Long-term observations hold the key to climate change impact assessment

[OKIsItJustMe]

https://noc.ac.uk/news/long-term-observations-hold-key-climate-change-impact-assessment

[font face=Serif][font size=5]Long-term observations hold the key to climate change impact assessment[/font]

January 08, 2016

[font size=4]Most ocean data sets are far too short for the accurate detection of trends resulting from global climate change, according to research published today in the journal Global Change Biology. This study, by scientists at the National Oceanography Centre (NOC), will help to make decisions about where, and for how long, we should monitor the ocean in order to spot climate trends in ocean biology.[/font]

[font size=3]Around fifteen years of continuous data is sufficient to detect changes in the ocean that are a direct response to increases in atmospheric CO2, such as sea surface temperature and ocean acidity. However, changes that are less directly related to increasing CO2 levels are harder to pick out from the ‘noise’ of natural variability. These include populations of tiny marine plants, known as phytoplankton, which are the base of the marine food web and help the ocean absorb CO2 from the atmosphere.

The lead author, Dr Stephanie Henson from the NOC, said that “Picking out some trends from natural variability is like listening for a soft sound in a noisy room. This research really highlights the importance of continuing long-term observations at a range of sites across the ocean, so we can better detect the way it is changing. Without long-term datasets, it’s difficult to understand how ocean biology may respond to global climate change. ”

Stephanie estimates the natural variability in phytoplankton populations is so great that spotting any trends relating to climate change will require thirty to forty years of continuous data. This is a problem, as currently the longest available continuous global data set is twenty years.

…[/font][/font]

http://dx.doi.org/10.1111/gcb.13152

[Binkie The Clown]

1. So once humans are extinct we will be able to say for sure: Yup, climate change causes extinctions.

I am a great believer in, and admirer of the scientific process, but when your house is on fire that's not the time proceed cautiously and wait for further data. It's time to call the damn fire department!

[Duppers]

2. Zactly!

[OKIsItJustMe]

3. I don’t believe the intent was to say we can’t do anything until the science is all in

http://onlinelibrary.wiley.com/doi/10.1111/gcb.13152/full

[font face=Serif][font size=5]Observing climate change trends in ocean biogeochemistry: when and where[/font]

Stephanie A. Henson, Claudie Beaulieu, Richard Lampitt

First published: 6 January 2016
DOI: 10.1111/gcb.13152

…

[font size=4]Implications for ocean observatories[/font]

[font size=3]Our results allow an initial assessment of the adequacy of the current BGC-SO network for climate change trend detection, in terms of space and time scale considerations. In many cases, long running SOs are close to having sufficiently long time series to distinguish climate change-driven trends from background natural variability, for example ALOHA (Table S2). However, these well-established SOs provide only limited spatial coverage (Fig. 3). Several of the BGC-SOs considered here are relatively new or still in the planning stages and so, although some of them fill in gaps in the spatial coverage of the network, they may require decades more data before a climate change trend can be detected.

If the opportunity arose to design a new ocean observatory, with the primary goal of detecting climate change trends, the optimal location would be in a region of large spatial length scales and rapidly detectable trends (ignoring logistical issues). As an example, the size of the footprint for chlorophyll concentration calculated at every grid point (in the same way as for individual BGC-SOs, see 'Materials and methods') is plotted in Fig. 4. Largest footprints for chlorophyll are located in the equatorial Pacific and Indian Oceans, and parts of the Southern Ocean. In the case of the equatorial regions, these also overlap with areas of relatively short n* (<35 years), marked with a black contour in Fig. 4. None of the existing BGC-SOs are located within these optimal trend detection regions.

Although existing BGC-SOs may not necessarily be in an ideal location if the primary aim is detecting climate change, for the well-established stations little is to be gained from relocating them. Generally, n* is >20 years (except for SST and pH), so if >20 years of data have already been collected, then in most cases more than half the required time series to detect climate change is already in hand (and in some cases much more than half). Importantly, some BGC-SOs do not have climate change detection as a primary goal, focusing instead on process understanding. In these cases, the discussion presented here of time and space scales is of less relevance.

Our results suggest that the current network of BGC-SOs is, in some cases, adequate to assess climate change trends at the local scale. Some BGC-SOs may already have, or will soon have, sufficiently long time series to detect climate change-driven trends. Care, however, needs to be taken when calculating trends. Autocorrelation, which is prevalent in geophysical time series, can lead to the detection of spurious trends if not accounted for, particularly when using short time series (Wunsch, 1999). In addition, any ‘interventions’ in the time series, such as due to changes in sampling methodology or instrumentation, gaps in the dataset, relocation of sampling site etc., will tend to increase the number of years of data needed to detect a trend as the intervention effect must then be estimated and accounted for (Beaulieu et al., 2013). Finally, careful choice of the appropriate statistical model to fit to the data must be made as trends may not be linear.

…[/font][/font]