Can scientists look at next year's climate?

[OKIsItJustMe]

http://newsroom.ucla.edu/portal/ucla/can-scientists-look-at-next-year-215064.aspx

Can scientists look at next year's climate?

By Stuart Wolpert | September 09, 2011

Is it possible to make valid climate predictions that go beyond weeks, months, even a year? UCLA atmospheric scientists report they have now made long-term climate forecasts that are among the best ever — predicting climate up to 16 months in advance, nearly twice the length of time previously achieved by climate scientists.

Forecasts of climate are much more general than short-term weather forecasts; they do not predict precise temperatures in specific cities, but they still may have major implications for agriculture, industry and the economy, said Michael Ghil, a distinguished professor of climate dynamics in the UCLA Department of Atmospheric and Oceanic Sciences and senior author of the research.

The study is currently available online in the journal http://www.pnas.org/content/early/2011/06/30/1015753108.abstract">Proceedings of the National Academy of Sciences (PNAS) and will be published in an upcoming print edition of the journal.

…

As is customary in this field, Ghil and his colleagues used five decades of climate data and test predictions retrospectively. For example, they used climate data from 1950 to 1970 to make "forecasts" for January 1971, February 1971 and beyond and see how accurate the predictions were. They reported achieving higher accuracy in their predictions 16 months out than other scientists achieved in half that time.

…

[ChandlerJr]

1. Weather is not climate

The whole premise of the article is invalid. The scientists are making general weather predictions for 16 months out.

The Old Farmers Almanac does it already.

[OKIsItJustMe]

2. I believe the researchers (and their peer reviewers) know the difference

http://dx.doi.org/10.1073/pnas.1015753108

Predicting stochastic systems by noise sampling, and application to the El Niño-Southern Oscillation

…

Interannual and interdecadal prediction are major challenges of climate dynamics. In this article we develop a prediction method for climate processes that exhibit low-frequency variability (LFV). The method constructs a nonlinear stochastic model from past observations and estimates a path of the “weather” noise that drives this model over previous finite-time windows. The method has two steps: (i) select noise samples—or “snippets”—from the past noise, which have forced the system during short-time intervals that resemble the LFV phase just preceding the currently observed state; and (ii) use these snippets to drive the system from the current state into the future. The method is placed in the framework of pathwise linear-response theory and is then applied to an El Niño–Southern Oscillation (ENSO) model derived by the empirical model reduction (EMR) methodology; this nonlinear model has 40 coupled, slow, and fast variables. The domain of validity of this forecasting procedure depends on the nature of the system’s path-wise response; it is shown numerically that the ENSO model’s response is linear on interannual time scales. As a result, the method’s skill at a 6- to 16-month lead is highly competitive when compared with currently used dynamic and statistic prediction methods for the Niño-3 index and the global sea surface temperature field.

ENSO forecasting has a decade-long history and relies mainly on two classes of models: dynamical and statistical (1, 2). Still, a further distinction has to be made within the latter class: Some of the statistical models do not make any use of dynamical information, like Lorenz’s method of analogues (3) and its followers (4–6), while others do use a dynamical model—previously fitted to the observations from the past—to drive the statistics in the future (2, 7, 8). Empirical stochastic models belong to this hybrid category, and linear versions of such models have been used in ENSO forecasting for two decades; see ref. 9 for a survey. More recently, Kravtsov et al. (10) have extended this approach to nonlinear models by developing an EMR methodology that can include quadratic nonlinearities as well as state-dependent noise that parameterizes small-scale effects, without assuming a priori scale separation (11).

The purpose of this paper is to show that, under suitable circumstances, a better understanding of the role of the fast processes, weather or noise, can help predict the slow ones—namely, the climate. To achieve this purpose, we proceed in two steps: (i) develop a special prediction methodology, called past noise forecasting (PNF), using EMR models; and (ii) provide a theoretical framework for applying the PNF method—or any other forecasting method based on perturbations of the noise—to other empirical stochastic models. Of late, probabilistic forecasts in weather and climate prediction have become fairly widespread: They are grounded in an estimation of the probability density function (PDF: 12–14).

We take here a distinct, pathwise approach instead, and will show that this approach is particularly well adapted to empirical stochastic models and to phenomena in which a considerable part of the variability exhibits some form of repetitive regularity (15). We focus mainly on the EMR models of ENSO introduced and studied by Kondrashov et al. (16).

…

http://nsidc.org/arcticmet/basics/weather_vs_climate.html

What is the Difference Between Weather and Climate?

Weather is the day-to-day state of the atmosphere, and its short-term (minutes to weeks) variation. Popularly, weather is thought of as the combination of temperature, humidity, precipitation, cloudiness, visibility, and wind. We talk about the weather in terms of "What will it be like today?", "How hot is it right now?", and "When will that storm hit our section of the country?"

Climate is defined as statistical weather information that describes the variation of weather at a given place for a specified interval. In popular usage, it represents the synthesis of weather; more formally it is the weather of a locality averaged over some period (usually 30 years) plus statistics of weather extremes.

We talk about climate change in terms of years, decades or even centuries. Scientists study climate to look for trends or cycles of variability (such as the changes in wind patterns, ocean surface temperatures and precipitation over the equatorial Pacific that result in El Niño and La Niña), and also to place cycles or other phenomena into the bigger picture of possible longer term or more permanent climate changes.

[AlecBGreen]

4. please re-read

Here it is again, in bold.

