Take one Alpine valley, add a small yet sufficiently diverse city complete with "local" shops and restaurants, as well as the obligatory Thai and Indian curry houses (and sadly a MacDonnalds and Irish pub!), include cable cars up the mountain for the sensible and paths for the insane, don't forget a smattering of "tat" shops (which provide a perfect vessel for venting frustrating about the difficulty and indeed futility of buying presents until desperation forces you into a last minute purchase for some lucky loved one back home) and what have you got? A holiday?! Oh hang on, I forgot... add one congress centre complete with microphones that actually work and rooms ranging from modest in volume up to cathedral proportions, the size of which, incidently, scales almost perfectly with the inverse of the number of people inside. And yes, here we have a perfect conference venue - in fact the top congress venue of 2000 and the home for the 8th Scientific Assembly of the International Association of Meteorology and Atmospheric Sciences (IAMAS) in July 2001. The main purpose of IAMAS is to "provide the scientific community with platforms to present, discuss and promote the newest achievements in the basic fields of research and in related areas".
A sizable contingent attended the meeting from the Hadley Centre and also from other parts of the Met Office (notably MRF who apparently stayed in luxury accommodation usually reserved for royal families, presidents, and past yodelling championships winners) and also the Meteorology Department at the University of Reading. But even these familiar faces were well diluted by the meteorological (and chemical) populous of IAMAS, making the conference a truly global affair. Here follows a description of the highlights from my campaign....
The main chapters of the "Scientific Basis" volume of the IPCC (2001) report were summarised:
Bryant McAvenney - Processes in Models: While the water vapour feedback is still thought to be strongly positive, the links between free-tropospheric humidity and surface temperature have been hotly debated over the last ten years and the increased complexity in the processes realised. Nevertheless, high resolution models can simulate the filamentary structure and very low values of humidity borne out by the observations suggesting good dynamical distribution of water vapour. Meanwhile, despite the increased sophistication of cloud parametrizations, the uncertainty in cloud feedback has not reduced. In the oceans there have been advances in the heat transport and latest models display improved large-scale circulation patterns including the variability associated with ENSO. Models pick up the decline in sea-ice seen in the observations even when flux adjustment is not applied. Over sub-continental scales, there is thought to be skill in seasonal to decadal prediction in coupled models which far surpasses those of 10 years ago. Finally, the 20th century simulation of global mean surface temperature (Ts) is reproduced by some models as long as natural forcing is applied in combination with greenhouse-gas and sulphate aerosol forcing.
John Mitchell - Detection and Attribution: An important finding is that similar global-mean radiative forcing, no matter how its spatial manifestation, produces a similar global mean surface temperature response. Despite good simulation of 20th century global mean Ts there remains uncertainty, for example indicated by the inability to simulate recent changes in temperature lapse rate. John also endured one of the longest questions to be posed in the history of vocabulary.
Chris Folland - Observations: Latest estimates at 20th century global mean warming is 0.6 +/- 0.2 K. Increases in ship heights in the last 10 years may cause the recent decadal warming to be adjusted upward. There is an observed reduction in the diurnal cycle of Ts over land (from 1958-98) and a 0.0037 K/decade increase in ocean 300m temperature. There appears to be mutil decadal variability in temperature lapse rate with the differences from year to year being highly correlated (indicating the variation is not random). Arctic sea ice is diminishing (particularly in Spring/Summer) while the Antarctic shows increases although the early satellite record shows a decrease. The warming over the last century is likely the largest seen this millennium based on proxy data which suggests cooling over much of the last 1000 years followed by rapid warming over the last 150 years. CH4 and C02 concentrations are strongly correlated with Ts proxy data over glacial cycles. Sudden climate changes, such as the termination of the Younger Dryas (~11 kyr BP), are evident in the proxy record, particularly on regional scales. Notably an 8.2 kyr BP event indicates a 2 K cooling over about a century in the N. Atlantic.
