded SM products, including single-sensor satellite, multi-sensor merged, and model-based products, using 68 CRNS measurement sites across the Europe. Our evaluation revealed that the SMAP-INRAE-BORDEAUX (SMAP-IB) SM retrievals showed the superior consistency with CRNS measurements among all analyzed products, demonstrating both high correlation (R = 0.80) and low unbiased root mean square error (ubRMSE = 0.050 m3/m3). The CCI/C3S combined active-passive SM products ranked second in performance (R > 0.75, ubRMSE <0.060 m3/m3). In the bias analysis, 17 products had negative bias (−0.003 m3/m3 to −0.190 m3/m3) against CRNS measurements, while AMSR2-LPRM at C1 and C2 bands and CCI/C3S at active and passive products had positive bias (0.011 m3/m3 to 0.161 m3/m3). It was also found that the capabilities of all SM products retrievals degraded in terms of R and ubRMSE with increasing vegetation density, topographic complexity and soil wetness. Most products showed the lowest ubRMSE and highest R values in cropland compared to other land cover types. Our study emphasizes the substantial potential of cosmic field-scale SM observations for the validation of satellite- and model-based SM products, and our findings have the potential to advance algorithm refinement, product improvement, and hydrometeorological applications.
Spectra-based predictive mapping of soil erodibility and analysis of its influence mechanism: A typical case study for Northeast China
Soil erosion in Northeast China’s black soil region poses serious challenges to agricultural productivity and ecosystem sustainability. This study proposes a novel framework for high-resolution (10 m) mapping of soil erodibility by integrating Sentinel-2 spectral data with a gradient boosting decision tree (GBDT) model. A comprehensive soil erodibility index (CSEI) was developed to represent the combined effects of soil texture, structure, and organic stability. The GBDT model was used to identify the dominant environmental drivers and their nonlinear relationships with CSEI. Results indicate that the normalized difference tillage index (NDTI), soil moisture, and mean annual precipitation are the key influencing factors, collectively explaining 69.3 % of the spatial variability in soil erodibility. Threshold effects were observed, including an inverse S-curve for soil moisture and an inverted-U response to precipitation, reflecting shifts in erosion mechanisms under varying surface conditions. These findings provide quantitative evidence for targeted soil conservation and land-use optimization, supporting management strategies such as conservation tillage, slope-specific terracing, and vegetation restoration to mitigate erosion risks in vulnerable landscapes.
Hydrologic connectivity as a predictor of degradation thresholds across semiarid sites with different vegetation patterns
Dryland landscapes typically display a two-phase mosaic consisting of densely vegetated patches interspersed with low-cover or bare soil areas. The extent and spatial patterns of these patches have a direct effect on ecosystem function and disturbances, such as over grazing, can disrupt the original structure of vegetation and lead to degradation. This work investigates changes in the hydrologic connectivity (i.e., the degree to which areas of the landscape connect to each other) of Mulga landscapes induced by land degradation. Mulga is a keystone ecosystem of the Australian drylands and is characterised by a patchy vegetation cover, which can vary considerably from site to site. We analyse 31 plots with different degrees of degradation (or vegetation cover) in four Mulga sites with different precipitation, slope and vegetation and we quantify hydrologic connectivity combining high-resolution binary vegetation maps and DEMs. Results indicate that connectivity increases as vegetation cover decreases, but this relation is significantly non-linear with a clear threshold at 38 % vegetation cover below which connectivity (and loss of resources due to runoff out of the system) increases dramatically leading to degradation. A site with a pattern of vegetation strands concentrated along drainage lines showed consistently higher connectivity (due to longer connected paths) compared to the other sites where vegetation was more uniformly scattered or presented banded pattern perpendicular to drainage lines. Outputs from a vegetation thinning algorithm on patch edges consistent with grazing effects confirm the existence of the observed threshold in vegetation cover and the influence of vegetation patterns on connectivity. Our results indicate that connectivity is a strong indicator to detect degradation thresholds over a variety of vegetation arrangements typical of dryland systems.
Large contribution of antecedent climate to ecosystem productivity anomalies during extreme events
Ecosystems are not only affected by current climate but are also shaped by antecedent climate through their influences on vegetation growth and environmental conditions. These lagged responses, known as memory effects, can either exacerbate or mitigate the impacts of climate extremes on ecosystem functions. However, the direction, strength and influential duration of memory effects on ecosystem productivity remain poorly understood. Here we implement an interpretable machine-learning framework based on eddy covariance data to model ecosystem gross primary productivity over the period 1995–2020 and further investigate the characteristics of memory effects on positive and negative extremes of ecosystem productivity. Our results show a large contribution from antecedent climate conditions (38.2%) to ecosystem productivity during extremes, with precipitation accounting for 42.2% of the memory effects, followed by temperature (22.1%) and vapour pressure deficit (20.8%). Extreme events conditioned by long-term climatic variations often cause higher productivity losses than short-term extremes, with semi-arid ecosystems exhibiting the largest productivity anomalies and prolonged memory effects. Our results highlight the role of memory effects in regulating carbon flux variations and provide an observation-constrained benchmark for these effects.
Upscaling Models for the Large-Scale Assessment of Soil Functions
The characterization and assessment of soil functions is a prerequisite for agricultural and environmental policies aimed at soil health. However, there is a lack of satisfactory models for the assessment of soil functions supply to support national and intergovernmental initiatives. In this study we fill this gap by restructuring models developed to assess the multifunctionality of agricultural soils at the field scale. The multi-criteria decision models rely on soil properties, site characteristics and management information to assess the following five soil functions: (1) water regulation, (2) climate regulation, (3) nutrient cycling, (4) primary productivity and (5) provision of habitat for biodiversity. We develop models to assess soil functions supply at regional and national scales by adapting their structure to cope with the general lack of information on soil management at larger geographical scales. The restructured models are verified and a sensitivity analysis of the new model structure is performed. We further applied a comparison of the upscaled models with results from validated field-scale models using real data from 94 sites spanning across 13 European countries. We found that the upscaled models showed a similar sensitivity to the variability of the input data from the 94 sampling sites as the base models from which they were developed and that their overall supply is expected to be comparable. We describe the model structure of the upscaled models as well as their qualitative scales and integration rules. We propose the application of the models can serve for large-scale assessment of soil functions supply as part of soil health assessment for regional and national environmental and agricultural policies.
Autistic voices are an overlooked minority in geosciences
Autism remains an under-recognized and under-represented aspect of inclusivity conversations in geosciences. We highlight an urgent need for support and recognition of autistic learners, alongside a need to integrate autistic voices in learning and teaching practices.