GEGilles Erkens

Gilles Erkens

*This is an algorithmically generated author record based on publicly available information.

Gilles Erkens has a total of 67 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Machine Learning-Based Hazard and Risk Assessment of Subsidence-Related CO2 Emissions in Dutch Peatlands

    by Muhannad Hammad, Kim Cohen, Gilles Erkens, Nicholas Hartono, Esther Stouthamer

    Dutch peatland drainage for intensive dairy farming purposes causes land subsidence and releases massive amounts of soil carbon as CO2, confronting policymakers with the challenge of sustaining dairy farming while fulfilling the reduction commitments of land subsidence and CO2 emissions under the Dutch Climate Act and the Paris Climate Agreement. This study presents a machine learning-based framework that allows predicting future land subsidence and assessing associated soil carbon loss and CO2 emission risk in Dutch peatlands, using Krimpenerwaard municipality as a case study. A Random Forest model was trained on irreversible subsidence data derived from the still-under-refinement Simple PArameterization for the Motion of Soils (SPAMS10 - Version1) empirical model for the 2016-2022 period, together with six geospatial predictors representing surface and subsurface conditions on a 5 m grid pixel scale. The trained model achieved good testing performance (R² = 0.903, RMSE = 0.173 mm/year). Feature importance analysis revealed that the most important controllable predictor is the soil moisture index. Land subsidence data were converted to hazard (oxidized volume loss) and integrated with vulnerability (soil organic carbon stock) and exposure (peat presence) within a normalized hazard-vulnerability-exposure framework to produce the 2016-2022 risk index map within a GIS environment. The resulting risk index map identifies managed grassland parcels near Stolwijk as priority areas for intervention. This framework enables future scenario testing by modifying controllable predictors to evaluate the effectiveness of potential interventions. This study provides a reproducible decision-support tool for identifying high-risk parcels and guiding sustainable mitigation of land subsidence across Dutch peatlands.

  • Open Access

    Assessing the viscous compression parameter for land subsidence modelling from material characteristics

    by Pepijn Van Elderen, Gilles Erkens, Cor Zwanenburg, Harry Van Essen, Esther Stouthamer

    Compaction of soft subsurface layers occurs through different mechanisms, namely con- solidation, viscous compression, autocompaction, and shrinkage. Viscous compression is a key long-term process often the most significant component of compaction over long time periods. The rate of viscous compression over time is defined by the viscous compression parameter (Cα), which describes the amount of strain per logarithmic unit of time with constant effective stress. Overall compressibility of different material can often be linked to their material characteristics, yet for viscous compression this is still problematic. Understanding the relationship between material characteristics and Cα is essential for improving compaction predictions in soft subsurface layers such as clay and peat. This study investigates how Cα varies with subsurface material characteristics and explores its predictability based on specific material characteristic values. An extensive dataset of one-dimensional compression tests on Holocene peat and clay samples from the Rhine-Meuse delta and coastal zone in the Netherlands was analysed. Using principal component and clustering analyses, four distinct lithological classes were identified and labelled based on organic fraction: Clay, Clayey peat, fibrous Peat & decomposed Peat. All four lithological classes show increasing median sample Cα values with decreasing bulk densities and increasing water content and void ratio, with Cα val- ues ranging from 0.0074 in the ‘Clay’ material type to 0.0293–0.0384 in the two ‘Peat’ ma- terial classes. Compared to currently applied traditional lithological classifications and default model values in the Netherlands, the material type-specific Cα values provide a more accurate and data-driven alternative. Regression analysis using the full dataset, with water content and dry bulk density as the predictors, yielded a better prediction of Cα (R² = 0.61) than for regressions per material type. Due to large variability within the peaty material types, upper and lower Cα bounds (0.02–0.06) are recommended in practi- cal applications for ‘Clayey Peat’ and ‘Peat’ subsurface layers, as these values are consis- tently observed across the full range of water content and dry bulk density within these clusters. While the Cα values for ‘Clay’ are similar to standard values (0.0035–0.0123), Cα values for ‘Clayey Peat’ and ‘Peat’ are up to three times higher than the standard values, which would result in twice or thrice the amount of viscous compression over the same time period. Finally, we emphasise the need for adaptable Cα values in models to reflect evolving material characteristics over time, driven by decomposition.

  • Open Access

    Iron and sulphur regulate carbon dioxide emissions in drained coastal peatlands of The Netherlands

    by Duygu Tolunay, Gizem Ergut, Levi Simon, Gilles Erkens, George A. Kowalchuk, Mariet M. Hefting

