# AMOC variabilities in PI, Paleo and historical runs
Three different types of simulation are analysed:
* Fully-coupled (BCCR_Fast version, NBF1850OC_ORGANIC1)
* Forced by NCEP (BCCR_Fast version,NOIIANOC_mr04_hist2, Schwinger)
* Forced by 20CR (NOI20C_sr10m60d50, differen model version)
* Fully-coupled paleo simulations (Guo and Bentsen)
Outline:
* Frequency/Period analysis of AMOC time series
* Decompoiste AMOC time series in to RapidMOC components
* AMOC in forced-simulations
* Compoisite of ocean and atmos circulation field with strong and weak AMOCs.
* EOF analysis of MOC, SPG, and PSL
* Cross-correlations between atmos/ocn cirulation with AMOCs
* Regression and Trend analysis
## AMOC timeseries, and frequency analysis
### AMOC timeseris
|Fully-coupled | NCEP forced | 20CR forced |
| --- | --- | --- |
|  |  |  |
AMOC in paleo-simulations
* Marine isotope stage 3b (mis3b)
* Eemian periods: 115ka, 120ka, 125ka, 130ka
* Control preindustrial simulation: 1000-2000 years
* Model grid: tnx1v1 (BCCR Fast version)
(20 years running mean)
### Power Spectrum
| Power spectrum | Wavelet spectrum |
| --- | --- |
|  | |
| 20CR forcing | NCEP forcing 1 | NCEP forcing 2|
| --- | --- | ---|
| total | total | total|
|  | |  |
| Eemian 115ka | Eemian 120ka |
| :---: | :---: |
|  | |
| Eemian 125ka | Eemian 130ka |
|  |  |
| MIS 3b | PI control 1001-2000 |
|  |  |
**Power spectrums of randonly selected 100 years timeseries from the 2000 years PI control run**
| Years 1401-1500 | Years 1701-1800 |
| --- | --- |
|  |  |
| Years 2001-2100 | Years 2301-2400 |
| --- | --- |
|  |  |
| Years 2601-2700 | Years 2901-3000 |
| --- | --- |
|  |  |
---
## AMOC components: FC, Ekman and UMO
|Components | model output vs components |
| --- | ---|
| | |
---
### Power spectrum of AMOC components
**PI control 1001-3000 (left figure), and its power spectrum (right figure)**
| Power spectrum | Filtered timeseries |
| :---: | :---: |
| Florida Current | |
|  |  |
| Ekman transport | |
|  |  |
| UMO transport | |
|  |  |
|  | |
## AMOC in forced simulations
### Compared to RapidMOC
| 20thC forcing | NCEP forcing 1 | NCEP forcing 2 |
| --- | --- | --- |
| AMOC from 1871-2009 | AMOC from 1948-2016 | AMOC from 1948-2016 |
| Model grid: tnx1v1 (not BCCR Fast version) | Model grid: mnp2 (old MICOM-ICE alone stand version) | Model grid: tnx1v1 (BCCR Fast vesion, by Jörg Schwinger) |
| | ||
| 20CR AMOC, 1995-2014,without SSS relax |
| |
---
### Power spectrum
| 20CR forcing | NCEP forcing 1 | NCEP forcing 2|
| --- | --- | ---|
| total | total | total|
|  | |  |
|Florica Current | | |
| | |
|Ekman transport | | |
| | |
|Upp-mid ocean transport| | |
| | |
## AMOC COMPOSITE: fully-coupled
**AMOC from PI years 10001-3000**
(Gray-dashed lines are $$\pm 1.5 \sigma$$ AMOC)
| AMOC@26.5N ($$1.5\sigma$$) | AMOC PC1 ($$1.0\sigma$$) |
| --- | --- |
|| |
| |
||
| mmflxd lowpass diff (yrsmax-yrsmin)
| |
| mmflxd highpass diff (yrsmax-yrsmin)|
||
||
|  | 
|  | 
|  | 
|  | 
|  | 
| March MLD|
|  | |
|  | |
|PSL, TS (November to April)|
|  | |
|  | 
Summary:
* Stronger overturning at about 1500 m, also more AABW water northward
* Stronger SPG and North Atlantic Current
* Warmer and Saliner water in the eastern Subpolar North Atlantic and the Greenland Sea
* Heat loss anomaolies in the NAC and the Nordic Seas, and heat gain anomaly in the subpolar North Atlantic
* MLD changes consistent with hflx, but is small.
