Summary

Many important decisions regarding water management, agriculture, and energy are made on weekly, monthly, seasonal, and annual timescales. These decisions can benefit from high quality, reliable predictions. Yet making useful predictions about the climate system on these timescales is a challenge. The purpose of this report is to examine current capabilities for making such intraseasonal to interannual (ISI)1 predictions for the climate system, to analyze how past improvements in these capabilities have been achieved, and to recommend opportunities for future improvement.

ISI climate predictions occupy an intermediate timescale between traditional weather forecasts, which are useful for the coming days, and global climate simulations associated with climate change, which relate to changes occurring over decades and centuries (see Box S.1). Predicting the climate at this intermediate timescale poses unique challenges since it involves many processes that operate among the atmosphere, ocean, and land surface. These processes are often incompletely understood and difficult to measure with available observational platforms. There are limits to the spatial and temporal resolution of our observations, and a “perfect” or complete observation of the climate system will never be achieved. Numerical models of the climate system demonstrate large sensitivity to initial conditions that cause errors or uncertainties to grow with increases in forecasting lead times. Moreover, models are known to have errors in formulation and are limited in resolution, which can also lead to forecast errors.

BOX S.1

MODELS FOR PREDICTING THE WEATHER VERSUS MODELS FOR PREDICTING CLIMATE

To understand climate prediction, it is useful to distinguish it from weather forecasting, which is a familiar concept to many from everyday experience. Weather models derive their prediction skill from accurate knowledge of initial conditions in the atmosphere. They produce deterministic forecasts, often with high enough skill that they can be used for simple everyday decisions, such as choosing proper clothing, or to warn us of short-term weather threats, such as lightning, severe winds, or intense precipitation. Climate models, on the other hand, derive much of their prediction skill from knowledge of the initial conditions in slowly evolving components of the climate system, such as the ocean or the cryosphere. The predictions produced by climate models are inherently probabilistic and have considerably lower skill than 1-2 day weather forecasts. They are usually of little use in planning everyday activities. However, climate predictions are very useful to government agencies, non-governmental organizations, and private companies for policy and longer-term planning purposes. Examples of applications include drought mitigation, malaria prevention, farming, pricing of insurance, and managing energy resources.

1

Intraseasonal to interannual is defined as extending from roughly two weeks to several years.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
Summary Many important decisions regarding water management, agriculture, and energy are made on weekly, monthly, seasonal, and annual timescales. These decisions can benefit from high quality, reliable predictions. Yet making useful predictions about the climate system on these timescales is a challenge. The purpose of this report is to examine current capabilities for making such intraseasonal to interannual (ISI)1 predictions for the climate system, to analyze how past improvements in these capabilities have been achieved, and to recommend opportunities for future improvement. ISI climate predictions occupy an intermediate timescale between traditional weather forecasts, which are useful for the coming days, and global climate simulations associated with climate change, which relate to changes occurring over decades and centuries (see Box S.1). Predicting the climate at this intermediate timescale poses unique challenges since it involves many processes that operate among the atmosphere, ocean, and land surface. These processes are often incompletely understood and difficult to measure with available observational platforms. There are limits to the spatial and temporal resolution of our observations, and a “perfect” or complete observation of the climate system will never be achieved. Numerical models of the climate system demonstrate large sensitivity to initial conditions that cause errors or uncertainties to grow with increases in forecasting lead times. Moreover, models are known to have errors in formulation and are limited in resolution, which can also lead to forecast errors. BOX S.1 MODELS FOR PREDICTING THE WEATHER VERSUS MODELS FOR PREDICTING CLIMATE To understand climate prediction, it is useful to distinguish it from weather forecasting, which is a familiar concept to many from everyday experience. Weather models derive their prediction skill from accurate knowledge of initial conditions in the atmosphere. They produce deterministic forecasts, often with high enough skill that they can be used for simple everyday decisions, such as choosing proper clothing, or to warn us of short-term weather threats, such as lightning, severe winds, or intense precipitation. Climate models, on the other hand, derive much of their prediction skill from knowledge of the initial conditions in slowly evolving components of the climate system, such as the ocean or the cryosphere. The predictions produced by climate models are inherently probabilistic and have considerably lower skill than 1-2 day weather forecasts. They are usually of little use in planning everyday activities. However, climate predictions are very useful to government agencies, non-governmental organizations, and private companies for policy and longer-term planning purposes. Examples of applications include drought mitigation, malaria prevention, farming, pricing of insurance, and managing energy resources. 1 Intraseasonal to interannual is defined as extending from roughly two weeks to several years. 1

