Homepage Thomas Opitz

Homepage of Thomas Opitz

I am a research associate (CR HDR) at the Biostatistics and Spatial processes lab of INRAE in Avignon.

Mail: thomas POINT opitz AT inrae POINT fr
Telephone: 04 32 72 21 86
Address: INRAE-BioSP – Domaine St. Paul – 228, route de l'Aérodrome – 84914 Avignon – France

General research interests

  • My research concerns spatio-temporal modeling and prediction of environmental, climatological, ecological and epidemiological processes and risks.
  • I develop and implement theoretical and statistical tools  at the inferface of Extreme-Value Theory, which provides a framework for predicting probabilities of events with very extreme magnitudes, and stochastic geometry, useful for studying geometric patterns in occurrence locations/times and in agricultural landscapes.
  • For inference in stochastic models, a blend of frequentist and Bayesian inference techniques is used, with a particular focus on composite likelihood techniques and the integrated nested Laplace approximation (INLA).
  • Statistical models are constructed to combine data available at multiple spatial and temporal scales (e.g., raster, lattice, irregularly spaced locations) and from multiple sources (validated data with strict observation / transformation protocol, citizen science programs, etc.), often large georeferenced datasets.

Current research topics

  1. Space-time modeling of extremes in climate and weatherdata:
    • New theory and inference tools for models with flexible joint tail decay rates, often based on scale or location mixture representation.
    • Stochastic geometry theory and tools for spatiotemporal extremes
    • Modeling  temporal trends and dependence in spatiotemporally indexed observations.
    • Semiparametric resampling techniques for spatial and spatio-temporal extremes.
    • Applications to meteorological and climatic processes (precipitation, wind speed, temperature, air pollution), with a view towards impacts of climate change.
  2. Bayesian modeling of wildfires and landslides using marked log-Gaussian Cox processes.
  3. Stochastic simulation for agricultural landscapes using stochastic geometry tools, coupled with population dynamics models, and sensitivity analysis of agroecological outputs.
  4. Spatial and spatiotemporal modeling of ecological processes (Asian hornet invasion and efficiency of capturing them; wolf attacks on sheep herds; opportunistic data for species distribution modeling).
  5. (Latent) Lévy convolutions: novel spatial-temporal models for non Gaussian data (extremes, count data) based on kernel-smoothed infinitely divisible distributions, with a focus on gamma processes.

Distinction

Young Researcher Award of INRAE in 2020 ("Laurier Espoir Scientifique"): General presentation and my portrait

Preprints

  • Allard, D., Clarotto, L., Opitz, T., Romary, T. Discussion on “Competition on Spatial Statistics for Large Datasets".
  • Koh, J., Pimont, F., Dupuy, J.-L., Opitz. T. Spatiotemporal wildfire modeling through point processes with moderate and extreme marks. arXiv preprint.
  • Zhang, Z., Huser, R., Opitz, T., Wadsworth, J. L. Modeling Spatial Extremes Using Normal Mean-Variance Mixtures.
  • Zamberletti, P., Papaïx, J., Gabriel, E., Opitz, T. Spatio-temporal point processes as meta-models for population dynamics in heterogeneous  landscapes.
  • Zamberletti, P., Sabir, K., Opitz, T.,  Bonnefon, O., Gabriel, E., Papaïx, J. More pests but less treatments: ambivalent effect of landscape complexity on Conservation Biological Control.
  • Zhong, P., Huser, R. and Opitz, T. Exact Simulation of Max-Infinitely Divisible Processes. arXiv preprint arXiv:2103.00533.
  • Allard, D.,  Hristopoulos, D. and Opitz, T. Linking Physics and Spatial Statistics: A New Family of Boltzmann-Gibbs Random Fields.
  • Simpson, E., Opitz, T. and Wadsworth J. L. High-dimensional modeling of spatial and spatio-temporal conditional extremes using INLA and the SPDE approach. [Link to arXiv preprint]
  • Opitz, T. et al. High-resolution Bayesian mapping of landslide hazard with unobserved trigger event. [Link to arXiv preprint]
  • Opitz, T.  Spatial random field models based on Lévy indicator convolutions, arXiv preprint arXiv:1710.06826.

