scholarly journals Inconsistent strategies to spin up models in CMIP5: implications for ocean biogeochemical model performance assessment

2015 ◽  
Vol 8 (10) ◽  
pp. 8751-8808 ◽  
Author(s):  
R. Séférian ◽  
M. Gehlen ◽  
L. Bopp ◽  
L. Resplandy ◽  
J. C. Orr ◽  
...  
Keyword(s):  
Performance Assessment ◽  
Initial Conditions ◽  
Model Performance ◽  
Cmip5 Models ◽  
Biogeochemical Model ◽  
Earth System ◽  
Model Intercomparison ◽  
System Models ◽  
Marine Biogeochemistry ◽  
Earth System Models

Abstract. During the fifth phase of the Coupled Model Intercomparison Project (CMIP5) substantial efforts were carried out on the systematic assessment of the skill of Earth system models. One goal was to check how realistically representative marine biogeochemical tracer distributions could be reproduced by models. Mean-state assessments routinely compared model hindcasts to available modern biogeochemical observations. However, these assessments considered neither the extent of equilibrium in modeled biogeochemical reservoirs nor the sensitivity of model performance to initial conditions or to the spin-up protocols. Here, we explore how the large diversity in spin-up protocols used for marine biogeochemistry in CMIP5 Earth system models (ESM) contribute to model-to-model differences in the simulated fields. We take advantage of a 500 year spin-up simulation of IPSL-CM5A-LR to quantify the influence of the spin-up protocol on model ability to reproduce relevant data fields. Amplification of biases in selected biogeochemical fields (O2, NO3, Alk-DIC) is assessed as a function of spin-up duration. We demonstrate that a relationship between spin-up duration and assessment metrics emerges from our model results and is consistent when confronted against a larger ensemble of CMIP5 models. This shows that drift has implications on their performance assessment in addition to possibly aliasing estimates of climate change impact. Our study suggests that differences in spin-up protocols could explain a substantial part of model disparities, constituting a source of model-to-model uncertainty. This requires more attention in future model intercomparison exercices in order to provide realistic ESM results on marine biogeochemistry and carbon cycle feedbacks.

scholarly journals Inconsistent strategies to spin up models in CMIP5: implications for ocean biogeochemical model performance assessment

2016 ◽  
Vol 9 (5) ◽  
pp. 1827-1851 ◽  
Author(s):  
Roland Séférian ◽  
Marion Gehlen ◽  
Laurent Bopp ◽  
Laure Resplandy ◽  
James C. Orr ◽  
...  
Keyword(s):  
Performance Assessment ◽  
Initial Conditions ◽  
Model Performance ◽  
Cmip5 Models ◽  
Biogeochemical Model ◽  
Earth System ◽  
Model Intercomparison ◽  
System Models ◽  
Marine Biogeochemistry ◽  
Earth System Models

Abstract. During the fifth phase of the Coupled Model Intercomparison Project (CMIP5) substantial efforts were made to systematically assess the skill of Earth system models. One goal was to check how realistically representative marine biogeochemical tracer distributions could be reproduced by models. In routine assessments model historical hindcasts were compared with available modern biogeochemical observations. However, these assessments considered neither how close modeled biogeochemical reservoirs were to equilibrium nor the sensitivity of model performance to initial conditions or to the spin-up protocols. Here, we explore how the large diversity in spin-up protocols used for marine biogeochemistry in CMIP5 Earth system models (ESMs) contributes to model-to-model differences in the simulated fields. We take advantage of a 500-year spin-up simulation of IPSL-CM5A-LR to quantify the influence of the spin-up protocol on model ability to reproduce relevant data fields. Amplification of biases in selected biogeochemical fields (O2, NO3, Alk-DIC) is assessed as a function of spin-up duration. We demonstrate that a relationship between spin-up duration and assessment metrics emerges from our model results and holds when confronted with a larger ensemble of CMIP5 models. This shows that drift has implications for performance assessment in addition to possibly aliasing estimates of climate change impact. Our study suggests that differences in spin-up protocols could explain a substantial part of model disparities, constituting a source of model-to-model uncertainty. This requires more attention in future model intercomparison exercises in order to provide quantitatively more correct ESM results on marine biogeochemistry and carbon cycle feedbacks.