OCR for page 1
2 Assessment of Intraseasonal to Interannual Climate Prediction and Predictability The committee was requested to assess how researchers and forecasters have addressed these challenges and to recommend avenues for further progress. Specifically, the committee was tasked to review the current understanding of ISI predictability, describe how past improvements in forecast systems have occurred, identify gaps in our current understanding of ISI predictability, assess the performance of current ISI forecast systems, and recommend strategies and best practices for future improvements to ISI forecasts and our overall understanding of ISI predictability. The committee begins from the premise that the ability to predict the climate accurately at ISI timescales stems from our knowledge of “sources of predictability,” the variables or processes operating within and among the atmosphere, ocean, and land that affect the state of the climate on ISI timescales. The sources of predictability are measured, represented, and simulated by ISI forecast systems through an assemblage of “building blocks,” namely observational systems, statistical and dynamical models, and data assimilation schemes. This report illustrates the relationship between the sources of predictability and the building blocks of ISI forecast systems. In addition, this report discusses techniques and protocols for the verification and dissemination of ISI forecasts by operational forecasting centers, highlighting the impact that these practices can have on forecast quality and opportunities for improvement. This report concludes with recommendations for improving ISI forecast systems, targeting both operational forecasting centers and the broader research community. SOURCES OF PREDICTABILITY This report explores three interrelated categories of predictability sources that exist within the climate system. The first of these sources of predictability is related to particular variables that exhibit inertia or memory, such as ocean heat content, in which anomalous conditions can take relatively long periods of time (days to years) to decay. The second type of source of predictability is related to patterns of variability or feedbacks. Coupling among processes in the climate system can give rise to characteristic patterns that explain some portion of the spatial and temporal variance exhibited by key climate variables, such as temperature or precipitation. An example is the El Niño-Southern Oscillation (ENSO), where anomalous conditions in the tropical Pacific Ocean influence seasonal climate in the mid-latitudes around the globe. The third source of predictability is due to external forcing. Volcanic eruptions, changes in solar activity, and the accumulation of greenhouse gases in the atmosphere are all examples of external forcing. These events or processes can affect the climate on ISI timescales in predictable ways that can be exploited for making climate predictions. It is important to note that the processes that affect the climate on ISI timescales can themselves operate on a variety of timescales. This is depicted in Figure S.1, which provides many examples of processes that affect the climate at ISI timescales and can serve as sources of predictability. These sources can be related to phenomena that occur in, on, or among the ocean, atmosphere, and land surface components of the climate system. The ability of ISI forecast systems to represent these sources of predictability accurately partially determines the quality of the predictions. Past improvements in prediction quality have accompanied increased understanding of the sources of predictability and incorporation of this understanding into forecast systems. Future advances in the quality of ISI predictions are closely

OCR for page 1
Summary 3 FIGURE S.1 Processes that act as sources of ISI climate predictability extend over a wide range of timescales and involve interactions among the atmosphere, ocean, and land. CCEW: convectively coupled equatorial waves; TIW: tropical instability wave; MJO/MISV: Madden- Julian Oscillation/Monsoon intraseasonal variability; NAM: Northern Hemisphere annular mode; SAM: Southern Hemisphere annular mode; AO: Arctic oscillation; NAO: North Atlantic oscillation; QBO: quasi-biennial oscillation, IOD/ZM: Indian Ocean dipole/zonal mode; AMOC: Atlantic meridional overturning circulation. For the y-axis, “A” indicates “atmosphere;” “L” indicates “land;” “I” indicates “ice;” and, “O” indicates “ocean.” tied to exploiting new sources of predictability or improving the representation of known sources of predictability in current forecast systems. THE BUILDING BLOCKS OF AN ISI FORECAST SYSTEM ISI forecasting systems are composed of several “building blocks:” observations, statistical and dynamical models, and data assimilations systems. Observations are required to measure the state of the variables that contain memory in the climate system; to monitor the evolution of key processes that operate within and among the atmosphere, ocean, and land; and to identify the magnitude of external forcing. These observations can be utilized by a data assimilation system to provide the current state (an “initial state”) of the climate system, and that information can be utilized by statistical and/or dynamical models to make predictions. Observations are required to validate models, verify forecasts, and expand understanding of

