Report of the International Institute for Theoretical and Applied Physics (IITAP)Regional-Scale Climate Model Intercomparison Workshop

Ames, Iowa
November 17 - 18, 1994
Contents

Welcoming Remarks by Dr. Takle

The group was welcomed on behalf of IITAP, Iowa State University, and the Department of Geological and Atmospheric Sciences by Dr. Gene Takle, IITAP Thrust Area Coordinator for Environmental and Earth Sciences (EES). He outlined the general goals of IITAP including the application of science (physics) to environmental (meteorological) problems, and the training of developing countries' scientists. He then gave the mission statement of the EES of IITAP which is:

To address basic and applied problems in environmental and earth sciences by use of basic physical science principles, and to use this mechanism for creating and expanding partnerships with scientists from developing countries to enhance local science infrastructure.

Dr. Takle noted that there are many environmental problems that are shared by both developed and developing countries where cooperation could help solve these problems. Also he noted that there are many unique databases in developing countries that local scientists have access to and that could be shared with scientists in developed countries. An institution like IITAP could help train scientists as they analyze these databases. The importance of trained scientists in providing local solutions to environmental problems and to provide guidance to local policymakers in their decision making, was also mentioned.

Takle then noted that in the above context the initial thrust activities of the earth and environmental sciences division would be regional-scale climate modeling, studies in shelterbelt turbulence, and the development of an electronic classroom. The main goal of the regional- climate modeling effort would be to improve the status of regional-scale meteorological models and to promote their use by developing countries for solving environmental problems. It was also made clear that the proposed activities were not intended to be performed exclusively by or at Iowa State.

Takle then noted that the main mission of the workshop was to design a first experiment for the proposed regional-climate model intercomparison. Dr. Martin Jischke (President of Iowa State and UCAR trustee) was then introduced to the workshop. He made a few general welcoming remarks and gave a brief historical background about the origins of IITAP. He also noted that this was the first technical workshop to be hosted by IITAP. Jischke then introduced the Keynote speaker Dr. Richard Anthes (Director of NCAR/UCAR).

Keynote address by Dr. Richard Anthes

Anthes started off by noting that climate is much more than 30 year averages of monthly mean temperatures and precipitation. It also includes a description of the frequency and intensity of extreme events such as hurricanes, droughts, floods, thunderstorms, and prolonged hot and cold spells. He also noted that climate includes a description of the temporal and spatial variability of climate variables. In short; climate is the integral over time of the weather, and in order to know the integral (climate) we have to know how the integrand (weather) behaves.

Anthes then presented a schematic of the components that are represented in 3-D climate models. He noted that the large-scale factors that determine the earth's climate (the radiation balance and the rotation of the earth) are included in 1-D and 2-D climate models that are adequate for describing the gross features of the earth's climate. The 3-D global models do a good job describing the large-scale spatial and temporal structure of the climate, but their coarse resolution is not adequate for describing climate on scales finer than a few hundred kilometers. Since information about climate is needed at these scales, motivation is provided for regional- scale climate modeling.

He then gave some hypotheses that form the basis for regional climate modeling, which were:

He concluded that if GCMs predict the large-scale well then mesoscale models may be embedded via a 1-way interaction to add mesoscale detail to regional climates. However, a cautionary note was added that if the GCM does not do well at the large-scale, then the mesoscale model will produce a poor or useless simulation.

Anthes then showed some results from western U.S. January climate simulations with MM4 (60 km grid) embedded in an R-15 GCM (Giorgi and Bates, (1989)). The results showed the refinements in the structure of precipitation that were possible using the mesoscale model. He also showed some recent work of Sousounis and Fritsch (1994) regarding the aggregate influence of the Great Lakes on that region's climate. Their modeling work showed that the Great Lakes tend to modify the large-scale flow by producing a regional low near the surface ("a lake-aggregate circulation") and an anticyclonic circulation near 700 mb.