    Fluctuating groundwater levels in drained peatlands create a transition zone with seasonally changing oxygen availability. This zone drives dynamic iron (Fe) and sulphur (S) cycling under alternating anoxic and oxic conditions, influencing decomposition rates. This study investigated how Fe and S affect decomposition rates and resulting carbon dioxide (CO 2 ) emissions under fluctuating redox conditions in transition zone. In a controlled laboratory experiment, peat samples from two drained Dutch coastal peatlands were amended with ferric iron (Fe 3+ ) and sulphate (SO 4 2− ) and incubated anoxically to mimic high groundwater tables. This was followed by an oxic phase simulating groundwater table drops. The cycle was repeated with lactate addition to replenish labile carbon. Carbon dioxide emission rates were monitored continuously throughout the anoxic–oxic cycles. Water soluble Fe and S concentrations, exoenzyme activities, and pH were measured before and after the experiment. Carbon dioxide emission rates increased under anoxic conditions with Fe 3+ and SO 4 2− amendments potentially due to stimulation of microbial activity using these compounds as alternative electron acceptors. Short-term oxygenation suppressed emissions compared to controls without amendments. Water-soluble Fe remained stable across treatments, while water-soluble S concentrations changed significantly from initial levels. Exoenzyme activities were primarily influenced by pH, with minimal effects from amendments. The findings show that transition zone is an active redox zone where decomposition dynamics are determined by available electron acceptors in the system, influencing greenhouse gas (GHG) emissions from managed peatlands. This zone should be integrated into future models to improve the accuracy of reporting national GHG emissions.


  • Open Access

    GotNL_Chapter23_GIS.zip

    by Peter Fokker, Thibault Candela, Gilles Erkens, R.F. Hanssen, Henk Kooi, Kay Koster, Freek van Leijen

    GIS files belonging to the figures of Chapter 23 - Subsidence, in Geology of the Netherlands (second edition).

  • Open Access

    GotNL_Chapter23_GIS.zip

    by Peter Fokker, Thibault Candela, Gilles Erkens, R.F. Hanssen, Henk Kooi, Kay Koster, Freek van Leijen

    GIS files belonging to the figures of Chapter 23 - Subsidence, in Geology of the Netherlands (second edition).

  • Open Access

    Shallow-subsidence vulnerability in the city of New Orleans, southern USA

    by Sanneke van Asselen, Gilles Erkens, Molly E. Keogh, Roelof Stuurman

    Land subsidence in the city of New Orleans (USA) and its surroundings increases flood risk, and may cause damage to buildings and infrastructure and loss of protective coastal wetlands. To make New Orleans more resilient to future flooding, a new approach for groundwater and subsidence management is needed. As a first step in developing such an approach, high-quality and high-resolution subsurface and groundwater information was collected and synthesized to better understand and quantify shallow land subsidence in New Orleans. Based on the collected field data, it was found that especially the low-lying areas north and south of the Metairie-Gentilly (MG) Ridge are most vulnerable to further subsidence; north of the MG Ridge, subsidence is mainly caused by peat oxidation and south of the MG Ridge mainly by peat compaction. At present, peat has compacted ~31% on average, with a range of 9–62%, leaving significant potential for further subsidence due to peat compaction. Phreatic groundwater levels drop to ~150 cm below surface levels during dry periods and increase to ~50 cm below surface during wet periods, on average. Present phreatic groundwater levels are mostly controlled by leaking subsurface pipes. Shallow groundwater in the northern part of New Orleans is threatened by salinization resulting from a reversal of groundwater flow following past subsidence, which may increase in the future due to sea-level rise and continued subsidence. The hydrogeologic information provided here is needed to effectively design tailor-made measures to limit urban flooding and continued subsidence in the city of New Orleans.


  • Open Access

    Probabilistic Estimation of InSAR Displacement Phase Guided by Contextual Information and Artificial Intelligence

    by Philip Conroy, Simon A. N. van Diepen, Sanneke Van Asselen, Gilles Erkens, Freek J. van Leijen, Ramon F. Hanssen

    Phase unwrapping, also known as ambiguity resolution, is an underdetermined problem in which assumptions must be made to obtain a result in SAR interferometry (InSAR) time series analysis. This problem is particularly acute for distributed scatterer InSAR, in which noise levels can be so large that they are comparable in magnitude to the signal of investigation. In addition, deformation rates can be highly nonlinear and orders of magnitude larger than neighboring point scatterers, which may be part of a more stable object. The combination of these factors has often proven too challenging for the conventional InSAR processing methods to successfully monitor these regions. We present a methodology which allows for additional environmental information to be integrated into the phase unwrapping procedure, thereby alleviating the problems described above. We show how problematic epochs that cause errors in the temporal phase unwrapping process can be anticipated by the machine learning algorithms which can create categorical predictions about the relative ambiguity level based on the readily available meteorological data. These predictions significantly assist in the interpretation of large changes in the wrapped interferometric phase and enable the monitoring of environments not previously possible using standard minimum gradient phase unwrapping techniques.


  • Open Access

    SOMERS: Monitoring greenhouse gas emission from the Dutch peatland meadows on parcel level

    by Gilles Erkens, Roel Melman, Siem Jansen, Jim Boonman, Ype van der Velde, Mariet Hefting, Joost Keuskamp, Merit van Den Berg, Jan van Den Akker, Christian Fritz, et al.