OCR for page 1
4 Assessment of Intraseasonal to Interannual Climate Prediction and Predictability underlying climate processes. These activities can then feedback into identifying model deficiencies and improving model formulations. The performance of ISI forecast systems can be enhanced through improvements to these building blocks, which are intimately connected. For example, if new observations are made available, then it is likely that new components of statistical and dynamical models or data assimilation algorithms need to be developed in order to incorporate these observations into forecasts. Conversely, a comparison of existing models to novel observations or models may identify underlying deficiencies in our understanding of important climate processes, motivating further model development. Thus, improvements to ISI forecast systems stem from synergistic improvements across each of the building blocks, where upgrades to one component enhance, or are enhanced by, upgrades to the other components. Based on its examination of the literature, the committee concludes that incremental increases in ISI forecasting quality are to be expected as the building blocks of ISI forecast systems are improved and upgraded. The committee also concludes that there are no “silver bullets;” there is no single action that will lead to a revolutionary leap forward in ISI predictions. As past improvements to ISI predictions and weather predictions have shown, progress forward can be achieved by a concerted effort to address the shortcomings of the various building blocks of forecast systems. CASE STUDIES Much can be learned about ISI predictions by exploring case studies for ENSO, the Madden-Julian Oscillation (MJO), and soil moisture. Such case studies demonstrate the role that observations, models, data assimilation techniques, and verification protocols play in making ISI forecasts. For ENSO, the perspective is somewhat historical, as many previous advances in ISI forecasting have come from an improved observational capacity that accompanied expanded understanding of physical processes and model development. For the MJO and soil moisture, the perspective is more forward-looking. There remains the potential to exploit the MJO and soil moisture to improve ISI forecasts. ISI FORECASTING INSTITUTIONS ISI forecasts, along with the building blocks for forecasting, are developed, produced, and processed by a variety of institutions around the world, including the National Centers for Environmental Prediction (NCEP), the European Centre for Medium Range Forecasting (ECMWF), the International Research Institute for Climate and Society (IRI), and many universities and research laboratories. The committee draws a distinction between “operational centers” and “research institutions.” The former issue forecasts in real time on a fixed, regular schedule and are associated with a national meteorological and hydrological service; the latter are more research-oriented and are often associated with universities and academic scientists. Programs that can foster collaboration between these two types of institutions have been successful in advancing ISI forecast quality, and several of the recommendations aim to encourage further collaboration and enhance existing mechanisms for cooperation.

OCR for page 1
Summary 5 USE OF FORECASTS Quality forecasts can contribute to societally-relevant decisions. However, a variety of metrics can be used to determine overall forecast quality; a single metric rarely encapsulates all the information regarding forecast quality. Likewise, different decision makers will rely on a diverse set of variables (e.g. agricultural planners may be most interested in precipitation, while wind power operators may be most interested in wind forecasts), and may have varying demands for the forecast quality associated with these variables. The report describes the procedure of making ISI forecasts, outlining how information from various sources, both objective (e.g., predictions from dynamical or statistical models) and subjective (e.g., expert opinions of forecasters), are combined. The wide range of forecast formats and accessible forecast documentation makes it challenging to compare the performance of forecast systems or detect how changes in forecast inputs and practices affect overall forecast performance. Similarly, the variables and formats of forecasts may not correspond to the needs of decision makers, acting as a barrier to the use of forecasts, regardless of their quality. RECOMMENDATIONS The committee identified three general categories of actions to advance ISI predictions: Best Practices, Improvements to the Building Blocks of ISI Forecast Systems, and Research for Sources of Predictability. Best Practices are largely focused on the activities of the operational forecast centers and aim to improve the delivery and dissemination of forecast information for both decision makers and researchers. Although adopting Best Practices may require some additional resources on the part of the operational centers, the barriers to adoption are relatively minimal; many of the recommendations involve modification to current protocols or expansion of current programs rather than a novel set of initiatives. The Improvements to the Building Blocks of ISI Forecast Systems pertain to both the operational and research communities and focus on the continued development of observations, statistical and dynamical models, and data assimilation systems. The benefits associated with these recommendations have a longer time horizon than those associated with Best Practices and may require several years to achieve. Research for Sources of Predictability, the final category of the recommendations, is aimed primarily toward the research community. These recommendations constitute specific goals for current and future academic exploration of ISI processes. Although the committee agrees that these goals should be pursued with the intent that they contribute to operational ISI forecasts, the initial efforts to investigate these unexploited sources of predictability fall largely on research scientists. Best Practices (1) The synergy between operational ISI forecasting centers and the research community should be enhanced. Establishing connections between the operational and research communities is critical to further progress in ISI forecasting. Fostering dialog and exchange between these communities permits identification of common problems and expands the sets of tools available for finding