scholarly journals Examining soil carbon uncertainty in a global model: response of microbial decomposition to temperature, moisture and nutrient limitation

2013 ◽  
Vol 10 (6) ◽  
pp. 10229-10269
Author(s):  
J.-F. Exbrayat ◽  
A. J. Pitman ◽  
Q. Zhang ◽  
G. Abramowitz ◽  
Y.-P. Wang
Keyword(s):  
Soil Carbon ◽  
Nutrient Limitation ◽  
Historical Change ◽  
Cmip5 Models ◽  
Earth System ◽  
Microbial Decomposition ◽  
Soil C ◽  
System Models ◽  
C And N ◽  
Earth System Models

Abstract. Reliable projections of future climate require land–atmosphere carbon (C) fluxes to be represented realistically in Earth System Models. There are several sources of uncertainty in how carbon is parameterized in these models. First, while interactions between the C, nitrogen (N) and phosphorus (P) cycles have been implemented in some models, these lead to diverse changes in land–atmosphere fluxes. Second, while the parameterization of soil organic matter decomposition is similar between models, formulations of the control of the soil physical state on microbial activity vary widely. We address these sources uncertainty by implementing three soil moisture (SMRF) and three soil temperature (STRF) respiration functions in an Earth System Model that can be run with three degrees of biogeochemical nutrient limitation (C-only, C and N, and C and N and P). All 27 possible combinations of a SMRF with a STRF and a biogeochemical mode are equilibrated before transient historical (1850–2005) simulations are performed. As expected, implementing N and P limitation reduces the land carbon sink, transforming some regions from net sinks to net sources over the historical period (1850–2005). Differences in the soil C balance implied by the various SMRFs and STRFs also change the sign of some regional sinks. Further, although the absolute uncertainty in global carbon uptake is reduced, the uncertainty due to the SMRFs and STRFs grows relative to the inter-annual variability in net uptake when N and P limitations are added. We also demonstrate that the equilibrated soil C also depend on the shape of the SMRF and STRF. Equilibration using different STRFs and SMRFs and nutrient limitation generates a six-fold range of global soil C that largely mirrors the range in available (17) CMIP5 models. Simulating the historical change in soil carbon therefore critically depends on the choice of STRF, SMRF and nutrient limitation, as it controls the equilibrated state to which transient conditions are applied. This direct effect of the representation of microbial decomposition in Earth System Models adds to recent concerns on the adequacy of these simple representations of very complex soil carbon processes.

scholarly journals Tracking Improvement in Simulated Marine Biogeochemistry Between CMIP5 and CMIP6

2020 ◽  
Author(s):  
Roland Séférian ◽  
Sarah Berthet ◽  
Andrew Yool ◽  
Julien Palmiéri ◽  
Laurent Bopp ◽  
...  
Keyword(s):  
Coupled Model ◽  
Climate Feedback ◽  
Full Potential ◽  
Earth System ◽  
Current Generation ◽  
System Models ◽  
Biogeochemical Models ◽  
Marine Biogeochemistry ◽  
Intercomparison Project ◽  
Earth System Models

Abstract Purpose of Review The changes or updates in ocean biogeochemistry component have been mapped between CMIP5 and CMIP6 model versions, and an assessment made of how far these have led to improvements in the simulated mean state of marine biogeochemical models within the current generation of Earth system models (ESMs). Recent Findings The representation of marine biogeochemistry has progressed within the current generation of Earth system models. However, it remains difficult to identify which model updates are responsible for a given improvement. In addition, the full potential of marine biogeochemistry in terms of Earth system interactions and climate feedback remains poorly examined in the current generation of Earth system models. Summary Increasing availability of ocean biogeochemical data, as well as an improved understanding of the underlying processes, allows advances in the marine biogeochemical components of the current generation of ESMs. The present study scrutinizes the extent to which marine biogeochemistry components of ESMs have progressed between the 5th and the 6th phases of the Coupled Model Intercomparison Project (CMIP).

