.. _ifs_nemo_model_historical_projection: Historical and Projection Simulations ##################################### The IFS-NEMO historical and projection experiment consists of two simulations: a historical simulation covering 1990–2014 and a projection simulation covering 2015–2049 (SSP3-7.0), spanning 60 years in total. The projection is a direct continuation of the historical simulation. The atmospheric component IFS used a horizontal resolution of 5 km (TCo2559 grid) while the ocean and sea ice components NEMO and SI3 used a resolution of 1/12º (<9 km, eORCA12 grid). Both simulations were conducted in MN5, as indicated in the table. +----------+---------------------------------------------+--------------+-------------+-------------------------------------------------+-----------------------------------+ | Model | Spatial Resolution | Time | Realization | Experiment ID | DestinE Data Lake (bridge site) | +==========+=============================================+==============+=============+=================================================+===================================+ | IFS-NEMO | 5km atmosphere; 1/12º (<9 km) ocean/sea ice | 1990-2014 | 1 | climatedt-gen2-ifs-nemo-baseline-hist-5km-r1 | MN5 data bridge | +----------+---------------------------------------------+--------------+-------------+-------------------------------------------------+-----------------------------------+ | IFS-NEMO | 5km atmosphere; 1/12º (<9 km) ocean/sea ice | 2015-2049 | 1 |climatedt-gen2-ifs-nemo-projections-ssp370-5km-r1| MN5 data bridge | +----------+---------------------------------------------+--------------+-------------+-------------------------------------------------+-----------------------------------+ For evaluation of these simulations, see :ref:`evaluation_ifs_nemo`. .. .. figure:: ../../../../evaluation/general_evaluation/figures/climate_metrics.performance_indices.climatedt-o25.1.IFS-NEMO.historical-1990.r1.png .. :name: ifs-nemo_hist_pi .. Performance Index table for the historical simulation. Performance values refer to the mean of CMIP6 models, with values below 1 indicating improved performance and values above 1 degraded performance. .. Global mean temperatures (:numref:`ifs-nemo_hist_proj_gregory`, left panel) over the historical period show a slightly weaker long-term trend than observations, which might be explained by the surface cooling drift already seen in the :ref:`ifs_nemo_model_control`. It also shows a small mean-state cold bias of ~0.25°C. The future projection depicts an acceleration of the historical warming trend by about 75%. .. Simulated values in the top of the atmosphere net radiation (:numref:`ifs-nemo_hist_proj_gregory`, middle panel) are within the range of observed values, although they seem to exhibit a substantially weaker trend. Part of these differences might be explained by internal variability, given the comparatively shorter observational period, and the larger year-to-year variations in IFS-NEMO, which introduce uncertainty in the computed trends. .. The Gregory plot (:numref:`ifs-nemo_hist_proj_gregory`, right panel) shows that the simulation stays largely within the observed ranges, with a potential overestimation of the response to Pinatubo in the years following to the eruption (i.e. 1991). .. .. figure:: ../../../../evaluation/mn5/figures/IFS-NEMO-Historical_SSP370_Tco2559_timeseries_Gregory_absoluteT.png .. :name: ifs-nemo_hist_proj_gregory .. Left\: Time series of the globally averaged annual surface air temperature in ERA5 and the historical and scenario IFS-NEMO simulations. Fit-lines representing the trends over the overlap period between ERA5 and the simulations are also included. Middle\: Same figure but for the timeseries of the net heat fluxes at the top of the atmosphere (TOA), including observations from CERES. Right\: Gregory plot of the combined IFS-NEMO simulations. The mean values and ranges of the observed TOA fluxes and global mean surface air temperatures are included for reference. .. Spatial maps of mean-state biases in annual surface air temperature (:numref:`ifs-nemo_hist_tas_bias`) show relatively small errors with respect to ERA5 over the three major ocean basins and most continental areas (global mean bias -0.10 K, RMSE 0.96 K). The global cold bias mostly arises from the Arctic region and Southern Ocean, likely linked to errors in sea ice representation. IFS-NEMO successfully mitigates the warm biases typical of eastern boundary coastal upwelling regions (e.g., Humboldt and Benguela currents) that commonly afflict coarser CMIP6 models. Over the continents, IFS-NEMO tends to be warmer than ERA5, like most CMIP6 models. Some local differences, evident in the main mountainous regions, might be due to the finer representation of orography in the model. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/tas_annual_bias_combined_cropped.png .. :name: ifs-nemo_hist_tas_bias .. Spatial maps of the climatological biases of annual surface air temperature in the historical IFS-NEMO simulation and the CMIP6 multi-model mean. Biases are computed against Berkeley Earth climatology over the period 1990–2014. .. In terms of mean sea level pressure (:numref:`ifs-nemo_hist_psl_bias`), IFS-NEMO performs well (global mean bias 0.08 Pa, RMSE 88.91 Pa), comparable to the CMIP6 multi-model mean (RMSE ~87 Pa). The polar regions show positive biases, characteristic of cold high-latitude mean states that are dynamically consistent with excessive surface cooling and sea ice, highlighting the strongly coupled nature of tropical and polar mean-state errors. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/psl_annual_bias_combined_cropped.png .. :name: ifs-nemo_hist_psl_bias .. Spatial maps of the climatological biases of annual mean sea level pressure in the historical IFS-NEMO simulation and the CMIP6 multi-model mean. Biases are computed against ERA5 climatology over the period 1990–2014. .. IFS-NEMO also shows good performance in terms of the climatological annual precipitation rate (:numref:`ifs-nemo_hist_pr_bias`), sharing the broad spatial structure of the CMIP6 Multi-Model Mean, with wet biases along the mid-latitude storm tracks and a pronounced dipole pattern in the tropical Pacific indicative of a double-ITCZ or ITCZ displacement error. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/pr_annual_bias_combined_cropped.png .. :name: ifs-nemo_hist_pr_bias .. Spatial maps of the climatological biases of annual precipitation in the historical IFS-NEMO simulation and the CMIP6 multi-model mean. Biases are computed against MSWEP climatology over the period 1990–2014. .. The Earth energy imbalance time series (:numref:`ifs-nemo_eei`) shows that IFS-NEMO accurately captures the observed annual mean energy imbalance of approximately 0.5–1.0 W/m² post-2000, significantly outperforming the CMIP6 multi-model mean which exhibits a systematic positive bias of approximately 1.0 W/m² relative to CERES observations. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/radiation_imbalance_timeseries.png .. :name: ifs-nemo_eei .. Time series of the global-mean net top-of-atmosphere radiation (Earth’s energy imbalance) from 1990 to 2014, comparing IFS-NEMO with CERES observations and the CMIP6 multi-model mean. .. While the sea ice extent (:numref:`ifs-nemo_hist_sea_ice_extent`) is slightly overestimated in the Northern Hemisphere compared to OSI-SAF observations, it shows a declining trend since the 1990s that is largely consistent with the observed one. In the Southern Hemisphere both IFS-NEMO and OSI-SAF show an increasing trend, although with some different features. IFS-NEMO starts close to OSI-SAF but develops positive bias in the late 1990s, which might reflect some initialization adjustment. CMIP6 models show a large inter-model spread, with the ensemble mean showing a slight sea ice decline, that opposes the observed one. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/sea_ice_extent_timeseries.png .. :name: ifs-nemo_hist_sea_ice_extent .. Time series of monthly and annual-mean sea ice extent for the Northern and Southern Hemispheres, comparing IFS-NEMO against OSI-SAF satellite observations and a CMIP6 multi-model ensemble. .. .. rubric:: Further evaluation .. Additional evaluation plots for the IFS-NEMO simulations are available .. in the `Climate DT Evaluation Charts `_. .. Sea ice volume (:numref:`ifs-nemo_sea_ice_volume`) is overestimated in the Northern Hemisphere by up to a factor of two compared to observation-constrained estimates of PIOMAS. An unphysical rapid increase during the 1990s prevents the model from capturing the observed steady downward trend, also simulated by the CMIP6 multi-model mean, strongly suggesting a model spin-up issue or a severe adjustment to the coupled state rather than a physical response to historical forcing. In the Southern Hemisphere, IFS-NEMO compares well with PIOMAS, both in terms of sea ice volume and the representation of a slight upward trend, not captured by the CMIP6 multi-model mean. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/sea_ice_volume_timeseries.png .. :name: ifs-nemo_sea_ice_volume .. Time series of monthly and annual-mean sea ice volume for the Northern and Southern Hemispheres, comparing IFS-NEMO against OSI-SAF satellite observations and a CMIP6 multi-model ensemble. .. The subsurface ocean conservative temperature Hovmöller diagram (:numref:`ifs-nemo_hist_thetao_hovmoller_anomref`) shows that IFS-NEMO rapidly develops a persistent cold bias of ~ 0.5°C in the upper 300 m and a compensating warm anomaly in the subsurface layer around 300–800 m depth, indicative of a thermocline that is too deep or too diffuse relative to EN4. This warm-over-cold dipole structure develops rapidly at the start of the simulation and remains quasi-stationary thereafter, suggesting it reflects a fast adjustment to a drifted equilibrium state rather than a progressive trend. A weaker but widespread warm anomaly also extends throughout the deep ocean below 1000 m, pointing to slow barotropic drift over the simulation period. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/en4_thetao_hovmoller_anomref_combined.png .. :name: ifs-nemo_hist_thetao_hovmoller_anomref .. Time-depth Hovmöller diagrams of IFS-NEMO global-mean ocean potential temperature (converted offline from conservative temperature) anomalies relative to the initial EN4 observations-based product from 1990 to 2014. .. Global ocean temperature time series integrated at different depths (:numref:`ifs-nemo_hist_thetao_depth_timeseries`) further illustrate the drifts hinted by the Hovmöller. In the 0–700 m range, IFS-NEMO rapidly develops a cold bias and remains flat throughout the simulation, failing to capture the steady warming trend observed in EN4 over the period. Further down, IFS-NEMO is too cold between 700-2000 m and too warm below 2000 m compared to EN5, and develops opposite trends after year 2002. These opposing trends point to a vertical redistribution of heat, with intermediate layers cooling and the deep ocean warming, possibly due to an insufficient spin-up period, though contributions from systematic biases in diapycnal mixing or deep water formation cannot be excluded. .. .. figure:: ../../../../evaluation/ifs_nemo_eval/en4_thetao_depth_timeseries.png .. :name: ifs-nemo_hist_thetao_depth_timeseries .. Volume-weighted mean ocean potential temperature (converted offline from conservative temperature) time series for three depth layers (0–700 m, 700–2000 m, 2000 m–bottom) comparing IFS-NEMO against EN4 v4.2.2 observations-based product from 1990 to 2014. .. Projection Assessment .. ^^^^^^^^^^^^^^^^^^^^^ .. To assess the projected climate changes from the IFS-NEMO Destination Earth simulations, a delta approach is adopted whereby anomalies are computed relative to the historical 1990–2014 climatology. This baseline period coincides with the historical simulation evaluated in the preceding sections, providing a consistent reference for quantifying change. The model's ability to reproduce observed recent changes is first evaluated by comparing the simulated 2015–2024 anomaly against ERA5 (for 2 m temperature) and MSWEP (for precipitation), before examining the projected 2025–2049 anomaly to characterise the expected evolution of the climate signal. .. The projected changes in annual mean 2 m temperature from IFS-NEMO (:numref:`ifs-nemo_proj_2t`) broadly reproduce the spatial pattern of recent observed warming derived from ERA5, with a predominance of positive anomalies relative to the 1990–2014 baseline across both land and ocean. However, IFS-NEMO underestimates the magnitude of the near-present warming signal, with a global mean anomaly of +0.34 K over 2015–2024 compared to +0.49 K in ERA5, suggesting the model may be slightly underestimating the pace of recent warming. The spatial structure is nonetheless well captured, with stronger warming over continental interiors and the Arctic consistent with observed patterns. In the period 2025–2049, IFS-NEMO projects a global mean warming of +0.78 K relative to the historical baseline, with a clear amplification of the signal particularly over land areas and the Arctic, this latter consistent with polar amplification. Notably, localised cooling anomalies emerge in the North Atlantic and around Antarctica in the future projection, which may reflect a weakening of the Atlantic meridional overturning circulation and changes in Southern Ocean dynamics respectively, features that are physically plausible under continued greenhouse gas forcing. .. .. figure:: ../../../../evaluation/levante/Fig/IFS-NEMO_bias_2D_2t_delta_proj_hist.png .. :name: ifs-nemo_proj_2t .. Spatial maps of the change in annual mean 2 m air temperature relative to the 1990–2014 climatology. The observed change (2015–2024) is shown from ERA5, alongside the corresponding IFS-NEMO simulated change for the same near-present period (2015–2024) and for a future period (2025–2049). Global mean changes (Δ_glb) are indicated for each panel. .. The projected changes in annual mean precipitation from IFS-NEMO (:numref:`ifs-nemo_proj_precip`) broadly follow the pattern of observed recent changes derived from MSWEP, both showing a modest global mean increase relative to the 1990–2014 baseline, though the model slightly underestimates the magnitude of the near-present signal (Δ_glb of 9.91×10⁻³ mm d⁻¹ in IFS-NEMO versus 1.41×10⁻² mm d⁻¹ in MSWEP over 2015–2024). Spatially, both the observed and simulated near-present anomalies are characterised by a wet-gets-wetter and dry-gets-drier pattern, with precipitation increases along the ITCZ and subtropical drying, though the model signal is somewhat weaker in amplitude. In the 2025-2049 period, IFS-NEMO projects an amplification of these tendencies, with a global mean increase of 2.79×10⁻² mm d⁻¹, and a clearer emergence of regional contrasts, with enhanced drying in the subtropics and increased precipitation in the deep tropics, consistent with the expected thermodynamic response to continued warming. .. .. figure:: ../../../../evaluation/levante/Fig/IFS-NEMO_bias_2D_pr_delta_proj_hist.png .. :name: ifs-nemo_proj_precip .. Spatial maps of the change in annual mean total precipitation relative to the 1990–2014 climatology. The observed change (2015–2024) is shown from MSWEP, alongside the corresponding IFS-NEMO simulated change for the same near-present period (2015–2024) and for a future period (2025–2049). Global mean changes (Δ_glb) are indicated for each panel