OCR for page 1
6 Assessment of Intraseasonal to Interannual Climate Prediction and Predictability solutions. Specific activities include holding workshops focused on specific areas of model and forecast development, encouraging scientists that work at operational centers to participate in scientific meetings focused on modeling and the use of observations, granting of short term positions in operational centers to academic researchers, and improving the speed and manner by which new data sets generated by operational centers are shared with the broader research community. (2) Operational ISI forecasting centers should establish public archives of all data used in forecasts including observations, model code, hindcasts, analyses, forecasts, re-analyses, re- forecasts, verifications, and official forecast outlooks. Archives of the inputs to, outputs of, and tools used in ISI forecasts are needed in order to quantify and identify sources of forecast error, provide the baseline for forecast assessment and model fidelity, develop metrics and diagnostics for model assessment, calibrate model predictions, and document model and forecast improvements. Archives can serve as an important mechanism for making ISI forecasts more readily useable for management decisions and societally relevant research. Although it is not possible for operational centers to foresee or address all possible needs of the forecast users, archives will permit the development of tailored forecast products for decision systems and risk management by users and researchers. Once engaged, these groups can also provide valuable feedback for further improvements in ISI forecasting. (3) Operational ISI forecasting centers should broaden and make available the collection of metrics used to assess forecast quality. No perfect metric exists that conveys all the information about a forecast. Multiple metrics should be used when assessing forecasts, including graphical techniques; metrics that assess the quality of probabilistic information and that from multi-model ensembles. Some of these metrics should include information on the distribution of forecast skill in space and time. (4) The subjective components of operational ISI forecasts should be minimized. Recent research suggests that the subjective component of many present-day forecasts can reduce forecast quality. The subjective component generally comes from qualitative discussion and interpretation by forecasters regarding the state of the climate system and forecasting tools. The subjective component also limits reproducibility, restricting retrospective comparison of forecast systems. Improvements to the Building Blocks of Forecast Systems (5) Statistical techniques, especially nonlinear methods, should be pursued in order to better characterize processes that contribute to ISI forecasts. Statistical methods provide important tools for comparing model predictions and observations and subsequently identifying model deficiencies. Historically, linear statistical analyses of observational data have provided an awareness of many patterns of variability that have been useful for making ISI forecasts. Recent research demonstrates that nonlinear methods can yield statistically significant increases in prediction skill on ISI time scales when compared to traditional linear techniques. However, these techniques have not been incorporated

OCR for page 1
Summary 7 operationally. Therefore, nonlinear alternatives should be explored to augment our current knowledge. (6) Systematic errors in dynamical models should be identified. Current state-of-the-art ISI prediction models have relatively large errors in their representation of the mean climate, the climate variability, and their interaction. These errors reduce prediction quality. Some classic examples include: (1) the so-called double intertropical convergence zone (ITCZ) problem, (2) the excessively strong equatorial cold tongue, (3) weak or incoherent intraseasonal variability, (4) failure to represent the multi-scale organization of tropical convection, and (5) poorly represented cloud processes, particularly low level stratus. These errors have both regional and global impacts and could be indicative of errors in the model formulations that are limiting prediction quality. Sustained observations are needed to quantify model errors. Examples of sustained observations include those related to describing the properties or fluxes among the atmosphere, ocean, and land surface (e.g., boundary layer humidity, exchange of heat between the atmosphere and ocean). (7) To reduce errors produced by dynamical models, the representation of physical processes should be improved. The physical processes underlying ISI variability are often poorly understood. Process studies that are closely tied to operational ISI model improvement should be carried out with the goal of transferring improvements into operational ISI forecasts. Targeted, novel observations will likely play a role in these types of studies. Studies could focus on specific components of the climate system (e.g., sea ice, aerosols, snow cover), specific processes and variability (e.g., triggering the onset of an MJO), and the interactions among components of the climate system (e.g., air-land coupling strength, stratosphere-troposphere interactions). The CLIVAR climate process teams (CPTs), which exist currently, provide a mechanism for accomplishing this. The CPTs focus modelers and process scientists on poorly-represented or unrepresented physical processes in models. Work should be carried out to move toward more complete inclusion of climate processes in the models. Computing capabilities should be improved to permit the explicit simulation of subgrid-scale processes and remove as much reliance on parameterization as possible. The role of increasing model resolution in improving ISI forecasts should continue to be explored. (8) Statistical and dynamical models should continue to be used in a complementary fashion by operational ISI forecasting centers. Using multiple prediction tools leads to improved and more complete ISI forecasts. Statistical tools should continue to be developed and employed in an effort to improve dynamical model output. Examples of statistical techniques include stochastic physics, interactive ensembles, empirical corrections or empirically-based parameterizations and process models. The use of statistical and dynamical downscaling methods is another application that should be explored to address the information mismatch between the coarse spatial resolution of operational climate forecasts and the fine resolution needs of some end users. (9) Multi-model ensemble (MME) forecast strategies should be pursued, but standards and metrics for model selection should be developed.

