Www.pol.ac.uk Initial results from a multi-year simulation of the NW European shelf seas Roger Proctor, Jason Holt, Graham Tattersall, Sarah Wakelin Proudman.

www.pol.ac.uk

Initial results from a multi-year simulation of the NW European shelf seas

Roger Proctor, Jason Holt, Graham Tattersall, Sarah WakelinProudman Oceanographic Laboratory

Liverpool, UK

Talk structure

• Why do it?

• What’s happened in last 50 years?

• The model study

• Initial results

• Summary

Why do it?

• Climate change is upon us• Need to understand its consequences• History best clue to the future• Need limits of predictability for future climate

change• Need baseline values against which to deduce

trend changes (for WFD* as well as planning)• Modelling challenge

*EC Water Framework Directive

What’s happened in the last 50 years

• Evidence from – weather– sea level (PSMSL)– Temperature & salinity (ICES database)– Continuous Plankton Recorder

ERA-40 period

After Hurrell: www.cgd.ucar.edu/cas/jhurrell/indices.html

Effects of NAO Lerwick, NW Scotland (winter)

1982-2000

94/95

95/96

+

- Extracts from the IACMST report – UK Marine Waters 2004: Marine Processes and Climate

http://www.oceannet.org/medag/reports/IACMST_reports/MCP_report/index.htm

Cypris station Irish Sea

Sea Level changes

• Absolute rise ~ 1mm/y

• Relative rise (mostly increasing, except N coasts)

• decrease in the rate of rise in the 20th Century

• no long-term trends in tides

Wakelin et al. (2003) have shown that winter-mean (December to March) sea levels and the NAO Index are significantly correlated over much of the northwest European shelf.

MECN partners

Marine Environmental Change Network (MECN)

http://www.mba.ac.uk/MECN/about.htm

• Establish a network to measure environmental change in marine waters

• Maintain and enhance existing long-term research programmes

• Restart important discontinued long-term research programmes

• Deliver and interpret long-term and broad scale contextual information to inform water quality monitoring

• Demonstrate the benefits of preserving and networking long-term time series programmes

• Provide information to policy makers and other end-users to enable them to produce more accurate accurate assessments of ecosystem state

• Have a clearer understanding of factors influencing change in marine ecosystems.

MECN Long time series measurements

TireePassageW Scotland

Temperature – SST Atlantic

ICES

FRS

TemperatureSST - shelf

Temperature – SST North Sea

Fair IsleN Scotland

FRS

ICES

NAO accounts for40-50% of winter SSTvariability in N Sea(Loewe, 1996)

Trends in Sea Surface Temperature

North Sea SSTNominal position

Data Span

Trend (°C/decade)

Spurn Point 53.5N 00.0E 1966-2002 + 0.17

Humber LV 53.5N 00.0E 1880-1989 + 0.01

Newarp LV 53.0N 02.0E 1880-1986 + 0.02

Lowestoft 52.5N 02.0E 1966-1991 + 0.16

Sizewell 52.0N 01.5E 1967-2002 + 0.77

Bradwell 52.0N 01.0E 1964-2001 + 0.15

English Channel, Celtic Sea and Bristol Channel SST

Dover 51.0N 01.5E 1926-2001 + 0.09

Eastbourne 51.0N 00.0E 1892-2002 + 0.06

Shoreham 51.0N 00.0E 1966-1997 + 0.31

E1 50.0N 04.5W 1921-1987 - 0.13

Sevenstones LV 50.0N 06.0W 1881-1986 + 0.04

St Gowan LV 51.5N 05.0W 1953-1987 - 0.17

Swansea 51.5N 04.0W 1976-1997 + 0.58

Irish Sea SST      

Port Erin, Isle of Man 54.0N 04.5W 1904-2003 + 0.07

Temperature – NBT winter (IBTS)

Temperature• No clear trend in summer SST in the eastern North Atlantic since the 1950s, but a warming in winter SST since the early 1990s.

• SST at the Continental Shelf Edge warmed between 0.12°C and 0.29°C over the past century. Temperatures in the Rockall Trough were relatively low in the early 1990s but then increased. The highest temperatures reached in the 1990s were similar to those in the 1960s.The Faroe Shetland Channel has become warmer over the last 40 years, with temperatures rising at a rate of approximately 0.3°C per decade from the late 1960s minimum.

• Waters around the UK have been warming since the 1980s, with the trend more pronounced in the southern North Sea and the Irish Sea (between 0.5°C and 1.0°C per decade) than elsewhere (between 0.0°C and 0.5°C per decade).

