• Energy Research
• 2025-2025close
• DE close
• FR close
• CN close

Predict+Optimize frameworks integrate forecasting and optimization to address real-world challenges such as renewable energy scheduling, where variability and uncertainty are critical factors. This paper benchmarks solutions from the IEEE-CIS Technical Challenge on Predict+Optimize for Renewable Energy Scheduling, focusing on forecasting renewable production and demand and optimizing energy cost. The competition attracted 49 participants in total. The top-ranked method employed stochastic optimization using LightGBM ensembles, and achieved at least a 2% reduction in energy costs compared to deterministic approaches, demonstrating that the most accurate point forecast does not necessarily guarantee the best performance in downstream optimization. The published data and problem setting establish a benchmark for further research into integrated forecasting-optimization methods for energy systems, highlighting the importance of considering forecast uncertainty in optimization models to achieve cost-effective and reliable energy management. The novelty of this work lies in its comprehensive evaluation of Predict+Optimize methodologies applied to a real-world renewable energy scheduling problem, providing insights into the scalability, generalizability, and effectiveness of the proposed solutions. Potential applications extend beyond energy systems to any domain requiring integrated forecasting and optimization, such as supply chain management, transportation planning, and financial portfolio optimization.

Tropical forests are of great ecological and climatological importance. Although they only cover about 6% of Earth’s surface, they are home to approx. 50% of the world’s animal and plant species. Their trees store 50% more carbon than trees outside the tropics. At the same time, they are one of the most endangered ecosystems on Earth: about 6 million of hectares per year are felled for timber or cleared for farming. Compared to the other components of the carbon cycle (i.e. the ocean as a sink and the burning of fossil fuels as a source), the uncertainties in the local land carbon stocks and the carbon fluxes are particularly large. This is especially true for tropical forests: more than 98% of the carbon flux generated by changes in land-use may be due to tropical deforestation, which converts carbon stored as biomass into emissions.In this context, the AfriSAR 2015/16 campaign, supported by ESA, was carried out over four forest sites in Gabon by ONERA (July 2015) during the dry season and by DLR (February 2016) during the wet season. From the data collected the innovative techniques applied to estimate forest height and biomass could be improved significantly and are summarized in a special issue ‘Forest Structure Estimation in Remote Sensing’ of IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.The motivation of the AfriSAR campaign was to acquire demonstration data for the soon to be launched ESA BIOMASS mission, that was selected as the 7th Earth Explorer mission in May 2013 in order to meet the pressing need for information on tropical carbon sinks and sources by providing estimates of forest height and biomass. AfriSAR focused on African tropical and savannah forest types (with biomass in the 100-300 t/ha range) and complements previous ESA campaigns over Indonesian and Amazonian forest types in 2004 (INDREX-II) and 2009 (TropiSAR).The present contribution concerns the GABONX campaign, the ESA supported successor to AfriSAR, which took place in May to July 2023. GABONX aims to detect and quantify changes that have occurred since the DLR acquisitions in February 2016. To this end, DLR’s F-SAR sensor acquired interferometric stacks of fully polarimetric L- and P-Band data over the same forest sites in the same flight geometry as in 2016. The results presented give an overview of campaign activities with particular emphasis on the calibration of the SAR instrument as well as the validation of forest parameters derived from polarimetric interferometry. The SAR sensor calibration is based on an innovative approach that leverages state-of-the-art EM simulation to accurately characterize the 5m trihedral reference target deployed for the campaign in Gabon. The validation of derived forest parameters uses lidar measurements obtained in the time frame of the GABONX campaign by NASA’s LVIS sensor. As an outlook, further collaborative calibration and validation activities will hopefully include the cross-calibration of DLR’s F-SAR and NASA’s UAVSAR, which is set to acquire L- and P-Band data over the GABONX sites in 2024.