OCR for page 1
8 Assessment of Intraseasonal to Interannual Climate Prediction and Predictability Continued work is necessary to develop techniques of optimally selecting and weighting ensemble members. Experimentation with MME should not compete with model improvement, but rather, should contribute to the process of identifying areas for model improvement. (10) To enable assimilation of all available observations of the coupled climate system, operational centers should implement state-of-the-art 4-D Var, Ensemble Kalman Filters, or hybrids of these in their data assimilation systems. The most advanced assimilation systems are typically not used in operational settings or are limited to atmospheric observations only. Assimilation systems should be upgraded. There are many available observations that are not currently being utilized in data assimilation schemes that could contribute to the initialization of dynamical models. More observations should be assimilated into operational ISI forecast systems. The expansion of the variables assimilated in weather forecasts has contributed to improvements in forecast quality. Analogous gains could be made for ISI forecasting. Priority should be given to expanding operational data assimilation to ocean observations such as sea surface heights. Research for Sources of Predictability (11) Many sources of predictability remain to be fully exploited by ISI forecast systems. To better understand key processes that are likely to contribute to improved ISI predictions, the committee recommends that the scientific community pursue the following six areas as research goals. Madden-Julian Oscillation (MJO) The path forward on understanding and forecasting the MJO should include focused process studies, model improvement, and close collaboration between research and operational communities. It will be necessary to develop and implement standardized diagnostics and metrics to gauge model improvements and track improvements in forecast quality. MJO influences on other important components of the climate system, such as ENSO, monsoon onsets and breaks, and tropical cyclone genesis should continue to be explored and exploited for additional predictability. Stratosphere-Troposphere Interactions Relatively long-lived (up to two months) atmospheric anomalies can arise from stratospheric disturbances. In sensitive areas such as Europe in winter, experiments suggest that the influence of stratospheric variability on land surface temperatures can exceed the local effect of sea surface temperature. Additionally, while our weather and climate models do not often resolve or represent the stratospheric Quasi-Biennial Oscillation very well, it is one of the more predictable features in the atmosphere, and it has been found to exhibit a signature in ISI surface climate. Ocean-atmosphere coupling Due to the very large heat capacity of sea water, anomalous sea surface temperatures and upper ocean heat content can have significant impacts on the atmosphere above. The impacts of the anomalies associated with ENSO are well-known. However, further research is needed to examine the role of extratropical atmosphere-ocean coupling, to investigate the need to more

OCR for page 1
Summary 9 realistically represent ocean-atmosphere coupling over a wide range of spatial scales (including down to the scales of the sharp SST gradients associated with fronts), and to better observe and more realistically represent air-sea fluxes in models. Land-atmosphere feedbacks The realistic initialization of soil moisture in models can increase the accuracy of precipitation and temperature predictions at intraseasonal timescales. The realistic initialization of snow amount may also yield better quality predictions, though this connection is relatively unexplored. To maximize the impact of land feedbacks on prediction quality, the mechanisms underlying the land-atmosphere coupling (e.g., evaporation, boundary layer dynamics, convection) need to be better understood and better represented in forecast systems. High impact events affecting atmospheric composition Research efforts should study the consequences on the climate system at ISI timescales of unusual but high impact events, such as volcanic eruptions, limited nuclear exchanges, or space impacts that cause a sudden, drastic change to the atmospheric burden of aerosols and trace gases. ISI forecasts from operational centers following these types of events could have significant societal ramifications. Non-stationarity Statistical and dynamical models for ISI forecasting should be improved to better capture the predictability associated with long-term trends in atmospheric composition (e.g., increases in greenhouse gas concentrations) and land cover change. Current statistical techniques and dynamical models do not adequately deal with this non-stationarity. Improved statistical techniques should be developed for exploiting the predictability associated with such non- stationary behavior. The use of dynamical models that include a more comprehensive treatment of radiative processes such as aerosol effects, and also incorporate trends in land use, could help improve the quality of dynamical ISI forecasts on longer timescales. CLOSING THOUGHTS For the short term, operational ISI forecast centers can increase the value of forecasts to decision makers and researchers by modifying procedures for archiving and disseminating forecast information and enhancing collaborations with the external research community. Over the next several years and coming decades, improvements to observational capabilities, statistical and dynamical models, and data assimilations systems should permit ISI forecast systems to better represent the variables and processes that serve as sources of predictability. Research to characterize sources of predictability that are poorly understood should also offer opportunities to improve ISI predictions as well as our understanding of important underlying climate processes.

OCR for page 1