Multi-model ensembles improve predictions of crop-environment-management interactions - INRAE - Institut national de recherche pour l’agriculture, l’alimentation et l’environnement Accéder directement au contenu
Article Dans Une Revue Global Change Biology Année : 2018

Multi-model ensembles improve predictions of crop-environment-management interactions

1 AGIR - AGroécologie, Innovations, teRritoires
2 LEPSE - Écophysiologie des Plantes sous Stress environnementaux
3 NAU - Nanjing Agricultural University
4 UF|ABE - Department of Agricultural and Biological Engineering [Gainesville]
5 INRES - Institute of Crop Science and Resource Conservation [Bonn]
6 ZALF - Leibniz-Zentrum für Agrarlandschaftsforschung = Leibniz Centre for Agricultural Landscape Research
7 CSIRO - Commonwealth Scientific and Industrial Research Organisation [Canberra]
8 Plant Production Systems Group
9 CCAFS - Agriculture and Food Security
10 Biological Systems Engineering
11 Department of Agronomy
12 Department of Earth and Environmental Sciences [East Lansing]
13 Plant Pathology
14 The James Hutton Institute
15 University of Leeds
16 CGIAR - Consultative Group on International Agricultural Research
17 GMO Unit
18 Department Terra & AgroBioChem, Gembloux Agro‐Bio Tech
19 Center for Development Research
20 Universidad de Córdoba = University of Córdoba [Córdoba]
21 Agriculture Victoria Research
22 University of Melbourne
23 Institute of Soil Science and Land Evaluation
24 GDEC - Génétique Diversité et Ecophysiologie des Céréales
25 UF - University of Florida [Gainesville]
26 Department of Geographical Sciences
27 Texas A and M AgriLife Research
28 Institute of landscape systems analysis
29 School of Earth and Environment
30 EFSA - European Food Safety Authority
31 PIK - Potsdam Institute for Climate Impact Research
32 CESCRA - Centre for Environment Science and Climate Resilient Agriculture
33 DEDJTR - Department of Economic Development, Jobs, Transport and Resources
34 LUKE - Natural Resources Institute Finland
35 AGROCLIM - Agroclim
36 UMG - University Medical Center Göttingen
37 Computational and Systems Biology Department
38 Water & Food and Water Systems & Global Change Group
39 IGSNRR - Institute of geographical sciences and natural resources research [CAS]
40 Plant Production Systems
41 State Key Laboratory of Earth Surface Processes and Resource Ecology
Davide Cammarano
  • Fonction : Auteur
  • PersonId : 968228
Giacomo de Sanctis
  • Fonction : Auteur
Gerrit Hoogenboom
C. Klein
  • Fonction : Auteur
Claas Nendel
Eckart Priesack
  • Fonction : Auteur
Dominique Ripoche
  • Fonction : Auteur
  • PersonId : 1203876

Résumé

A recent innovation in assessment of climate change impact on agricultural production has been to use crop multi model ensembles (MMEs). These studies usually find large variability between individual models but that the ensemble mean (e‐mean) and median (e‐median) often seem to predict quite well. However few studies have specifically been concerned with the predictive quality of those ensemble predictors. We ask what is the predictive quality of e‐mean and e‐median, and how does that depend on the ensemble characteristics. Our empirical results are based on five MME studies applied to wheat, using different data sets but the same 25 crop models. We show that the ensemble predictors have quite high skill and are better than most and sometimes all individual models for most groups of environments and most response variables. Mean squared error of e‐mean decreases monotonically with the size of the ensemble if models are added at random, but has a minimum at usually 2‐6 models if best‐fit models are added first. Our theoretical results describe the ensemble using four parameters; average bias, model effect variance, environment effect variance and interaction variance. We show analytically that mean squared error of prediction (MSEP) of e‐mean will always be smaller than MSEP averaged over models, and will be less than MSEP of the best model if squared bias is less than the interaction variance. If models are added to the ensemble at random, MSEP of e‐mean will decrease as the inverse of ensemble size, with a minimum equal to squared bias plus interaction variance. This minimum value is not necessarily small, and so it is important to evaluate the predictive quality of e‐mean for each target population of environments. These results provide new information on the advantages of ensemble predictors, but also show their limitations.

Dates et versions

hal-02625468 , version 1 (26-05-2020)

Identifiants

Citer

Daniel Wallach, Pierre Martre, Bo Liu, Senthold Asseng, Frank Ewert, et al.. Multi-model ensembles improve predictions of crop-environment-management interactions. Global Change Biology, 2018, 24 (11), pp.5072-5083. ⟨10.1111/gcb.14411⟩. ⟨hal-02625468⟩
105 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More