* Fice change is consistent with sst change
* Weaker zonal wind stress in the subpolar North Atlantic and the Southern Ocean, stronger zonal wind stress in the subpolar North Pacific
COMMENT: Bentsen2004 (figur19) shows positive NAO anomaly and warm NA/cool spg pattern. Here, shows negative NAO pattern (also wind stress)
------
For AMOC EOF PC1 composites
| yrsmax - mean | yrsmin - mean |
| --- | --- |
|  | 
|  | 
|  | 
|  | 
|  | 
|  | 
|  | 
PSL, TS (November to April), for EOF PC1
| yrsmax - mean | yrsmin - mean |
| --- | --- |
|  | |
|  | 
PSL, TS (November to April), for AMOC@26.5
| yrsmax - mean | yrsmin - mean |
| --- | --- |
|  | |
|  | 
## AMOC COMPOSITE: 20CR-forced
**Strong vs Weak AMOC, years 1900-2009**
(Gray-dashed lines are $$\pm 1.0 \sigma$$ AMOC)
| AMOC@26.5N | AMOC PC1 |
| --- | --- |
| |  |
| yrsmax-yrsmin | yrsmax-yrsmin |
| | 
| | 
| | 
| | 
| |  |
## EOF Analysis
### EOF of MOC
**NBF1850OC_ORGANIC1**
| PC1 | PC2 | PC3 |
| :---: | :---: | : ---: |
| |  | |
**NOIIANOC_mr04_hist2**
| PC1 | PC2 | PC3 |
| :---: | :---: | : ---: |
| |  | |
**NOI20C_sr10m60d50**
| PC1 | PC2 | PC3 |
| :---: | :---: | : ---: |
| |  | |
Power spectrum
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  |  |
COMMENT:
* EOF1 represents whole Atlantic basin varaibility; EOF2 represents North and South Atlantic variability; EOF3 represents subtropical gyre and subpolar gyre variability
* How to extract other features connecting this?
* Intepretation of power spectrum?
* Bentsen2004, EOF explains 79% and 57% variance. von Storch et al., 2000, in the four 1000-yrs integration, explains 41-56% variance.
---
### EOF of SPG
NBF1850OC_ORGANIC1,1001-3000
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  | |
NBF1850OC_ORGANIC1, 2890-3000
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  | |
NOI20C_sr10m60d50
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  | |
NOIIANOC_mr04_hist2
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  | |
Power spectrum
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  |  |
COMMENT:
* EOF1 represents North Atlantic Current varaibility; EOF2 represents subtropical gyre and subpolar gyre variability; EOF3 represents variabilities of western and eastern basin
* EOF2 of spg is connected to MOC PC3? No in-phase correlation, but opposite correlation at immediate 1 year lead and lag.
---
### EOF of PSL
| PC1 | PC2 | PC3|
| :---: | :---: | :---:|
| |  ||
Power spectrum
| PC1 | PC2 | PC3 |
| :---: | :---: | :---: |
| |  |  |
COMMENT:
* Connections with MOC, SPG and covectivive activities?
## Correlation Analysis
### AMOC PC1
| AMO | MLD_LS | MLD_IS | MLD_GIN|
| --- | --- | --- |
|  |  | | 
| NAO1 | NAO2 | NAO3 |
| --- | --- | --- |
|  |  | |
| NHICEA | SHICEA |
| --- | --- |
|  |  |
| SPG11 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
COMMENTS:
* SH sea ice positive lead AMOC1 by about 20a. NH sea ice is more complicate lag/lead?