DeepOWT (deep learning derived global offshore wind turbines) is an independent and openly accessible data set of offshore wind energy infrastructure locations and their temporal deployment dynamics on a global scale. It is derived by applying deep learning based object detection on ESA's spaceborne Sentinel-1 synthetic aperture radar (SAR) archive. DeepOWT provides OWT locations along with their quarterly deployment stages from 2016Q1 until 2025Q1. It differentiates between platforms under construction, OWTs which are readily deployed and offshore wind farm substations, such as transformer stations.The dataset continues the work of 10.5194/essd-14-4251-2022. File metadata File Time Geometry Spatial extent DeepOWT.geojson (Dataset) 2016Q1-2025Q1 points Global gt_2021Q2_nsb.geojson (Ground Truth Location) 2021Q2 polygons North Sea Basin gt_2021Q2_ecs.geojson (Ground Truth Location) 2021Q2 polygons East China Sea gt_2021Q2_vtn.geojson (Ground Truth Location) 2021Q2 polygons Southeast Vietnamese Coast gt_nsb_gridded.geojson (Ground Truth Region) - polygon North Sea Basin gt_ecs_gridded.geojson (Ground Truth Region) - polygon East China Sea gt_ecs_gridded.geojson (Ground Truth Region) - polygon Southeast Vietnamese Coast Used semantic label open sea under construction offshore wind turbine offshore wind farm substation

Based on the findings of present investigation, it can be inferred that physico-chemical properties and microbial population in soil was also varied with land use whereas the physico-chemical properties and microbial population were generally improved in kharif as compared to rabi. Plantation land use was the best land use system followed by forest land use system for sustainable improvement of soil health may be recommended in the eastern region of Uttar Pradesh or similar agro ecoregions.

Abstract Photovoltaic (PV) energy grows rapidly and is crucial for the decarbonization of electric systems. However, centralized registries recording the technical characteristics of rooftop PV systems are often missing, making it difficult to monitor this growth accurately. The lack of monitoring could threaten the integration of PV energy into the grid. To avoid this situation, remote sensing of rooftop PV systems using deep learning has emerged as a promising solution. However, existing techniques are not reliable enough to be used by public authorities or transmission system operators (TSOs) to construct up-to-date statistics on the rooftop PV fleet. The lack of reliability comes from deep learning models being sensitive to distribution shifts. This work comprehensively evaluates distribution shifts’ effects on the classification accuracy of deep learning models trained to detect rooftop PV panels on overhead imagery. We construct a benchmark to isolate the sources of distribution shifts and introduce a novel methodology that leverages explainable artificial intelligence (XAI) and decomposition of the input image and model’s decision regarding scales to understand how distribution shifts affect deep learning models. Finally, based on our analysis, we introduce a data augmentation technique designed to improve the robustness of deep learning classifiers under varying acquisition conditions. Our proposed approach outperforms competing methods and can close the gap with more demanding unsupervised domain adaptation methods. We discuss practical recommendations for mapping PV systems using overhead imagery and deep learning models.

Improving our comprehension of infiltration processes in karst systems is crucial for a better adaptation to the global change regarding water resources availability and management. In this work, the effective recharge under different meteorological conditions and its transfer along the vertically distributed compartments of a geologically complex karst aquifer in southern Spain have been evaluated. Continuous records of soil moisture and temperature values (at 5 and 10 cm depth and the soil-rock transition -average depth of 28 cm-) have been combined with hourly hydrodynamic and hydrothermal responses recorded at two springs with a marked influence of the unsaturated zone (UZ) and the saturated zone (SZ), respectively.Most rainfalls generate soil moisture signal in the shallowest probes. However, a mean increase of soil water content of 10.5% in summer (from background values of 2.5%) and 6.1% in autumn-winter (from 9.6%) at the soil-rock interface were needed to produce hydrodynamic responses in the system: first in the spring related to the UZ, with a time delay of 4-9 hours after moisture peaks, and then (14-18 hours) in the spring draining the SZ, but only during autumn-winter recharge events. In addition, recharge caused decreases (up to 0.9°C) in the temperature of the water drained by the first spring, while lagged rises (up to 0.6°C) occurred in the second outlet.Transmission of the input signal would be favoured by stronger karstification, but the presence of inter-bedded detrital formations in the lithological sequence of the aquifer (partially confined in the SE border) filter and buffer groundwater flows before being drained by the spring related to the SZ. These findings will help to assess thresholds for effective infiltration and to predict groundwater recharge in karst aquifers under different climate change scenarios.