Publications

  • Zhong, P., Huser, R. and Opitz, T. (2021+) Modeling Non-Stationary Temperature Maxima Based on Extremal Dependence Changing with Event Magnitude. Accepted for Annals of Applied Statistics.
  • Zamberletti, P., Papaïx, J., Gabriel, E., Opitz, T. (2021) Landscape allocation: stochastic generators and statistical inference. Accepted for Annals of Applied Statistics.Link to arXiv preprint.
  • Yadav, R., Opitz, T. and Huser, R. (2021)  Spatial hierarchical modeling of threshold exceedances using rate mixtures. Environmetrics.
  • Pimont, F. et al. (2021) Prediction of regional wildfire activity with a probabilistic Bayesian framework. Ecological applications.
  • Grente, O. et al. (2020) Tirs dérogatoires de loups en France : état des connaissances et des enjeux pour la gestion des attaques aux troupeaux. Faune Sauvage.
  • Palacios-Rodriguez, F. et al. (2020) Semi-parametric generalized Pareto processes for simulating space-time extreme events. To appear in Stochastic Environmental Research and Risk Assessment.
  • Castro-Camilo, D., Mhalla, L. and Opitz, T. Bayesian space-time gap filling for inference on hot spots: an application to Red Sea surface temperatures. Extremes. Link to arXiv preprint.
  • Huser, R., Opitz, T. and Thibaud, E. (2020) Max-infinitely divisible models and inference for spatial extremes. Scandinavian Journal of Statistics. Link to arXiv preprint.
  • Lombardo, L. et al. (2021) Space-Time Landslide Predictive Modelling. Earth Science Reviews. Link to arXiv preprint.
  • Opitz, T., Allard, D. and Mariethoz, G. (2020) Semi-parametric resampling with extremes. Spatial Statistics. doi: 10.1016/j.spasta.2020.100445.
  • Opitz, T., Bonneu, F. and Gabriel, E. (2020) Point-process based modeling of space-time structures of forest fire occurrences in Mediterranean France’, Spatial Statistics, In press. doi: 10.1016/j.spasta.2020.100429.
  • Bacro, J.-N. et al. (2019) ‘Hierarchical Space-Time Modeling of Asymptotically Independent Exceedances With an Application to Precipitation Data’, Journal of the American Statistical Association. Taylor & Francis, 0(0), pp. 1–26. doi: 10.1080/01621459.2019.1617152.
  • Engelke, S., Opitz, T. and Wadsworth, J. L. (2019) ‘Extremal dependence of random scale constructions’, Extremes.
  • Lombardo, L., Opitz, T. and Huser, R. (2019) ‘Numerical Recipes for Landslide Spatial Prediction Using R-INLA: A Step-by-Step Tutorial’, in Spatial Modeling in GIS and R for Earth and Environmental Sciences. Elsevier, pp. 55–83.
  • Mhalla, L., Opitz, T. and Chavez-Demoulin, V. (2019) ‘Exceedance-based nonlinear regression of tail dependence’, Extremes. Springer, pp. 1–30.
  • Bakka, H. et al. (2018) ‘Discussion of ``Using Stacking to Average Bayesian Predictive Distributions" by Yao et. al’, Bayesian Analysis.
  • Fargeon, H. et al. (2018) ‘Assessing the increase in wildfire occurrence with climate change and the uncertainties associated with this projection’, in 8th International conference on forest fire research.
  • Lombardo, L., Opitz, T. and Huser, R. (2018) ‘Point process-based modeling of multiple debris flow landslides using INLA: an application to the 2009 Messina disaster’, Stochastic environmental research and risk assessment. Springer, 32(7), pp. 2179–2198.
  • Opitz, T. et al. (2018) ‘INLA goes extreme: Bayesian tail regression for the estimation of high spatio-temporal quantiles’, Extremes. Springer, 21(3), pp. 441–462.
  • Tapi Nzali, M. D. et al. (2018) ‘Reconciliation of patient/doctor vocabulary in a structured resource’, Health Informatics journal. SAGE Publications Sage UK: London, England.
  • Gabriel, E., Opitz, T. and Bonneu, F. (2017) ‘Detecting and modeling multi-scale space-time structures: the case of wildfire occurrences’, Journal of the French Statistical Society (Special Issue on Space-Time Statistics).
  • Huser, R., Opitz, T. and Thibaud, E. (2017) ‘Bridging asymptotic independence and dependence in spatial extremes using Gaussian scale mixtures’, Spatial Statistics. Elsevier, 21, pp. 166–186.
  • Mornet, A. et al. (2017) ‘Wind storm risk management: sensitivity of return period calculations and spread on the territory’, Stochastic Environmental Research and Risk Assessment. Springer, 31(8), pp. 1977–1995.
  • Nzali, M. D. T. et al. (2017) ‘What patients can tell us: topic analysis for social media on breast cancer’, JMIR Medical Informatics. JMIR Publications Inc., 5(3).
  • Opitz, T. (2017) ‘Latent Gaussian modeling and INLA: A review with focus on space-time applications’, Journal of the French Statistical Society (Special Issue on Space-Time Statistics), 158(3).
  • Opitz, T. (2016) ‘Modeling asymptotically independent spatial extremes based on Laplace random fields’, Spatial Statistics, 16, pp. 1–18.
  • RESSTE network (2017). Analyzing spatio-temporal data with R: everything you always wanted to know-but were afraid to ask. Journal of the French Statistical Society (Special Issue on Space-Time Statistics), 158(3).
  • Mornet, A. et al. (2015) ‘Index for Predicting Insurance Claims from Wind Storms with an Application in France’, Risk Analysis. Wiley Online Library, 35(11), pp. 2029–2056.
  • Opitz, T., Bacro, J.-N. and Ribereau, P. (2015) ‘The spectrogram: A threshold-based inferential tool for extremes of stochastic processes’, Electronic Journal of Statistics. Institute of Mathematical Statistics, 9(1), pp. 842–868.
  • Tapi Nzali, M. D. et al. (2015) ‘Construction d’un vocabulaire patient/médecin dédié au cancer du sein à partir des médias sociaux’, 26. Journées Francophones d’Ingénierie des Connaissances (IC), Rennes.
  • Thibaud, E. and Opitz, T. (2015) ‘Efficient inference and simulation for elliptical Pareto processes’, Biometrika, 102(4), pp. 855–870.
  • Opitz, T. et al. (2014) ‘Breast cancer and quality of life: medical information extraction from health forums’, in Medical Informatics Europe Conference 2014, pp. 1070–1074.
  • Opitz, T. (2013) ‘Extremal t processes: Elliptical domain of attraction and a spectral representation’, J. Multivar. Anal., 122, pp. 409–413.