V. Ramaswamy - Chemistry, Aerosols and Radiative Forcing: There have been reductions in the uncertainties in radiative forcing due to ozone. Understanding radiative forcing due to sulphate aerosol has improved (including the direct optical properties effect and indirect effects linked to precipitation efficiency and cloud lifetime) but still remains a considerable uncertainty. There is no good global mean estimate although a negative forcing is probable. Compared to previous reports the future sulphate emissions have been reduced. By 2100 it is expected that CO2 forcing will strongly dominate over sulphate forcing. Ram was bullied by a questioner asking what the radiative forcing of water vapour was and correctly stated that this was a feedback and not considered a forcing (apart from due to water emissions in the stratosphere which may exert a forcing after chemical interactions). Nevertheless it is estimated that water vapour enhances the temperature response to radiative forcing by over 30%.
Karl Taylor - AMIP model performance: Used now fashionable quarter circle plots of correlation verses standard deviation to show improvement from AMIP I to II of cloud amount in the tropics (reduced) although there is still too little stratocumulus. The upper tropospheric cold bias is still present. Column water vapour is well simulated.
William Collins - Improved Radiation in NCAR model: Lookup tables based on GENLN 2.0 produced improved longwave radiation in the water vapour rotational bands and the continuum (upper troposphere due to foreign continuum and lower troposphere for self-broadened continuum). The cloud overlap approximation has been improved from random overlap to Maximum Random overlap (ie clouds in adjoining vertical layers display maximum overlap). This strongly affects the surface radiative fluxes and the new radiation treatment improves surface land temperature, polar stratospheric temperatures and polar surface longwave fluxes. There were questions raised as to the uncertainty attached to the temperature dependence of the CKD foreign broadened continuum and how this could substantially affect the diabatic heating.
Richard Sommerville - "Can we trust Climate Models?": The different responses to CO2 for different cloud parametrizations was noted. There is a big variability in the height dependent cloud amount in models which all tend to underestimate upper tropospheric cloud amount (based on measurements with cloud radar) due mainly to over-prediction of detrainment rates rather than relative humidity (RH). More realistic low altitude cloud amount is produced by models which include detrainment schemes. It was also noted that the cloud height feedback is the strongest cloud feedback in models. Results of single column models and simulations of TOGA COARE situations are published in Journal of the Atmospheric Sciences: Vol. 57, No. 2, pp. 161-183.
John Edwards talked on Cloud Overlap and the effects of incorporating the method of Hogan & Illingworth (2000) based on radar observations. Most models use maximum overlap of vertically continuous clouds while the observations suggest a reduction from maximum to random overlap with increased altitude for continuous cloud. Results from the Met Office model show an increase in cloud amount causing increased reflected shortwave radiation and reduced OLR with increased SW heating rates above where the cloud increases are largest. Dan Bleaken talked on diagnosing tropical storms in climate models and how they respond to climate change. Rich Neale showed that by removing tropical west Pacific warm pool land points improved the simulation of rainfall and further exerted a strong influence on Europe and Canadian Ts indicating the importance of this region for global climate. The results of Yang and Slingo (2001) suggest an error in the diurnal cycle of convection over land in the Met Office model which could explain some of the errors in rainfall in this region. By increasing the time resolution of radiation calls (to 30 minutes from the usual 3 hours) it was found that the amplitude of convection increased and impacted the time mean precipitation but the phase of the diurnal cycle was not improved.
Martin Werner - Glacial-interglacial changes in Mineral Dust:
Mineral dust is emitted to the atmosphere from areas with sparse vegetation, low soil moisture, and sufficiently intense surface winds. The dust content of the atmosphere has varied in association with natural climate change and was massively increased during glacial periods, as shown by stratigraphic records from polar and tropical ice caps, marine sediments, and loess deposits. Thus, due to the radiative properties of dust and the effect on the nutrient ecosystem balance and chemistry, this constitutes a complex feedback to the climate system. Much of the source of mineral dust is from low latitudes. One potential sources is the drying of lakes (and also the increase in coast-lines as sea level falls in glaciation??). Just a dry near surface layer is required for dust emission.
Ellie Highwood showed, using the Reading Intermediate General Circulation Model, that the impact of a 1783 "effusive" (non-stratospheric aerosol injecting) volcanic eruption (gradual emissions of SO2 into the troposphere) was to produce a cold northern Hemisphere summer as borne out by proxy observations.
