SAP

A tag for the old pages ported over from SAP (other than the main page).

Ocean cooling geoengineering experiment

in
Summary

This experiment uses exactly the same setup as the Seasonal Attribution Experiment but instead looks at the effect of a sudden artificial cooling of the ocean surface.

Status

  • Some simulations used in this experiment were run previously under the Seasonal Attribution Experiment.
  • The first simulation dedicated to this experiment was started on 11 December 2008.
  • The final simulation was started on 11 March 2009. Thanks to all who volunteered their computers!
  • This experiment will enter the analysis phase when all simulations are completed.
Motivation

On the one hand people are becoming increasingly concerned about climate change caused by human emissions of greenhouse gases, but on the other hand they are also realising both that a lot of greenhouse gases have already been emitted and that reducing future emissions substantially could be quite difficult. One idea for how to deal with this is to actively cool the ocean surface, supposedly to "natural" temperatures.

For instance, James Lovelock, of Gaia fame, and Chris Rapley, director of the Science Museum in London, have proposed inserting large vertical pipes in the ocean which would passively use the energy of waves to force cool water to rise from the deep ocean to the surface, thereby cooling the surface. (It could also get more carbon dioxide to be absorbed from the atmosphere and stored in the the deep ocean, but we will only worry about the cooling effect here.) This experiment will examine the effects of any method that would cool the ocean surface but leave other conditions the same. Will inland areas be cooled? Will rainfall change?

Method

This new experiment consists of two components. The first is a set of simulations of the year 2000 (actually March 2000 to April 2001) using the actual conditions experienced in year 2000. These conditions include observed values of the atmospheric concentrations of greenhouse gases and other gases and aerosols, as well as observed ocean surface temperatures and sea ice coverage. Simulations differ in the initial weather state set for midnight on 1 March 2000, such that each simulation gives an example of what the weather could have been like in year 2000. Altogether, these simulations give an estimate of the climate for that year. These simulations were completed as part of the Seasonal Attribution Experiment.

The second component involves re-running these simulations, but with ocean surface temperatures cooled (and sea ice extent expanded) according to an estimate of what year 2000 conditions would have been had we not been emitting greenhouse gases (note this is not quite "natural" because we would still have been doing other things, like emitting sulphate aerosols). These ocean conditions were estimated as part of the "non-industrial" component of the Seasonal Attribution Experiment. (We will use only only one of the 40 estimates used in that experiment though, the one with the pattern estimated from the HadCM3 model and the amplitude as the 6th decile estimate based on a comparison of that pattern with the observed record.) Note though that unlike those "non-industrial" simulations, these simulations will use the true observed greenhouse gas concentrations.

Investigators
Dáithí Stone (University of Cape Town)
Tolu Aina, Milo Thurston (University of Oxford)
Pardeep Pall (ETH-Zürich)

Sudden carbon sequestration experiment

in

Summary

This experiment uses exactly the same setup as the Seasonal Attribution Experiment but instead looks at the effect of a sudden sequestration of greenhouse gases.

Status

  • Some simulations used in this experiment were run previously under the Seasonal Attribution Experiment.
  • The first simulation dedicated to this experiment was started on 2 May 2008.
  • The final simulation was started on 11 December 2008. Thanks to all who volunteered their computers!
  • This experiment will enter the analysis phase when all simulations are completed.

Motivation

On the one hand people are becoming increasingly concerned about climate change caused by human emissions of greenhouse gases, but on the other hand they are also realising both that a lot of greenhouse gases have already been emitted and that reducing future emissions substantially could be quite difficult. Therefore, an idea is gaining favour that we should actively extract greenhouse gases from the atmosphere and sequester them somewhere out of harm's way. This should preferably be done very soon, the faster the better.

But is faster really better? Consider the extreme case that we were suddenly able to extract the entire anthropogenic greenhouse gas contribution from the atmosphere, thus returning to a "natural" state. In this scenario, we would expect to find that the land would have cooled nicely (away from the coasts anyway), which seems ideal. But the ocean would have been much slower to cool, so we also think that the warm ocean/cool land contrast would have produced more severe storms and rainfall events for some time afterward, which is not so ideal. This experiment will examine this extreme case.

Method

This new experiment consists of two components. The first is a set of simulations of the year 2000 (actually March 2000 to April 2001) using the actual conditions experienced in year 2000. These conditions include observed values of the atmospheric concentrations of greenhouse gases and other gases and aerosols, as well as observed ocean surface temperatures and sea ice coverage. Simulations differ in the initial weather state set for midnight on 1 March 2000, such that each simulation gives an example of what the weather could have been like in year 2000. Altogether, these simulations give an estimate of the climate for that year. These simulations were completed as part of the Seasonal Attribution Experiment.