• There is a warming trend in winter and summer SST averaged over the northern North Sea since the early 1980s, with a warming of about 1°C and 0.5°C respectively. North Sea winter bottom temperatures increased by about 0.3°C and 0.6°C per decade since a cool period in the late 1970s.

• Irish Sea annual mean SST increased by about 0.7°C over the last 100 years. Winter SST from 1950 to 2002 shows a clear warming since the 1980s. An apparent cooling in summer SST since the 1980s may be due to sparse data.

Salinity – SSS Atlantic

ICES

Salinity – SSS North Sea

Fair IsleAlmost cyclical variability since the end of the GSA in the 1970s

GSA

ICES

Salinity – SSS shelf

Salinity

• Atlantic waters and adjacent shelf areas had low winter and summer sea surface salinity (SSS) in the mid-late 1970s (associated with the passage of the Great Salinity Anomaly (GSA)), followed by three decades of large inter-annual variability.

• Salinity records from the Faroe Shetland Channel and the Ellett line indicate a recent trend to high salinity.

• SSS averaged over the northern North Sea from 1950 to 2002 shows decreasing salinity since the 1970s and is reflected by observations at fixed locations in the Fair Isle Current and the North Sea fishing grounds.

• There is no discernible trend in mean SSS in the English Channel from 1900 to the early 1980s.

• SSS averaged over the Irish Sea from 1950 to 2002 shows a decrease in both winter and summer.

Reid et al. (1998) Nature

Long-term changes at the base of the food web

• Phytoplankton abundance in North Sea and NE Atlantic

The Continuous Plankton Recorder

Biogeographic Changes in the Northeast Atlantic

Warm temperate slope species

The Continuous Plankton Recorder

Beaugrand et al. Science 2002

The model study

• Using POLCOMS – POL Coastal Ocean Modelling System

• Configured for ocean/shelf (AMM)

• ERA-40 surface flux (SLP, 10m wind, bulk heat flux, E-P flux)

• Daily river flux (mix of climatological and measured)

• FOAM repeating annual cycle (2000/1)

POLCOMS• 3D Shallow Water equations

– Horizontal finite difference discretization on an Arakawa B-grid– Spherical polar co-ordinates– Terrain following co-ordinates - horizontally varying S levels– Forward Time-Centred Space scheme in the horizontal– Equations split into depth-mean and depth-fluctuating components

• Horizontal Advection

– Piecewise Parabolic Method (PPM) for accurate, conservative representation of sharp gradients, fronts, thermoclines etc.

• Horizontal pressure gradients

– calculated by interpolation onto horizontal planes, improving numerical accuracy of S co-ordinates over steep topography

• Vertical diffusion

– Mellor-Yamada-Galperin level 2.5 turbulence closure scheme

– GOTM (k-ε), Canuto stability functions

N Atlantic FOAM 1/9o

(T,S, ζ, Q)

Physics (T,S, ζ, U,V) No data assimilation or relaxation to climatologyModules for spm, (light, nutrients, biology) - ERSEM

AMM-12kmAtlantic Margin Model

Tides(9C, ζ, Q)

ECMWFERA-40

1957-20021 degree6-hourly

SLP, 10m wind, 2m AT, RH, CC,

E-P

River dischargeDaily

323 rivers

Lack of ocean time varying bc means features likeGSA will not be simulated

Initial results

• Point series comparisons (Cypris, …)

• ICES CTD database comparisons– Errors– Seasons– Trends

Cypris station – Irish Sea

Cypris station – Irish Sea Irish Sea only (Holt&Young06)

OB (Sevenstones) station – Celtic Sea

OB (Sevenstones) station – Celtic Sea

ICES CTD database analysis

Total61,000 CTDs

20,000 this period (95-99)

Temperature errors – Oct-Dec 1995-99

Surface Near-bed

Salinity errors seasonal 1995-99

WinterSummer

Surface Near-bed

SST trends

Like-for-like trends ‘81-99 (min of 8 years & 5 obs/season / 1o cell)

Summary

• 40 year simulation of NW European shelf without data assimilation successfully reproduces many of the trends seen in the ICES and MECN observed data

• Model errors highest in spring (T) and NCC (S)• Effect of ‘climatic’ oceanic circulation not

apparent• Model agrees better with obs when like-for-like

measure used. Spatial variability allows error estimates to be made

• Just the beginning, also started ecosystem runs

POLCOMS-ERSEM 15yr (1986-2001)

Mean chl errors mg/m3 (against ICES data)

NBT trends