---
### AMOC PC2
| SPG11 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
| NHICEA | SHICEA |
| --- | --- |
|  |  |
### AMOC PC3
| SPG11 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
### AMOC@26.5N
| MLD_LS | MLD_IS | MLD_GIN|
| --- | --- | --- |
|  | | 
### AMOC@45.0N
| MLD_LS | MLD_IS | MLD_GIN|
| --- | --- | --- |
|  | | 
COMMENT:
* MLD_GIN is correlated to AMOC265 and AMOC450, because of overflow water?
* AMOC3 and SPG2 have similar spatial pattern, and interesting temporal correlations
### MLD
| LS vs IS | LS vs GIN | IS vs GIN |
| --- | --- | --- |
|  |  | |
COMMENT:
* These convective indices are derived as Bentsen et al., 2004;
* LS and IS is in phase, with IS leads the LS for about 1 year (consistent with Bentsen2004); GIN lead LS by 10-15 years (Bentsen2004, GIN and LS in opposite phase, linked to NAO as observations suggest)
---
MLD LS vs ...
| SPG1 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
MLD IS vs ...
| SPG1 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
MLD GIN vs ...
| SPG1 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
MLD_xx vs NAO1
| LS | IS | GIN |
| --- | --- | --- |
|  |  | |
COMMENT: BENSTEN2004 shows positive simutaneous correlation between NAO1 with LS and IS, correlation > 0.4; which GIN is anti-correlated with NAO1. In this simulation, LS and IS is positively correlated, but only with efficient as 0.2. GIN is not negative.
### FC
| DrakePass | spectrum| vs FC|
| --- | --- | --- |
||  ||
| NHICEA | SHICEA |
| --- | --- |
|  |  |
| MLD PC1 | MLD_GIN | MLD_IS | MLD_LS|
| --- | --- | --- | --- |
|  |  |  | |
| NAO 1 | NAO 2 | NAO 3 |
| --- | --- | --- |
|  |  | |
| SPG 1 | SPG 2 | SPG 3 |
| --- | --- | --- |
|  |  | |
Comment:
* FC is correlated with 1st and 3rd component of SPG
* Note the SPG1 pattern in the 1000-3000 yr run is not present in other short simulations; SPG1 reflects the Florida current and NAC, SPG2 reflects STG and SPG, SPG3 reflects westward/eastward shift of NAC?
* FC transport is positively correlated with all (3) PC of NAO
### Ekman
| NAO 1 | NAO 2 | NAO 3 |
| --- | --- | --- |
|  |  | |
Comment: Ekman transport is negatively correlated with NAO?
| SPG 1 | SPG 2 | SPG 3 |
| --- | --- | --- |
|  |  | |
### UMO
| MLD1 | MLD LS | MLD IS | MLD GIN |
| --- | --- | --- | --- |
|  |  | | |
| NAO1 | NAO2 | NAO3 |
| --- | --- | --- |
|  |  | |
| SPG1 | SPG2 | SPG3 |
| --- | --- | --- |
|  |  | |
## Regression
### TS vs AMOC265
| 90a high-pass | 90a low-pass|
| --- | --- |
| ||
| ||
COMMENT:
* Park and Latif 2008 (figure 4ab) shows negative regression in the southern ocean on multicentennial scale and positive on multidecadal scale variability in the North Atlantic region. Here not.
* This regressions does not pass significance test at all
### PSL vs AMOC265
| 90a high-pass | 90a low-pass|
| --- | --- |
| ||
| ||
## Trend
### AMOC
| 8a | 10a | 12a|
| --- | --- | --- |
|  |  |  |
|20a | 50a | 100a |
| --- | --- | --- |
|  |  |  |  |
COMMENT: The RAPID AMOC Index at 26°N shows a decline trend ( 0.265 Sv/yr) over the 12-year period of 2004–2015(Yan et al, GRL2018)
10 year trend
|mis3b | 115ka | 120ka |
| --- | --- | --- |
|  |  | 
| 125ka | 130ka | ctrl02|
|  |  | |
12 year trend
|mis3b | 115ka | 120ka |
| --- | --- | --- |
|  |  | 
| 125ka | 130ka | ctrl02|
|  |  | |
20CR
| window 12 | window 8|
| --- | --- |
| | |