How to understand and reason about uncertainty in climate science is a topic that is receiving increasing attention in both the scientific and philosophical literature. This paper provides a perspective on exploring ways to understand, assess, and reason about uncertainty in climate science, including application to the Intergovernmental Panel on Climate Change (IPCC) assessment reports. Uncertainty associated with climate science and the science–policy interface presents unique challenges owing to the complexity of the climate system itself, the potential for adverse socioeconomic impacts of climate change, and the politicization of proposed policies to reduce societal vulnerability to climate change. The challenges to handling uncertainty at the science– policy interface are framed using the “monster” metaphor, whereby attempts to tame the monster are described. An uncertainty lexicon is provided that describes the natures and levels of uncertainty and ways of representing and reasoning about uncertainty. Uncertainty of climate models is interpreted in the context of model inadequacy, uncertainty in model parameter values, and initial condition uncertainty. This article examines the challenges of building confidence in climate models and, in particular, the issue of confidence in simulations of the twenty-first-century climate. The treatment of uncertainty in the IPCC assessment reports is examined, including the IPCC Fourth Assessment Report conclusion regarding the attribution of climate change in the latter half of the twentieth century. Ideas for monster-taming strategies are discussed for institutions, individual scientists, and communities.

Abstract

Using the NCEP–NCAR reanalysis for the 1950–2000 period, differences in the atmospheric hydrological cycle between the extremes of ENSO (i.e., La Niña minus El Niño) are examined. Zonal vertically integrated moisture transport (VIMT) across 100°E accounts for about half of the variability in net moisture convergence in the north Indian Ocean region between ENSO extremes when all ENSO events are considered. Changes in VIMT across 100°E are associated with large changes in the strength of the Pacific Ocean trade wind regime during ENSO. The bulk of the remaining VIMT anomalies are from the Arabian Sea and appear to be associated with sea level pressure variations in the northern and western parts of the Indian Ocean Basin. This initial analysis, therefore, suggests that the interaction between the monsoon and ENSO may be more complex than the direct modulation of VIMT by the Pacific Ocean trade winds alone.

The analysis is refined further by comparing the differences of the Indian and Pacific Ocean hydrological cycles between ENSO extremes when they occur concurrently with anomalous monsoons [ENSO–anomalous monsoon years (EAM)], and when the monsoon is normal [ENSO–normal monsoon years (ENM)]. For both EAM and ENM years, similar differences exist in VIMT across 100°E between ENSO extremes. However, major differences are noted in VIMT anomalies from the west and south into the north Indian Ocean region. Thus, the principal difference in moisture convergence in the north Indian Ocean between EAM and ENM years is associated primarily with VIMT anomalies in the western Indian Ocean region and not those in the eastern Indian or Pacific Oceans.

To test the hypothesis that Pacific Ocean SST anomalies occurring prior to the monsoon may be important in influencing the eventual nature of the monsoon, the analysis is extended backward to the spring period. While May SST differences in the Niño-3 region between ENSO extremes are found to be similar for both EAM and ENM years, VIMT differences in both the Indian Ocean and the central and western Pacific Oceans are significantly larger during EAM years than ENM years. May SST differences in the central subtropical Pacific Ocean are also significantly larger during EAM than ENM years. These results show that the anomalous SST gradient between the eastern equatorial and the central subtropical Pacific Ocean prior to the monsoon onset, together with its associated VIMTs anomalies, may be important factors in determining the degree of connection between monsoon and ENSO. In addition, the circulation in the Indian Ocean prior to and during the monsoon onset shares a strong association with the eventual intensity of the monsoon–ENSO coupling.

