Skip to: [content] [navigation]

The purpose of these meetings was to present analysis of our most recent 100 year 1860 control run of ESM2p1 to demonstrate the qualitative similarity of it's climate with CM2p1, and to present the current status of the land and ocean carbon cycles as ESMDT moves from a focus on climate issues to carbon issues. The run is on the database as eperiment: ESM2p1 -> ESM2.1U_Control-1860_lm3_O_v1.

ESM2.1 Development: Climate and carbon

Agenda:
I. Background and physical climate (Ron Stouffer; presentation can be downloaded here)
II. Ocean carbon cycle (John Dunne; presentation can be downloaded here)
III. Vegetation and carbon (Elena Shevliakova; presentation can be downloaded here)

I. Background and Physical Climate

A) Experiment
ESM2p1 -> ESM2.1U_Control-1860_lm3_O_v1 (database)
/archive/ms2/esm2.1/ESM2.1U_Control-1860_lm3_O_v1 (output)

B) Background: Mixed-Layer Model
CM2.1 - few changes as possible to climate.
Swapped LM2 with LM3V.
Tuning and problems (physical climate and biosphere).
Two water leak issues resolved: MLM-LM2 ans LM2-LM3V.
Computing new qflux adjustments for LM3V with Omsk base code in progress.
Overall, happy with LM3V.

C) ESM2.1
Everything that was originally slated is included. Any additional items will go into ESM3.
CM2.1 1860 control.
100 year integration completed.
Atmosphere carbon is held fixed and does not interact with land or ocean.
Overall:
-happy with the run
-physical climate is similar to CM2.1
-carbon drifts need tuning (soil and ocean carbon)
-it is possible to get to the goal of drift less than 10 ppm pCO2/century

D) Comparison of the physical climate
----------------------------------
(Compared experiment with CM2.1U_Control-1860_D4)
Major differences between ESM2.1 and CM2.1 physical climate are:
1) swapping LM2 with LM3V
2) ocean color changes
Overall, physical climate looks similar for CM2.1 and ESM2.1.
Global mean time series plots of t_ref, ice extent, SSTs, net radiation TOA, heat flux into the ocean, volume annual mean ocean temperature are similar and show no big trends.
t_ref average difference maps show springtime warming and summertime cooling in the high northern latitudes. This makes the early spring melt bias worse. For the most part, the rest of the maps show no changes in the seasonal cycle.
The precipitation changes are not outrageously large.
NINO3 SST spectra (obs are for the 20th century) shows an improvement (possibly because of ocean colors).

E) Runtime
3 hours/year on 156 processors
split: 96 ocean / 60 atmosphere

II. Ocean carbon cycle

A) CO2 variability
Mauna Loa CO2 variability is approximately 6 ppm.
286 ppm corresponding to 493 Pg/C total corresponds to a net seasonal change of 5.4 PgC over 6 months = 1 Pg/month.

B) Ocean biogeochemistry development
Highlights:
-TOPAZ
-river flux assessment showed no effect on productivity
-model more phosphorous limiting than it should be

C) Model analysis
Model is initialized with preindustrial CO2.
SEAWIFS data for chlorophyll compared to model on linear and logarithmic scales.
On the linear scale, low chlorophyll in gyres and high chlorophyll in upwelling zones.
Average chlorophyll is just about right but too low in high chlorophyll regions and too high in low chlorophyll regions.
Phosphate is too low.
Plot of surface chlorophyll vs. latitude shows bias in north Atlantic where you do not get chlorophyll.
Plot of surface nitrate vs. latitude shows movement of high nitrate region to the north. This happens immediately.
Comparison of global profiles compared with initialized observations shows too high surface nitrate and too low subsurface nitrate, phosphorous is right on and oxygen is too low.
The CO2 flux from Takahashi data (2002 map) and the model are very different.
The model has too much convection in the Southern Ocean and too much mixing in the north Atlantic.
Air-to-sea CO2 flux time series plots show a great amount of variability on the annual scale.
At year 120:
-N2 fixation = 473.8 TgN/yr
-water column denitrification = 584.7
-sediment denitrification = 150.2

D) Biological parameter tuning
The model is very sensitive to biological parameter tuning.
After tuning, at year 10:
-the open ocean dynamic range of chlorophyll is better
-nitrate is too high
-phosphate is good
-chlorophyll is lower but resolving open ocean chlorophyll
-N2 fixation = 93.3 TgN/yr
-water column denitrification = 51.0
-sediment denitrification = 93.8
Note: chlorophyll stabilizes in 3-4 years.

III. Vegetation and carbon

A) Model Background
LM3 and LM3V are different models with respect to soil.
All vegetation is simulated (not prescribed).
LM3V has 5 vegetation types that are simulated according to climate:
-warm grasses
-cold grasses
-tropical forests
-deciduous trees
-coniferous forests
Warm and cold grasses have different photosynthetic pathways.
Vegetation and soil carbon are stable.
Total vegetation carbon is within estimates but on the low side.
We need to examine seasonal cycle, inter-annual variability and vegetation CO2 sensitivity further.

B) Model vegetation type
Overall, the model simulated vegetation types are pretty good except for the Amazon, which is represented by warm grassland rather than tropical forest.
The Amazon region needs some tuning to get a seasonal forest.
The major issue is that the vegetation types have a sharp boundary between grasses and trees.

C) Biomass
The time series plot of vegetation biomass does not show much drift in the global values.
Annual biomass, compared with ORNL/DAAC observations, shows that the overall Northern Hemisphere biomass pattern is not bad.
The lack of biomass in the Amazon is a major problem (this is also a problem with carbon).

D) Soil carbon
The modeled soil carbon, compared with ISRIC observation data, is higher in the northern high latitudes:
-introduced snow on trees in the model
-soil is cold and biological processes are small
-small respiration rates
Total soil carbon is low. We need to equilibrate to a higher value.
The soil carbon is 880 Gt after equilibrium.
LM3V does not represent a vertical structure of soil carbon. However, LM3 will have a vertical structure.

E) NPP & carbon from fires
NPP observations are from the CASA model. This is a data driven (prescribed) model that is very well tuned.
There is a pretty high NPP rate in the Amazon considering it is represented by grassland.
The spatial pattern of NPP is pretty close to the observations.
The map of carbon from fires shows that there might be too much fire in the Amazon (?).

• About us
  • Welcome
  • Visiting GFDL
  • Overview
  • Tour
  • Gallery
  • People
  • Employment
  • Help
  • Notices
• Research
  • Atmospheric Physics and Chemistry
  • Biospheric Processes
  • Climate Diagnostics
  • Climate Dynamics and Prediction
  • Oceans and Climate
  • Weather and Atmospheric Dynamics
• Products and Services
  • Flexible Modeling System
  • Our Data Portal
  • Our GForge
  • Technical Services
  • Visualization
• Reference
  • Our Online Bibliography
  • Our Meetings and Seminars
  • Our Library
• GFDL Only
  • External forms and links
  • External web mail
  • Help Desk
  • Internal Administrative Services
  • Internal Technical Services
  • NOAA's Hot Items
  • webTA Login

smaller bigger reset