The second component involves re-running these simulations, but with greenhouse gas concentrations reduced to 1900 levels. In effect, this supposes that in 1999 we had suddenly found a very, very cheap and fast way of removing all of the anthropogenic greenhouse gases from the atmosphere and that this method was adopted in early 2000. Because the ocean responds slowly, we can plausibly suppose that this did not affect the ocean over the 2000-2001 period of the simulations, so we will leave the ocean surface temperatures and sea ice coverage as observed in 2000 (and as in the first set of simulations).

Investigators
Dáithí Stone (University of Oxford; University of Cape Town)
Tolu Aina, Milo Thurston (University of Oxford)
Pardeep Pall (ETH-Zürich)

The UK Autumn 2000 floods

in
  • Introduction
  • Antecedent conditions
  • Associated meteorology
  • References Flooding in Nottingham during Autumn 2000


    Introduction

    The United Kingdom floods of October and November 2000 occurred during the wettest autumn in England & Wales since records began in 1766, with a total of 503mm of rain, exceeding the previous maximum by almost 50mm (Marsh and Dale, 2002), and resulting in almost double the average (1961-1990) seasonal precipitation (Alexander and Jones, 2001).

    Just under 10,000 properties, 58% of which had no flood defences, were flooded at over 700 locations and there was widespread disruption to road and rail services. Train services were cancelled, major motorways closed, power supplies disrupted and 11,000 people were requested to evacuate their homes. The total costs are of the order £ 1.0 bn (EA, 2001). Figure 1 illustrates the nationwide scale of the events, in terms of properties flooded. The most heavily affected areas were the South East, North and South Wales and Yorkshire, and Figure 2 shows these regions received well above the climatalogical (1961-90) September-December rainfall average, as did most others.

    Fig 1. Flood events by number of properties affected for the Autumn 2000 floods in England and Wales.Taken from EA (2001) Fig 2. September-December 2000 regional rainfall (in mm), as a percentage of the 1961-1900 average.Taken from Marsh and Dale (2002)


    The national picture was of persistent, widespread heavy rain, rather than very strong localized events, although these did also occur. Synoptic data for the period showed that the October-November period was dominated by persistent and repeated frontal depressions, with broad bands of rain enclosing notable high-intensity rain cells across the British Isles. The resultant flooding was hydrologically complex with large local and regional variations in its severity and was the cumulative effect of the series of exceptional rainfalls affecting different parts of the country at different times and in many cases more than once. Some areas flooded two or three times in the autumn and some even five times during 2000 (DEFRA, 2001).

    Back to top ^

    Antecedent conditions and main flood events

    Rainfall is most hydrologically effective over the November-April period when evaporation losses are very modest and soils generally moist. The rainfall for this period in the three preceding years had been above the 1961-1990 average (and largely explains the transformation in water resources since the drought conditions in the mid-1990s), but the 1999-2000 recovery was by far the greatest and most rapid. In 2000, particularly unsettled weather patterns in April and May ensured that soils did not begin to dry out until the later spring, prolonged the aquifer recharge season, and meant that flows in rivers draining the most permeable catchments remained healthy throughout the summer, even withstanding the below average summer rainfall (CEH, 2001). Then during the Autumn, the main periods of prolonged rainfall were the 29-30th October and 2nd, 5th, 6th November for Wales, South West, Central and Northern England; 9th-12th, 15th-18th October, 5th and 6th November for Southern England (EA, 2001). There were distinct phases to the flooding associated with these periods, summarized here from Marsh and Dale (2002):

    The second half of September and early October saw localized and short-lived events with heavy frontal rainfall resulting in early spate conditions in Northern Ireland and western Scotland, and urban flooding in parts of southern England in response to intense convective rainfall. This heralded severe flooding in the South East in the second week of October associated with thunderstorms forming along a near-stationary front, causing flooding in some impermeable South East catchments. Storms around the 19th October substantially increased the number of saturated or near-saturated catchments and extended the area which was at high risk of flooding. River flows significantly increased at the end of the month following widespread and persistent rainfall, and a low-pressure system produced rainfall totals of 20-40 mm across most of southern Britain. This was followed by the passage of an extremely active system which tracked north-east from the Bristol Channel to the North sea. Flood plain inundations were especially extensive in Yorkshire, and by the second week of November, severe flood warnings (issued by the Environment Agency) covered rivers across most of England.

    Back to top ^

    Associated meteorology

    Blackburn and Hoskins (2001) find that the Autumn 2000 season was characterized on average by a displacement of the Atlantic Jet-stream eastward from its climatological position, with the region in which the air exits from the jet being more marked than usual over western Europe and displaced south of the UK from its normal elongated position near 55oN (Figure 3). This brought intense systems into the area, where they slowed, repeatedly leading to prolonged precipitation events.