Sub-Saharan Africa (SSA) faces significant food security risks, primarily due to low soil fertility leading to low crop yields. Climate change is expected to worsen food security issues in SSA due to a combined negative impact on crop yield and soil fertility. A common omission from climate change impact studies in SSA is the interaction between change in soil fertility and crop yield. Integrated soil fertility management (ISFM), which includes the combined use of mineral and organic fertilizers, is expected to increase crop yield but it is uncertain how this advantage is maintained with climate change.   We explored the impact of scenarios of change in soil fertility and climate variables (temperature, rainfall, and CO2) on rainfed maize yield in four representative sites in SSA with no input and ISFM management. To do so, we used an ensemble of 15 calibrated soil-crop models. Reset and continuous simulations were performed to assess the impact of soil fertility vs climate change on crop yield. In reset simulations, SOC, soil N and soil water were reinitialized each year with the same initial conditions. In continuous simulations, SOC, soil N and soil water values of a given year were obtained from the simulation of the previous year, allowing cumulative effects on SOC and crop yields.Most models agreed that with current baseline (no input) management, yield changed by a much larger order of magnitude when considering declining soil fertility with baseline climate (-39%), compared with considering constant soil fertility but changes in temperature, rainfall and CO2 (from -12% to +5% depending on the climate variable considered). The interaction between change in soil fertility and climate variables only marginally influenced maize yield (high agreement between models). The model ensemble indicated that when accounting for soil fertility change, the benefits of ISFM systems over no-input systems increased over time (+190%). This increase in ISFM benefits was greater in sites with low initial soil fertility. We advocate for the urgent need to account for soil-crop long-term feedback in climate change studies to avoid large underestimations of climate change and ISFM impact on food production in SSA.

Precise and reliable climate projections are required for climate adaptation and mitigation, but Earth system models still exhibit great uncertainties. Several approaches have been developed to reduce the spread of climate projections and feedbacks, yet those methods cannot capture the nonlinear complexity inherent in the climate system. Using a Transfer Learning approach, we show that Machine Learning can be used to optimally leverage and merge the knowledge gained from global temperature maps simulated by Earth system models and observed in the historical period to reduce the spread of global surface air temperature fields projected in the 21st century. We reach an uncertainty reduction of more than 50% with respect to state-of-the-art approaches while giving evidence that our method provides improved regional temperature patterns together with narrower projections uncertainty, urgently required for climate adaptation.

Eastern Boundary Upwelling Systems (EBUS) are characterised by wind-triggered upwelling of deep waters along the coast. They are hotspots of biological productivity and therefore have a high economic, ecological and social importance. Here we investigate the evolution of the two Atlantic EBUS during the historical period and in a future high-emission scenario in CMIP6 models from two modelling centres, with spatial resolutions ranging from 1° to 1/12° in the ocean. The decomposition of the upwelling systems into subregions reveals differences between the equatorward and poleward parts. Our analysis is focused on the modelled vertical transport, which is shown to be consistent with the wind-derived Ekman index. Integrating the vertical transport provides a synthetic view of the upwelling cells, their strength, depth and distance to the coast. The models show high interannual variability over the 21st century century, which explains why significant trends could only be found in few subregions of the Atlantic EBUS. The results suggest a poleward migration of upwelling systems with climate change and a change of the upwelling cells, rather than the uniform intensification which had been hypothesised by Bakun in 1990.

EMSO-Western Ligurian Site (European Multidisciplinary See floor Observatory and water column, Western Ligurian Site) is a second generation permanent submarine observatory deployed offshore of Toulon, France, as a follow up of the pioneering ANTARES neutrino telescope located nearby. This submarine network is part of KM3NeT (https://www.km3net.org/) which has a modular topology designed to connect up to 120 neutrino detection units, i.e. ten times more than ANTARES. The Earth and Sea Science (ESS) instrumentation connected to KM3NeT is based on two complementary components: an Instrumented Interface Module (MII) and an autonomous mooring line (ALBATROSS). The Module Interface Instrumented "MII" was deployed in May 2019 and cabled to the MEUST Node#1. Node#1 is cabled to the shore via the KM3NeT cable. The MII provides data for temperature, conductivity/salinity, pressure, particles proxy (deduced from beam attenuation measured with a CSTAR transmissiometer). The MII collects data through an acoustic link from the instrumented mooring line ALBATROSS (https://doi.org/10.17882/74513) deployed at a distance of 2-3 kilometers. These data are being transferred daily for near real time visualisation.