Anthes then demonstrated the ability of mesoscale models to capture extreme events such as a rapidly intensifying Atlantic cyclone by showing video of MM5 simulations of the January 4, 1989 ERICA IOP 4 storm (Reed et al. (1994)). The video was very interesting and it included a demonstration of the sensitivity of the model simulations of the storms intensity and life-cycle to the inclusion of friction, horizontal diffusion and surface physical processes.

Global climate model intercomparison; AMIP experience

The next speaker was Jerry Potter from Lawrence Livermore who shared with the workshop the Program for Climate Model Diagnosis and Intercomparison (PCMDI) experiences with the Atmospheric Model Intercomparison Project (AMIP). The purpose of AMIP is to document model performance under realistic conditions, facilitate model diagnosis and validation, and provide a benchmark for sensitivity studies. The first experiment that was conducted to help achieve the three goals was a simulation of the global climate between 1979- 1988 with prescribed monthly mean SST's, sea ice, CO2, and the solar constant. Potter noted that this was basically an "El Nino" experiment since the overwhelming climate signal (perturbation) in that period was the El Nino of 1982-1983. He also stated that they were planning to extend the simulations 6 more years through 1994. After the simulations were done by the groups involved (30 major climate modeling groups in the world) an in-depth analysis was to be performed via diagnostic sub-projects. Part of the AMIP experience was that many of the initial simulations had to be redone because many of the variables required by the subprojects had not been saved, or had been saved too infrequently. He stressed the importance of careful planning in deciding what variables should be saved and at what frequency ("Save as much as you can!" was the catch cry).

Potter noted that Lawrence Livermore had offered computer time to some of the modeling groups involved and that the need for ensemble runs had placed constraints on the resolution at which the models could be run. He also noted that 50% of their staff time (20 scientist/computer technicians) was devoted to AMIP, and as an example he stated the case of one of their staff who had spent a year documenting all the models involved in the experiment (T.J. Phillips, (1994)). Potter noted the support that the DOE and WGNE/WCRP had given AMIP.

The most important aspect of AMIP has been the identification of systematic errors amongst the models. As an example Potter noted that diagnosis of the tropics revealed anomalously low short-wave radiation absorption in all the model atmospheres.

The data structure had to be self-describing for atmospheric modeling use, and they had to design an interface between the database (of model results) and a diagnostic library which included visualization tools capable of handling various formats (netCDF, GRIB, HDF, DPS, and GraDS). They also have a Mosaic home-page (URL http://www-pcmdi.llnl.gov/) where people will be able to extract data to their site or perform some calculations with some variables at data central. He noted that consistency among models as to units used and levels at which output was available should be agreed upon, or procedures to handle differences should be sorted out before any simulations are done. Storage limitations and data access could also be a problem (model histories- 6 hr state of model = 100 GB per run). They have a facility for mass storage with a capacity of 20 Terabytes of storage at data central which is needed to handle the data efficiently. The analyses used for validation of the simulations will come from NMC, NASA, NRL, and ECMWF analyses.

A list of some of the subprojects was then given. They could basically be divided into 3 categories involving the study of processes such as cloud radiative forcing, regional climates, and phenomena such as blocking. The support of these sub-projects was one of the tasks assigned to AMIP management. Other tasks assigned were communication with the groups (newsletters), quality control of the analyses and model output, and the organization of conferences and workshops. The importance of quality control of the model output and the need for it to be converted to standard format could not be understated. Finally, he noted that it was important this intercomparison was not meant to be a "beauty contest" among the models. A reference for the AMIP project was given as Gates, (1992).

Mesoscale forecast model intercomparison; COMPARE experience

The next speaker was Clement Chouinard of the Canadian Atmospheric Environment Service group that organized the first case of the Comparison Of Mesoscale Prediction And Research Experiments (COMPARE) project, which was designed to compare the performance of mesoscale models in forecasting synoptic events such as maritime cyclogenesis. The objectives of COMPARE are to "propose and perform experiments in a collaborative manner to further understanding and predictive capability at the mesoscale", to "identify important issues of mesoscale research and prediction that may be addressed by numerical experimentation", and "to establish a testbed of representative mesoscale cases using raw data, assimilation systems and analyses of the highest possible quality".