Abstract

A diagnostic criterion that retrospectively assesses the onset and withdrawal dates of the Indian monsoon is derived from variability in the large-scale hydrologic cycle. The method is proposed as an improved means with which to understand interannual variability in the monsoon transitions as compared to criteria that rely heavily on rainfall variability over limited spatial domains (e.g., individual Indian districts). The hydrologic cycle is chosen as a key physical basis for monitoring the monsoon due to the essential roles played by zonal and meridional gradients in water vapor, clouds, and rainfall in driving the large-scale monsoon circulation. Moreover, as rainfall is greater than evaporation in wet monsoonal areas, lateral transports of water vapor are required for the existence of monsoonal rains. To diagnose onset and withdrawal, vertically integrated moisture transport (VIMT) is therefore used instead of rainfall, which over the large scale is often poorly measured and modeled. In contrast to rainfall, VIMT is generally well modeled and observed, and its variability, particularly over the Arabian Sea, is substantial during both monsoon onset and withdrawal. An index, named the hydrologic onset and withdrawal index (HOWI), is thus formed from those regions where VIMT variability is pronounced at the beginning and end of the monsoon season. The HOWI offers several advantages as the index is based on fields that are better modeled and measured than rainfall, and the index is indicative of the transition in the large-scale monsoon circulation rather than being highly sensitive to synoptic variability and the spatial complexity of the monsoon transitions. The HOWI is shown to be both robust to bogus monsoon onsets and reflective of the timing, rather than the spatial character, of the transitions.

Analysis of interannual variability in monsoon onset and withdrawal dates based on the HOWI reveals robust associations that are weak and insignificant when assessed using other onset criteria. For example, the associations between total June–July–August–September (JJAS) rainfall and both monsoon onset and withdrawal are weak (correlations are weaker than −0.11) when onset dates from the Indian Meteorological Department (IMD) or other objective methods are considered. However, the HOWI criterion shows strong correlations between total JJAS rainfall and both onset (0.30) and withdrawal (−0.49). Thus, the length of the monsoon season is shown to be strongly related to its overall strength. In addition, while the correlation between IMD onset date and Niño-3 SST is insignificant, the correlation based on HOWI is 0.41. The associations between HOWI and both ENSO and the overall monsoon season exceed significance at the 99% confidence level. Moreover, the associations are shown to be robust to the scale of the region selected in compiling the HOWI. It is speculated that the influence of synoptic variability and the spatially variable nature of the monsoon transitions mask, in part, the existence of the climate associations that are revealed by the HOWI.

The 2005 Atlantic hurricane season was the most active and costly season on record. Recent publications linking an increase in hurricane intensity to increasing tropical sea surface temperatures have fueled the debate on whether or not global warming is causing an increase in hurricane intensity. Because of the substantial implications of the hurricane–global warming issue for society and the immediate policy relevance associated with decision making related to Hurricane Katrina, attacks and rebuttals related to this research are being made in the media and on the World Wide Web without the rigor or accountability expected of scientific discourse. In this paper, we aim to promote a balanced and thoughtful examination of this subject by

• clarifying the debate surrounding the subject as to whether or not global warming is causing an increase in global hurricane intensity,

• illustrating a methodology of hypothesis testing to address multiple criticisms of a complex hypothesis that involves a causal chain, and

• providing a case study of the impact of politics, the media, and the World Wide Web on the scientific process.

Abstract

Intraseasonal oscillations (ISOs) are important large-amplitude and large-scale elements of the tropical Indo-Pacific climate with time scales in the 20–60-day period range, during which time they modulate higher-frequency tropical weather. Despite their importance, the ISO is poorly simulated and predicted by numerical models. A joint diagnostic and modeling study of the ISO is conducted, concentrating on the period between the suppressed and active (referred to as the “transition”) period that is hypothesized to be the defining stage for the development of the intraseasonal mode and the component that is most poorly simulated.

The diagnostic study uses data from the Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE). It is found that during the transition period, the ocean and the atmosphere undergo gradual but large-scale and high-amplitude changes, especially the moistening of the lower troposphere caused jointly by the anomalously warm sea surface temperature arising from minimal cloud and low winds during the suppressed phase and the large-scale subsidence that inhibits the formation of locally deep convection. Using a cloud classification scheme based on microwave and infrared satellite data, it is observed that midtop (cloud with a top in the middle troposphere) nonprecipitating clouds are a direct response of the low-level moisture buildup.

To investigate the sensitivity of ISO simulations to the transitional phase, the European Centre for Medium-Range Weather Forecasts (ECMWF) coupled ocean–atmosphere climate model is used. The ECMWF was run serially in predictive ensemble mode (five members) for 30-day periods starting from 1 December 1992 to 30 January 1993, encompassing the ISO occurring in late December. Predictability of the active convective period of the ISO is poor when initialized before the transitional phases of the ISO. However, when initialized with the correct lower-tropospheric moisture field, predictability increases substantially, although the model convective parameterization appears to trigger convection too quickly without allowing an adequate buildup of convective available potential energy during the transition period.