scholarly journals Solar irradiance reduction to counteract radiative forcing from a quadrupling of CO<sub>2</sub>: climate responses simulated by four earth system models

2012 ◽  
Vol 3 (1) ◽  
pp. 63-78 ◽  
Author(s):  
H. Schmidt ◽  
K. Alterskjær ◽  
D. Bou Karam ◽  
O. Boucher ◽  
A. Jones ◽  
...  
Keyword(s):  
Radiative Forcing ◽  
Surface Air Temperature ◽  
Co2 Concentration ◽  
Northern Eurasia ◽  
Earth System ◽  
Model Intercomparison ◽  
Regional Patterns ◽  
System Models ◽  
Eu Project ◽  
Earth System Models

Abstract. In this study we compare the response of four state-of-the-art Earth system models to climate engineering under scenario G1 of two model intercomparison projects: GeoMIP (Geoengineering Model Intercomparison Project) and IMPLICC (EU project "Implications and risks of engineering solar radiation to limit climate change"). In G1, the radiative forcing from an instantaneous quadrupling of the CO2 concentration, starting from the preindustrial level, is balanced by a reduction of the solar constant. Model responses to the two counteracting forcings in G1 are compared to the preindustrial climate in terms of global means and regional patterns and their robustness. While the global mean surface air temperature in G1 remains almost unchanged compared to the control simulation, the meridional temperature gradient is reduced in all models. Another robust response is the global reduction of precipitation with strong effects in particular over North and South America and northern Eurasia. In comparison to the climate response to a quadrupling of CO2 alone, the temperature responses are small in experiment G1. Precipitation responses are, however, in many regions of comparable magnitude but globally of opposite sign.

scholarly journals Oceanic primary production decline halved in eddy-resolving simulations of global warming

2021 ◽  
Vol 18 (14) ◽  
pp. 4321-4349
Author(s):  
Damien Couespel ◽  
Marina Lévy ◽  
Laurent Bopp
Keyword(s):  
Climate Change ◽  
Global Warming ◽  
Primary Production ◽  
Nutrient Transport ◽  
Biogeochemical Model ◽  
Earth System ◽  
System Models ◽  
Marine Biomass ◽  
Fish Catch ◽  
Earth System Models

Abstract. The decline in ocean primary production is one of the most alarming consequences of anthropogenic climate change. This decline could indeed lead to a decrease in marine biomass and fish catch, as highlighted by recent policy-relevant reports. Because of computational constraints, current Earth system models used to project ocean primary production under global warming scenarios have to parameterize flows occurring below the resolution of their computational grid (typically 1∘). To overcome these computational constraints, we use an ocean biogeochemical model in an idealized configuration representing a mid-latitude double-gyre circulation and perform global warming simulations under an increasing horizontal resolution (from 1 to 1/27∘) and under a large range of parameter values for the eddy parameterization employed in the coarse-resolution configuration. In line with projections from Earth system models, all our simulations project a marked decline in net primary production in response to the global warming forcing. Whereas this decline is only weakly sensitive to the eddy parameters in the eddy-parameterized coarse 1∘ resolution simulations, the simulated decline in primary production in the subpolar gyre is halved at the finest eddy-resolving resolution (−12 % at 1/27∘ vs. −26 % at 1∘) at the end of the 70-year-long global warming simulations. This difference stems from the high sensitivity of the sub-surface nutrient transport to model resolution. Although being only one piece of a much broader and more complicated response of the ocean to climate change, our results call for improved representation of the role of eddies in nutrient transport below the seasonal mixed layer to better constrain the future evolution of marine biomass and fish catch potential.

scholarly journals Variable particle size distributions reduce the sensitivity of global export flux to climate change

2021 ◽  
Vol 18 (1) ◽  
pp. 229-250
Author(s):  
Shirley W. Leung ◽  
Thomas Weber ◽  
Jacob A. Cram ◽  
Curtis Deutsch
Keyword(s):  
Climate Change ◽  
Particle Size ◽  
Euphotic Zone ◽  
Nutrient Concentrations ◽  
Biogeochemical Model ◽  
Size Distributions ◽  
Earth System ◽  
Particle Size Distributions ◽  
System Models ◽  
Earth System Models