Rowan Sutton - N. Atlantic decadal climate variability: The N. Atlantic variability is dominated by solar forcing during the early part of the 20th century and greenhouse gas forcing in the latter part. However in the mid-20th century, some other effect is thought to be important (i.e. changes in thermohaline circulation (THC)). It was found that the tropical sea surface temperatures are important in determining the predictability of N. Atlantic climate. It is argued that because global climate does not seem strongly affected by the N. Atlantic sea surface temperatures, interactions between this region and the THC and tropical sea surface temperatures are the key link between global climate and the N. Atlantic. Tropical convection and its sensitivity to the THC may be important in this process.
Colin Price talked on the correlations between global lightening activity, which may be measured accurately from a single site, and upper tropospheric water vapour. Problems with using such a parameter to monitor water vapour are: 90% of lightning occurs over land, the links between amount of lightening and upper tropospheric water vapour need to be established on more time and space scales. A 5-day periodicity in global lightening is apparent.
Also, I talked on the variability of the tropical top of atmosphere radiation budget (1979-2000) and a possible decadal variability of cloud radiative effect not simulated by climate models.
Bob Cess Surface Downwelling Longwave Flux Algorith Development: Using ARM measurements, the clear-sky surface downwelling LW flux (SDLc) can be parametrized in terms of the surface emitted LW flux (SUL) and the precipitable water content (PWC):
Joseph Michalsky ARESE II SW Radiometer Calibration: Using single BSRN pyranometers to measure the total SW flux is problematic; it is better to use diffuse+(directxzenith). The uncertainty in aircraft SW obs is of order 3% while ARM surface direct measures are accurate to +/- 3 Wm-2 and for diffuse about +/- 5 Wm-2 (although there is no standard calibration as yet). When it is cloudy the diffuse beam is more accurate as there are no problems with the alignment of the shading apparatus.
Rolf Philipona talked about the greenhouse effect over the Alps. However this study suggested the surface downwelling LW flux as a good measure of the greenhouse effect because it has direct impact on the surface temperature. This appears misguided as much of this flux comes from the near surface layers and therefore does not account for the atmospheric adjustment to radiative cooling/heating important to feedbacks.
Paul Stackhouse talked on the WCRP/GEWEX surface radiation budget project (SRB) while the session convenor, Atsumu Ohmura talked on observed GEBA/ERBE estimates of the global radiation budget (Wm-2), SURFACE: Sensible Heat= 17, Latent Heat=85, Incoming SW=169, Reflected SW=27, SDL=345, Surface LW Emission=385, TOP OF ATMOSPHERE: OLR=235, Incoming SW=342, Reflected SW=102. This gives atmospheric SW absorption of nearly 100 Wm-2, generally larger than climate models. A further paper was presenting trying to estimate anomalous cloud absorption of a statocumulus deck using surface BSRN data, Meteosat top of atmosphere fluxes and modelling the in-between radiative transfer using MODTRAN. The estimate of cloud absorption (88 Wm-2) was generally thought to be erroneous by the audience with rather inaccurate cloud albedo assumed.
Finally, I talked on methods for using collocated surface BSRN and top of atmosphere satellite fluxes to validate climate models. (Powerpoint talk)
This session inevitably began the night before with a now regular trip to the casino instigated by Jim Haywood (see GORDON RESEARCH CONFERENCE trip report for detailed discussion of casino etiquette - no convertible this time though).
Brian Soden - Sensitivity Studies of Water Vapour Feedback
First the IPCC statements on water vapour feedback over the last 11 years were presented. These appear to exhibit a cyclical variability! In 1990 water vapour feedback was judged one of the best understood of all feedbacks. In the 1990's the uncertainty relating to free-tropospheric humidity begged questions as to the reliability of the simulations. By 2001, while the extra complexity involved was noted, the essence of the 1990's statement was essentially intact. Models generally all predict an increase in column water vapour with temperature of 7% per Kelvin (in line with the Clausius Clapeyron relationship). Providing that the conservation of relative humidity predicted by climate models is robust, the water vapour feedback is almost certain to be strongly positive. Sensitivity studies were presented for LW and SW fluxes to changes in water vapour with and without clouds. When clouds are included the tropics and low altitudes become less important. The strongest sensitivity is attached to the free tropical troposphere although the extra tropics are still highly important.