Abstract

The seasonal dependence of atmospheric short-term climate (i.e., seasonal to interannual) predictability is studied. This is accomplished by analyzing the output from ensemble integrations of the European Centre for Medium-Range Weather Forecasts model. The integrations use the observed evolution of sea surface temperature (SST) as prescribed boundary forcing. Forced by the interannual variation of SST, the short-term climate predictability of the atmospheric circulation is geographically and seasonally dependent. In general, the predictability is larger in the Tropics than the extratropics and is greater in the Pacific–Atlantic Ocean sector compared to the Indian Ocean–Asian monsoon region. Predictability is also higher in the winter hemisphere than in the summer hemisphere. On average, the weakest predictability in the Northern Hemisphere occurs during the northern autumn. However, it is noted that the 1982/83 strong El Niño event produced stronger atmospheric predictability than the 1988/89 strong La Niña event during the northern spring, and the predictability pattern is reversed during the northern autumn.

Predictability is further partitioned into its internal and external components. The external component is defined as the interannual variation of ensemble average, and the internal component is the sample-to-sample variance. The temporal and spatial structure in the external variability accounts for most of the structure in the SST-forced atmospheric predictability. However, there are regions in the Tropics, such as over the monsoon region, where the external and internal variabilities show roughly the same magnitude. Overall, internal variability is largest in the extratropics. Specifically, the internal variability is larger in the northern extratropics during the northern autumn and larger in the southern extratropics during the northern spring. In contrast, the external variability is smaller (larger) in the northern extratropics during the northern autumn (spring).

It is concluded that major features of the SST-forced atmospheric predictability are determined by the external variability in the Tropics. In the extratropics, the predictability is determined by seasonal variations in both internal and external variabilities. The weakest predictability that occurs in the northern extratropics during the northern autumn is the result of a conjunction of a local increase in internal variability and a decrease in external variability at the same time.

Furthermore, the external variability is controlled by seasonality in the forcing over the tropical Pacific Ocean, which is largely determined by the following two mechanisms: 1) the annual cycle–ENSO interaction over the tropical Pacific Ocean and 2) nonlinear effects of hydrological processes associated with the annual cycle–ENSO interaction. Also, it is interesting that the annual cycle–ENSO interaction can be summarized into a conceptual model that shows some analogy to the quark model in nuclear physics.

Abstract

The interaction of the Indian Ocean dynamics and the tropospheric biennial oscillation (TBO) is analyzed in the 300-yr control run of the National Center for Atmospheric Research (NCAR) Climate System Model (CSM). Sea surface temperature (SST) anomalies and equatorial ocean dynamics in the Indian Ocean are associated with the TBO and interannual variability of Asian–Australian monsoons in observations. The air–sea interactions involved in these processes in the coupled ocean–atmosphere model are analyzed, so as to diagnose the causes of the SST anomalies and their role in the development of a biennial cycle in the Indian–Pacific Ocean region.

By using singular value decomposition (SVD) analysis, it is found that the model reproduces the dominant mechanisms that are involved in the development of the TBO's influence on the south Asian monsoon: large-scale forcing from the tropical Pacific and regional forcing associated with both the meridional temperature gradient between the Asian continent and the Indian Ocean, as well as Indian Ocean SST anomalies. Using cumulative anomaly pattern correlation, the strength of each of these processes in affecting the interannual variability of both Asian and Australian monsoon rainfall is assessed.

In analyzing the role of the Indian Ocean dynamics in the TBO, it is found that the Indian Ocean zonal mode (IOZM) is an inherent feature of the Asian summer monsoon and the TBO. The IOZM is thus a part of the biennial nature of the Indian–Pacific Ocean region. The coupled ocean–atmosphere dynamics and cross-equatorial heat transport contribute to the interannual variability and biennial nature of the ENSO–monsoon system, by affecting the heat content of the Indian Ocean and resulting SST anomalies over multiple seasons, which is a key factor in the TBO.