Abstract. Recent earth system models predict a 10 %–20 % decrease in particulate organic carbon export from the surface ocean by the end of the 21st century due to global climate change. This decline is mainly caused by increased stratification of the upper ocean, resulting in reduced shallow subsurface nutrient concentrations and a slower supply of nutrients to the surface euphotic zone in low latitudes. These predictions, however, do not typically account for associated changes in remineralization depths driven by sinking-particle size. Here we combine satellite-derived export and particle size maps with a simple 3-D global biogeochemical model that resolves dynamic particle size distributions to investigate how shifts in particle size may buffer or amplify predicted changes in surface nutrient supply and therefore export production. We show that higher export rates are empirically correlated with larger sinking particles and presumably larger phytoplankton, particularly in tropical and subtropical regions. Incorporating these empirical relationships into our global model shows that as circulation slows, a decrease in export is associated with a shift towards smaller particles, which sink more slowly and are thus remineralized shallower. This shift towards shallower remineralization in turn leads to greater recycling of nutrients in the upper water column and thus faster nutrient recirculation into the euphotic zone. The end result is a boost in productivity and export that counteracts the initial circulation-driven decreases. This negative feedback mechanism (termed the particle-size–remineralization feedback) slows export decline over the next century by ∼ 14 % globally (from −0.29 to −0.25 GtC yr−1) and by ∼ 20 % in the tropical and subtropical oceans, where export decreases are currently predicted to be greatest. Our findings suggest that to more accurately predict changes in biological pump strength under a warming climate, earth system models should include dynamic particle-size-dependent remineralization depths.

scholarly journals Oceanic Primary production decline halved in eddy-resolving simulations of global warming

2021 ◽  
Author(s):  
Damien Couespel ◽  
Marina Lévy ◽  
Laurent Bopp
Keyword(s):  
Global Warming ◽  
Primary Production ◽  
Nutrient Transport ◽  
Biogeochemical Model ◽  
Earth System ◽  
System Models ◽  
Coarse Resolution ◽  
Marine Biomass ◽  
Fish Catch ◽  
Earth System Models

&lt;p&gt;The decline in ocean primary production is one of the most alarming consequences of anthropogenic climate change. This decline could indeed lead to a decrease in marine biomass and fish catch, as highlighted by recent policy-relevant reports. Because of computational constraints, current Earth System Models used to project ocean primary production under global warming scenarios have to parameterize flows occurring below the resolution of their computational grid (typically 1&amp;#176;). To overcome these computational constraints, we use an ocean biogeochemical model in an idealized configuration representing a mid-latitude double-gyre circulation, and perform global warming simulations under increasing horizontal resolution &amp;#160;(from 1&amp;#176; to 1/27&amp;#176;) and under a large range of parameter values for the eddy parameterization employed in the coarse resolution configuration. In line with projections from Earth System Models, all our simulations project a marked decline in net primary production in response to the global warming forcing. Whereas this decline is only weakly sensitive to the eddy parameters in the eddy-parametrized coarse resolution, the simulated decline in primary production is halved at the finest eddy-resolving resolution (-12% at 1/27&amp;#176; vs -26 at 1&amp;#176;). This difference stems from the high sensitivity of the subsurface nutrient transport to model resolution. Our results call for improved representation of the role of eddies on nutrient transport below the seasonal mixed-layer to better constrain the future evolution of marine biomass and fish catch potential for decision-making.&lt;/p&gt;

scholarly journals CPMIP: measurements of real computational performance of Earth system models in CMIP6

2017 ◽  
Vol 10 (1) ◽  
pp. 19-34 ◽  
Author(s):  
Venkatramani Balaji ◽  
Eric Maisonnave ◽  
Niki Zadeh ◽  
Bryan N. Lawrence ◽  
Joachim Biercamp ◽  
...  
Keyword(s):  
Climate Models ◽  
Climate Model ◽  
Initial Conditions ◽  
Performance Model ◽  
Earth System ◽  
System Models ◽  
Performance Engineering ◽  
Computational Performance ◽  
Sensitive Dependence ◽  
Earth System Models