I talked on a diagnostic analysis of clear-sky feedbacks in HadAM3 and GFDL AGCM runs(Figures), including discussion on problems with the ECMWF 15-year reanalysis moisture variability(Figures). The conclusions of this work are: While both models give similar clear-sky feedbacks (diagnosed from the top of atmosphere OLR regression with Ts) they produce different water vapour and temperature lapse rate feedbacks.
Unfortunately John Bates didn't talk on Dynamics of the Upper Tropospheric Water Vapour as he never turned up.
Anand Inamdar - New Earth Obs in the water vapour bands and continuum: The changes in relative humidity in the lower troposphere affect the atmospheric greenhouse effect due to changes in transmission in the longwave window region due to continuum absorption. Upper tropospheric humidity changes alter the transmission in the main water vapour bands which also affect the greenhouse effect. Using measurements in the window and non-window region of the LW spectrum therefore provide information of the height dependent water vapour greenhouse effect and it variability. Here observations from CERES on TRMM were presented. Absorption in the window increase quadratically with the scaled water vapour amount while in non-window regions the absorption increases with the logarithm of the scaled water vapour amount in line with theory.
Dennis Hartmann tried to explain why the top of atmosphere net radiation is relatively stable throughout the tropics despite radically different dynamical/cloud regimes. Personally I wasn't convinced...
Bob Ellingson - The ARM Enhanced Shortwave Experiment II (ARESE II): This was the last talk I saw and probably one of the most interesting. There has been considerable debate on whether clouds absorb more shortwave radiation than predicted from radiative transfer theory. if they do, this could have a large impact on our climate simulations. ARESE I was conducted in 1995 using aircraft, satellites and ground-based instruments and found solar absorption by clouds to be greater than model predictions. ARESE II, conducted early 2000 hoped to overcome the limitations encountered from ARESE I and to resolve some of the conflicting results. This was achieved
by (a) increasing the number of samples of extensively cloudy days, (b) using multiple instruments to measure the same thing, (c) conducting before and after intercomparison and/or callibration, (d) including spectrally resolved radiometric measures, (e) comparing aircraft measured albedo's with broadband satellite data. Uncertainty due to horizontal solar flux divergence were reduced by measuring 500 nm fluxes. Flights close to noon were used to reduce the zenith angle dependence biases. The average profile absorptance was found to be 0.225. For clear-only profiles the value was just 0.12 and overcast only 0.25 suggesting excess solar absorption by clouds. However, again there were inconsistent results between different instruments. Generally all observations indicated at least some excess cloudy solar absorption compared to models. The data is available at http://armuav.atmos.colostate.edu.
Overall it was an excellent conference which helped cement existing collaborations as well as initiate new ones. The format was also excellent, with plenty of free time over the weekend. This also allowed busy people to fly in just for the pre-weekend or post weekend session (or just afew fleeting hours in John Mitchell's case!) and also meant people giving talks at the end of the conference managed to retain their audiences. Apart from this there was some great walking to be had, great food to be eaten (incidently one Hadley Centre member managed a new world record in the number of puddings eaten in a week) and great casino's to visit so all in all a thumbs up.
SDLc=SUL+74ln(PWC)+25(ln(PWC))2.
The all-sky flux needs information on cloud base height. However this is problematic in situations of broken cloud and it was found that Liquid Water Path is a good surrogate for the "radiative" cloud base height:
SDL=SUL+136ln(PWC)-33(ln(PWC))2+10ln(1+2272LWP).
Artctic regions (with temperature inversions) and uncertainties involving aerosol and near surface minus surface temperature differences introduce inaccuracy in the parametrization.
Final Comments
![[Winnings=donuts!]](http://www-hc/~hadri/JPG/gordon_donut.jpg)