Responsibilities and research networks

  • Steering committee member of the Metaprogramme CLIMAE of INRAE ("Adaptation and mitigation of climate change").
  • Steering committee member of RESSTE ("RESeau Statistique pour données Spatio-TEmporelles"), one of INRAE's current research networks.
  • Elected member and webmaster of the "Groupe Environnement et Statistique" of the French Statistical Society.
  • Co-organizer of BioSP's seminar.

 

Teaching

  • 2020/2021: Course "Introduction to extreme-value analysis" at École Centrale Marseille, Master Climaths 
  • since 2018: Course "Statistique spatiale et écologie", M2 Data Science, Marseille
  • 2019,2021: One-day master course on Multivariate Extremes, ATHENS network, MinesParisTech
  • 2016/2017: Statistique Descriptive 2, L1 STID, IUT Avignon

Projects with funding

  • PhD project of Ryan Cotsakis (2021-2024), co-supervised with Elena di Bernardino (Université Côte d'Azur, 3IA Côte d'Azur). Funding: 3IA Côte d'Azur.
    "Stochastic geometry theory and tools for spatiotemporal extremes".
  • Co-Investigator of a KAUST Competitive Research Grant  (2018-2021) coordinated by Raphael Huser, with partners at KAUST and Lancaster University:
    "Statistical Estimation and Detection of Extreme Hot Spots, with Environmental and Ecological Applications".
  • PhD project of Patrizia Zamberletti (2018-2021), co-supervised with Julien Papaix and Edith Gabriel at BioSP. Funding: INRAE divisions MIA (25%) and SPE (25%), and Provence-Alpes Côtes d’Azur region (50%).
    "Simulation and inference of agricultural landscapes using stochastic geometry; agroecological analysis of numerical simulations of spatially explicit population dynamics model (sensitivity analysis, statistical learning)".
  • Post-doc project (2017-2019) of Fátima Palacíos-Rodriguez, with funding from MUSE, Labex Numev and Inria:
    "Semiparametric resampling of extreme events over space and time, with an application to precipitation data, and with a view towards extreme risk measures".
  • LEFE-CERISE, LEFE-FRAISE projects (2016-2021) funded by INSU:
    "Simulation de scénarii intégrant des champs extrêmes spatio-temporelle avec éventuelle indépendance asymptotique pour des études d'impact en science de l'environnement".
  • Pari scientifique EA, coordinated by Hocine Bourennane and Nicolas Saby (2018-2020):
     "Innovative approaches for space-time prediction and mapping of soil properties using INLA" ("Approches innovantes de prévisions géo-datées des propriétés des sols").