Is it possible to make valid climate predictions that go beyond weeks, months, even a year? UCLA atmospheric scientists report they have now made long-term climate forecasts that are among the best ever — predicting climate up to 16 months in advance, nearly twice the length of time previously achieved by climate scientists.

Forecasts of climate are much more general than short-term weather forecasts

[Nederland]

3. Amazing concession if true

While we've been looking at predictions of what climate will be like 100 years from now, this group is admitting that the best we can do is 16 months. Refreshing.

[OKIsItJustMe]

5. That’s not exactly what they’re saying…

… Interannual and interdecadal prediction are major challenges of climate dynamics. …

Making longer term projections (i.e. like for 100 years) is easier than making shorter term projections (like for 1-10 years.) This has been known to be a problem, because policy makers don’t like thinking about 100 years in the future (they tend to think about next year, or 5 years or so down the line.)

Modeling for the long run is easier than modeling for the short run.

Let me give you an analogy if I may. I am willing to predict that the temperature in 4 months will be colder than it is today. However, I am unwilling to predict what the temperature will be at the end of this month (relative to today.)

In 4 months time, the normal seasonal variation in the Northern Hemisphere virtually guarantees that it will be colder. However, since the seasonal variation between now and the end of the month isn’t as great, all sorts of other variables come into play.


Here’s an article you may find of interest:

http://onlinelibrary.wiley.com/doi/10.1002/wcc.69/pdf

Prospects for decadal climate prediction

Noel S. Keenlyside∗ and Jin Ba

During the last decade, global surface temperatures did not increase as rapidly as in the preceding decades. Although relatively small compared to the observed centennial scale global warming, it has renewed interest in understanding and even predicting climate on time scales of decades, and sparked a community initiative on near-term prediction that will feature in the fifth intergovernmental panel on climate change assessment report. Decadal prediction, however, is in its infancy, with only a few publications existing. This article has three aims. The first is to make the case for decadal prediction. Decadal fluctuations in global climate similar to that of recent decades were observed during the past century. Associated with large regional changes in precipitation and climate extremes, they are of socioeconomic importance. Climate models, which capture some aspects of observed decadal variability, indicate that such variations might be partly predictable. The second aim is to describe the major challenges to skilful decadal climate prediction. One is poor understanding of mechanisms of decadal climate variability, with climate models showing little agreement. Sparse observations in the past, particularly in the ocean, are also a limiting factor to developing and testing of initialization and prediction systems. The third aim is to stress that despite promising initial results, decadal prediction is in a highly experimental stage, and care is needed in interpreting results and utilizing data from such experiments. In the long-term, decadal prediction has the potential to improve models, reduce uncertainties in climate change projections, and be of socioeconomic benefit. © 2010 John Wiley & Sons, Ltd. WIREs Clim Change 2010 1 627–635

INTRODUCTION

Although defined as the prevailing weather conditions in a region, climate varies over a wide range of space and time scales. At the global scale, surface temperature increased by almost 1◦C during the last century (Figure 1). The increase is attributed largely to anthropogenic greenhouse gas emissions¹ (Ch. 9). Although these rose monotonically, global surface temperature did not, exhibiting clear interdecadal fluctuations: pronounced warming occurred during 1910–1940 and 1970–2000, and weak cooling from 1940 to 1970. Decadal-to-interdecadal fluctuations are also prominent in many regions. North Atlantic sea surface temperature (SST), for example, exhibited warming and cooling periods coherent with global surface temperature² (Figure 1). Eastern Tropical Pacific SST also exhibited decadal fluctuations³ (Figure 1). These show some correspondence to global changes, but have a shorter time scale and are less prominent compared to interannual variability.

Of direct relevance to society, decadal-to-interdecadal fluctuations are also found in atmospheric circulation patterns, precipitation, and climate extremes. For example, the North Atlantic Oscillation (NAO), a vacillation in sea level pressure between Iceland and the subtropical North Atlantic, underwent pronounced interdecadal variations⁶ (Figure 1). These were associated with strong changes in wintertime storminess, and European and North American surface temperature and precipitation, and thus had major economic impacts.⁶ Large interdecadal fluctuations were also seen in summertime Sahel rainfall⁷ (Figure 1), with profound consequence for people living in the region. For example, the drought of the 1970s to 1980s caused the death of at least 100,000 people, and displaced many more⁸ (Ch. 2). North America too suffered from persistent droughts, the 1930s ‘dust bowl’ is an example.⁹ North Atlantic Hurricane activity¹⁰ (Figure 1) and European temperature extremes¹¹ also exhibit multidecadal variations.

CAUSES OF OBSERVED INTERDECADAL CLIMATE VARIATIONS

Climate variations result from process external to the Earths climate system, internal to it, or a combination of both. The first category of variability, often referred to as externally forced, covers variations caused by factors considered external to the climate system, such as variations in solar forcing, and anthropogenic changes in greenhouse gas concentrations and aerosol loadings. The seasonal cycle and anthropogenic global warming are familiar examples. The second category, referred to as internal climate variability, encompasses variations that arise naturally from interactions within the atmosphere itself and with other components of the climate system, such as the ocean. Weather is a good example, arising from the atmosphere’s inherent nonlinearity, it can be essentially treated as stochastic on time scales longer than 14 days, and hence can produce variability on all time scales. Another example is the El Niño phenomenon, arising from ocean–atmosphere interaction in the Tropical Pacific. It has global impacts and explains much of the interannual variations in global surface temperature. It contributed to 1998 being one of the warmest years on record (Figure 1).

…

It is this shorter-term internal climate variability that the researchers in the OP are trying to address.