The Aerosonde is a small robotic aircraft designed for highly flexible and inexpensive operations. Missions are conducted in a completely robotic mode, with the aircraft under the command of a ground controller who monitors the mission. Here we provide an update on the Aerosonde development and operations and expand on the vision for the future, including instrument pay loads, observational strategies, and platform capabilities. The aircraft was conceived in 1992 and developed to operational status in 1995–98, after a period of early prototyping. Continuing field operations and development since 1998 have led to the Aerosonde Mark 3, with ~2000 flight hours completed. A defined development path through to 2002 will enable the aircraft to become increasingly more robust with increased flexibility in the range and type of operations that can be achieved. An Aerosonde global reconnaissance facility is being developed that consists of launch and recovery sites dispersed around the globe. The use of satellite communications and internet technology enables an operation in which all aircraft around the globe are under the command of a single center. During operation, users will receive data at their home institution in near-real time via the virtual field environment, allowing the user to update the mission through interaction with the global command center. Sophisticated applications of the Aerosonde will be enabled by the development of a variety of interchangeable instrument payloads and the operation of Smart Aerosonde Clusters that allow a cluster of Aerosondes to interact intelligently in response to the data being collected.

An integrated approach is presented for determining from several different satellite datasets all of the components of the tropical sea surface fluxes of heat, freshwater, and momentum. The methodology for obtaining the surface turbulent and radiative fluxes uses physical properties of the atmosphere and surface retrieved from satellite observations as inputs into models of the surface turbulent and radiative flux processes. The precipitation retrieval combines analysis of satellite microwave brightness temperatures with a statistical model employing satellite observations of visible/infrared radiances. A high-resolution dataset has been prepared for the Tropical Ocean Global Atmosphere Coupled Ocean–Atmosphere Response Experiment (TOGA COARE) intensive observation period (IOP), with a spatial resolution of 50 km and temporal resolution of 3 h. The high spatial resolution is needed to resolve the diurnal and mesoscale storm-related variations of the fluxes. The fidelity of the satellite-derived surface fluxes is examined by comparing them with in situ measurements obtained from ships and aircraft during the TOGA COARE IOP and from vertically integrated budgets of heat and freshwater for the atmosphere and ocean. The root-mean-square differences between the satellite-derived and in situ fluxes are dominated by limitations in the satellite sampling; these are reduced when some averaging is done, particularly for the precipitation (which is from a statistical algorithm) and the surface solar radiation (which uses spatially sampled satellite pixels). Nevertheless, the fluxes are determined with a useful accuracy, even at the highest temporal and spatial resolution. By compiling the fluxes at such high resolution, users of the dataset can decide whether and how to average for particular purposes. For example, over time, space, or similar weather events.

The methods and initial results of an extensive pilot study, the Joint Air–Sea Monsoon Interaction Experiment (JASMINE) held in the Indian Ocean during the summer of 1999, are described. The experimental design was based on the precept that the monsoon sways back and forth from active to inactive (or break) phases and that these intraseasonal oscillations are coupled ocean–atmosphere phenomena that are important components of the monsoon system. JASMINE is the first comprehensive study of the coupled ocean–atmosphere system in the eastern Indian Ocean and the southern Bay of Bengal. Two research vessels, the NOAA ship Ronald H. Brown and the Australian research vessel Franklin, totaled 52 days of surveillance in April–June and September, with 388 conductivity–temperature–depth (CTD) casts and 272 radiosonde ascents. In addition, both ships carried identical flux systems to measure the ocean–atmosphere interaction. The Brown had five radar systems and profilers, including a cloud radar and a Doppler C-band rain radar.

Active and break periods of the monsoon, and the transitions between these phases, and the onset of the 1999 South Asian summer monsoon occurred during JASMINE. The undisturbed and disturbed periods had vast differences in the net heating of the ocean, ranging from daily averages of +150 W m⁻² during the former to −100 W m⁻² in the latter. Accompanying these changes in the monsoon phase were distinct states of the upper ocean and the atmosphere, including complete reversals of the near-equatorial currents on the timescales of weeks. Diurnal variability occurred in both phases of the monsoon, particularly in near-surface thermodynamical quantities in undisturbed periods and in convection when conditions were disturbed. The JASMINE observations and analyses are compared with those from other tropical regions. Differences in the surface fluxes between disturbed and undisturbed periods appear to be greater in the monsoon than in the western Pacific Ocean. However, in both regions, it is argued that the configuration of convection and vertical wind shear acts as a positive feedback to accelerate low-level westerly winds. Outstanding questions and tentative plans for the future are also discussed.