Abstract. A climate model represents a multitude of processes on a variety of timescales and space scales: a canonical example of multi-physics multi-scale modeling. The underlying climate system is physically characterized by sensitive dependence on initial conditions, and natural stochastic variability, so very long integrations are needed to extract signals of climate change. Algorithms generally possess weak scaling and can be I/O and/or memory-bound. Such weak-scaling, I/O, and memory-bound multi-physics codes present particular challenges to computational performance. Traditional metrics of computational efficiency such as performance counters and scaling curves do not tell us enough about real sustained performance from climate models on different machines. They also do not provide a satisfactory basis for comparative information across models. codes present particular challenges to computational performance. We introduce a set of metrics that can be used for the study of computational performance of climate (and Earth system) models. These measures do not require specialized software or specific hardware counters, and should be accessible to anyone. They are independent of platform and underlying parallel programming models. We show how these metrics can be used to measure actually attained performance of Earth system models on different machines, and identify the most fruitful areas of research and development for performance engineering. codes present particular challenges to computational performance. We present results for these measures for a diverse suite of models from several modeling centers, and propose to use these measures as a basis for a CPMIP, a computational performance model intercomparison project (MIP).

Using objective comparisons of observed and simulated precipitation to help guide the improvement of Earth System Models

2020 ◽  
Author(s):  
Peter Gleckler ◽  
Angeline Pendergrass
Keyword(s):  
Model Development ◽  
Model Performance ◽  
Source Analysis ◽  
Earth System ◽  
Frequency Distributions ◽  
System Models ◽  
Simulated Precipitation ◽  
Tiered System ◽  
Earth System Models

&lt;p&gt;In this presentation we discuss a community-based effort to establish the benchmarking of simulated precipitation in Earth System Models.&amp;#160;&amp;#160; We first summarize the impetus and outcomes of a recent workshop dedicated to the topic.&amp;#160;&amp;#160;&amp;#160; This includes the identification of a tiered system of objective tests (metrics) for the following climatological characteristics:&amp;#160; the mean state, seasonal cycle, variability across time scales, intensity/frequency distributions, extremes and drought.&amp;#160;&amp;#160; Preliminary results are shown gauging model performance changes across multiple generations of CMIP.&amp;#160;&amp;#160; The performance tests we describe are part of an open-source analysis framework being made available to model developers to help them make judgements about the quality of simulated precipitation during the model development process.&lt;/p&gt;

scholarly journals CPMIP: Measurements of Real Computational Performance of Earth System Models

2016 ◽  
Author(s):  
V. Balaji ◽  
E. Maisonnave ◽  
N. Zadeh ◽  
B. N. Lawrence ◽  
J. Biercamp ◽  
...  
Keyword(s):  
Climate Models ◽  
Climate Model ◽  
Initial Conditions ◽  
Earth System ◽  
System Models ◽  
Performance Engineering ◽  
Multi Scale ◽  
Computational Performance ◽  
Sensitive Dependence ◽  
Earth System Models

Abstract. A climate model represents a multitude of processes on a variety of time and space scales; a canonical example of multi-physics multi-scale modeling. The underlying climate system is physically characterized by sensitive dependence on initial conditions, and natural stochastic variability, so very long integrations are needed to extract signals of climate change. Algorithms generally possess weak scaling and can be I/O and/or memory bound. Such weak-scaling, I/O and memory-bound multi-physics codes present particular challenges to computational performance. Traditional metrics of computational efficiency such as performance counters and scaling curves do not tell us enough about real sustained performance from climate models on different machines. They also do not provide a satisfactory basis for comparative information across models. We introduce a set of metrics that can be used for the study of computational performance of climate (and Earth System) models. These measures do not require specialized software or specific hardware counters, and should be accessible to anyone. They are independent of platform, and underlying parallel programming models. We show how these metrics can be used to measure actually attained performance of Earth system models on different machines, and identify the most fruitful areas of research and development for performance engineering. We present results for these measures for a diverse suite of models from several modeling centres, and propose to use these measures as a basis for a CPMIP, a computational performance MIP.

Sign in / Sign up

Export Citation Format

Share Document