    Following the Paris Agreement (2015) that aims to limit climate warming, the Dutch government presented a National Climate Agreement in 2019. This agreement stated the overall ambition of reducing the national greenhouse gas emission by 49% in 2030 (compared to 1990) and allocates this reduction target to different sectors, such as industry, mobility, agriculture or land use. Within the latter sector, the peatland meadows are currently estimated to contribute ~4.6 to 7 Mton per year of CO2 to the national Dutch greenhouse gas emission. In the National Climate Agreement, the aim is to reduce the net CO2 emission from the peatland meadows with 1 Mton per year by 2030. To comply with the greenhouse gas emission reduction targets for peatlands, a set of measures that raise groundwater levels are currently being proposed and tested in pilots. The Dutch National Research Programme on Greenhouse Gas Emissions from Peat Meadows (NOBV) investigates the effects of the proposed measures on the greenhouse gas emission balance under different environmental conditions. In the National Climate Agreement, it was decided that annual progress made in reducing greenhouse gas emissions needs to be monitored. The NOBV consortium is developing a registration system for this monitoring and presents it current status and ideas for future development in this contribution.The registration system SOMERS (Subsurface Organic Matter Emission Registration System) is based on a multi-model ensemble approach. Using numerical models that simulate groundwater and carbon dynamics, the CO2 emission as a result of peat decomposition is calculated. Within SOMERS, existing models are supplemented by two newly developed models for assessing groundwater dynamics and peat decomposition, that require limited data input and have a short runtime. The new models simulate at parcel resolution and together are used to make a multi-model ensemble estimate of annual, national peatland greenhouse gas emissions since 2016 (the reference year). The new models are tested with annual carbon flux estimates. In the long run, we envisage to fully couple the modelling approach with the automated field measurements that are being collected in a new national measurement network.In this contribution, SOMERS will be introduced, and the calibration and validation approach will be discussed. We present predictions, under idealized average weather conditions, to establish effects of proposed mitigation measures. This directly serves policy development in regional spatial plans for the Dutch peatland meadows. Lastly, a first national peatland CO2 emission budget based on SOMERS is presented, which after some further development may support LULUCF-sector reporting in the Netherlands.


  • Open Access

    Organic Matter Accretion, Shallow Subsidence, and River Delta Sustainability

    by Molly E. Keogh, Torbjörn E. Törnqvist, Alexander S. Kolker, Gilles Erkens, Jonathan G. Bridgeman

    Globally, mineral sediment supply to deltaic wetlands has generally decreased so these wetlands increasingly rely on accretion of organic matter to keep pace with relative sea‐level rise (RSLR). Because organic‐rich sediments tend to be more compressible than mineral‐dominated sediments, deltaic wetland strata are vulnerable to compaction and drowning. Using an unprecedented data set of almost 3,000 discrete bulk density and organic‐matter measurements, we examine organic‐rich facies from coastal Louisiana to quantify the thickness lost to compaction and investigate whether sediments are able to maintain sufficient volume for the associated wetlands to keep pace with RSLR. We find that organic content as well as overburden thickness and density (which together determine effective stress) strongly control sediment compaction. Most compaction occurs in the top 1–3 m and within the first 100–500 years after deposition. In settings with thick peat beds, successions up to 14 m thick have been compacted by up to ∼50%. We apply geotechnical modeling to examine the balance between elevation gained from accretion and elevation lost to compaction due to renewed sediment deposition over a 100‐year timescale. Wetlands overlying mineral‐dominated lithologies may support the weight of deposition and allow net elevation gain. Model results show that reintroduction of sediment to a representative Mississippi Delta wetland site will likely cause another ∼0.35–1.14 m of compaction but leave a net elevation gain of ∼0.01–1.75 m, depending on the sediment delivery rate and stiffness of underlying strata.


  • Open Access

    Altered surface hydrology as a potential mechanism for subsidence in coastal Louisiana

    by Jaap H. Nienhuis, Torbjörn E. Törnqvist, Gilles Erkens

    The natural wetlands of coastal Louisiana are experiencing rapid subsidence rates averaging 9±1 mm yr−1. Recent measurements based on GPS data and CRMS surface elevation tables (SETs) have shown that most of the subsidence is shallow and occurs in the uppermost 5 meters. Sources of subsidence and the origin of their spatial variability are strongly debated. Here we use CRMS SETs together with historic maps of coastal Louisiana to explore two hypotheses: (i) shallow subsidence is a result of accommodation created by (long-term) deep subsidence processes and self-weight consolidation, and (ii) changes in marsh hydrology (groundwater and surface water flows) have led to a recent increase in shallow subsidence. First, we find that, although self-weight consolidation would result in generally high observed shallow subsidence rates, it does not explain the rates nor the spatial variability of the CRMS SET data. Second, based on historic maps, we find that shallow subsidence rates are significantly higher for CRMS sites where shipping canals have reduced their distance to the marsh edge. This is potentially a result from increased sediment deposition, but CRMS data also show altered groundwater levels near the marsh edge. We find some indication that prolonged periods of low water could have led to increases in effective stresses that explain some of the rapid rates of shallow subsidence observed along Louisiana's coastline.