    Fig 3. See original caption for details.Taken from Blackburn and Hoskins (2001)


    Furthermore, it was argued that the anomalous jet-stream was associated with a train of anomalies in geopotential height extending from the subtropical Atlantic across Eurasia, with a cyclone over the UK and strong anticyclone over Scandinavia (Figure 4, left). The wave-like pattern was indicative of quasi-stationary Rossby waves, and the persistence and coincidence of this pattern with the mid-September to mid-December wet period in the UK were notable. Moreover, a correlation of autumn England & Wales precipitation with global geopotential height, for the preceding 42 years produced a remarkably similar pattern to the anomalous pattern in Autumn 2000 (Figure 4, right), and very similar to the Scandinavia or Eurasia-1 pattern. Correlations in excess of 0.4, significant at the 99% level, were found in the four peaks of the pattern over Eurasia.

    Fig 4. See original caption for details.Taken from Blackburn and Hoskins (2001)


    Blackburn and Hoskins (2001) concluded that wet UK Autumns have not only been associated (unsurprisingly) with low pressure in situ, but also with the above pattern of anomalies stretching over a wide area. In proposing a tropical catalyst for this pattern, in the form of anomalous deep convection overlying tropical sea surface temperature (SST) anomalies, they then identify idealized modelling studies showing upper-troposphere horizontal divergence above a region of converging tropical convection driving an eastward propagating wave-packet-like response along great circles. With their own idealized global barotropic model, with a suitable forcing to represent the observed convective anomalies in Autumn 2000, the main result is the generation of a Rossby wave-train propagating into the extra-tropics remarkably similar to those in Figure 4.

    Back to top ^

    References

    • Alexander, L. and Jones, P. (2001). Updated precipitation series for the U.K. and discussion of recent extremes. Atmos. Sci. Lett., 1, 142\u2013150.
    • Blackburn, M. and Hoskins, B. (2001). Atmospheric variability and extreme Autumn rainfall in the UK. Available from http://www.met.rdg.ac.uk/ mike/autumn2000.html.
    • CEH (2001). Hydrological review of 2000. Technical report, Centre for Ecology and Hydrology, Wallingford/British Geological Survey.
    • DEFRA (2001). To what degree can the October/November flood events be attributed to climate change? Technical Report FD2304 (Final Report), Centre for Ecology and Hydrology, Wallingford/UK Met Office report to DEFRA.
    • EA (2001). Lessons Learned: Autumn 2000 Floods. Technical report, Environment Agency.
    • Marsh, T. and Dale, M. (2002). The UK floods of 2000-2001: A hydrometeorological appraisal. J. Chartered Inst. of Water and Env. Man., 16, 180\u2013188.

  • Research collaborations

    in

    Overview:

    In addition to investigating the effect of human-induced climate change on the risk of an event like the United Kingdom floods of Autumn 2000, we are also collaborating with researchers who will use data concurrently generated by this project for other regions of the world to investigate effects on seasonal-timescale events in those regions.

    Specifically, as well as returning a simulation of daily England and Wales total precipitation for the April 2000 - March 2001 period, the results uploaded back to us from participants also include concurrent simulations of other variables for that time over various other regions of the world (since the climate model being used has global coverage).

    There is daily precipitation and temperature returned for the Atlantic-European, Northwest USA, South African and Indian regions, and these are illustrated in the figure below:

    Also concurrently simulated are monthly precipitation, temperature, sea-level pressure, winds and land soil moisture over the whole globe.

    If you are at an academic research institution and would like to use this data from our project to investigate the effects of climate change in your region, then please get in touch with us via our Message Boards.

    Details of current collaborations:

    • Mountain snowpack decline in western North America
    • Heatwave occurrence in South Africa and India
    • Mountain snowpack decline in western North America
      Mountain snowpack in western North America is important for storing water from the winter (when most precipitation falls) and releasing it in spring and early summer, when economic, environmental, and recreational demands for water throughout the American West are frequently greatest. It has recently been found that widespread declines in spring-time snowpack have occurred in much of the region over the period 1925-2000, and that several factors argue for a climatic role in this (Mote et al 2005).

      By using precipitation and temperature data generated by the climate model simulations in this Seasonal Attribution project, and feeding it into a hydrological model, researchers can investigate the effects of human-induced climate change on mountain snowpack. The lead investigator in this collaboration is Dr Philp W. Mote in the Climate Impacts group at the University of Washington.

      The figure below shows the region in the climate model over which we are outputting daily precipitation and temperature data. The actual data plotted is for average conditions (ensemble mean from Beta test) around the time of year most important to the investigation.

      Example daily precip and temp output for NWUS


      Heatwave occurrence in South Africa and India

      Both these regions have suffered severe heatwaves recently. More details will follow.

      In the meantime, here's a little movie of the type of output we have for the Indian region. Notice how summer temperatures for an Industrial climate are generally a few degrees warmer than for the corresponding Non-Industrial climate:

      Example daily precip and temp output for NEPIND

    Syndicate content