The following scientific issues at the mesoscale have been identified:

Chouinard then described the set of standard procedures and methodology for producing the analyses that would be used to drive the models. A minimum frequency of 6 hours was to be used for the data assimilation cycle. Analyses were to be provided for at least the 24 hour period up to the initial forecast time. The analyzed data were to be formatted on 2 regular lat/lon grids, the first being on a 1 degree global grid, the 2nd being "semi- hemispheric" at 0.5 degrees. They also recommended that the models should be run at the highest possible resolution compatible with the data density. Minimum requirements for vertical resolution at which data to be supplied to the models was set at 25 hPa between 1000 and 100 hPa, plus the analyzed mandatory levels above 100 hPa. Values of p-s, z-s, u-10m, v-10m, T-2m, q-2m coherent with the data assimilation procedure were also provided along with a coherent set of surface geophysical fields such as SST, z0, albedo, soil-moisture, ice-cover etc, and participants were highly recommended to use them for their simulations. Some of the guidelines for the participation were to insist that the most recent version of models be used, and that the global and regional data fields used to drive the models be consistent between models so as to allow for a good comparison of the LBC driven and LBC free approaches.

The first COMPARE case was based on a marine cyclogenesis event from CASP (Canadian Atlantic Storms Project). Analyses were prepared using the RPN/CMC regional data assimilation system driven by NMC global analyses. A description of these techniques is given Chouinard et al. (1994). They produced a participation package consisting of an analysis tape; reading and decoding software; ECMWF's documentation of the GRIB format; a tape-reading guide; a description of suggested algorithms for interpolation and relations between thermodynamic and moisture variables; and a questionnaire on attributes of the participants models. Chouinard noted that an important aspect of the first experiment was to standardize procedures so as to facilitate the running of future experiments. He then described the set of proposed simulations that was to make up the first experiment. These were basically sensitivity tests to resolution and initial and lateral boundary conditions.

Chouinard then presented some results from the first project experiment, and noted the achievements thus far of COMPARE. The evaluation of the results of 11 modeling groups has been completed and presented at a workshop, they developed a software package to verify the forecasts against a set of quality-controlled observations and analyses, and consolidated all the results from the mesoscale model runs into a database which was distributed, along with visualization software to an independent evaluation group, and then to each participant.

Finally he noted that it had taken 5 years to get to this stage, they had many problems with data exchange and had to validate, separately at least 100 experiments (visual inspection, they had a staff of 4.)

Federico Mayor, Director-General of UNESCO, then briefly addressed the workshop. A transcript of that address is included in Appendix A. Mayor made these three points: (1) we often talk about applied science, but basic science is also very significant: "there is no applied science if there is no science to apply!" (2) National politicians and ministers need more scientific input for decision making and these scientific elements must be presented in an understandable way. (3) Thirty thousand Ph.D. scientists from sub-Sahelian Africa are now working in developed countries, where they have better opportunities to continue their research. UNESCO is trying to reverse this trend by bringing scientists from advanced countries to the south for periods of 2-3 months to facilitate rapid transfer of technology.

Climate needs of developing countries (Semazzi)

The next speaker was Frederick Semazzi (North Carolina State) who talked about the climate needs of developing countries and some interesting work relating to aspects of African climate. Semazzi first noted the hierarchy of climate scales which are important to Africa ranging from the continental to watershed scales. He presented some observational results from Nicholson (1986) which showed rainfall departures over Africa for the periods 1950 - 1959 and 1968 - 1973. These results seemed to suggest that there were 2 modes for rainfall anomalies in Africa, the first where central Africa is drier than normal, and the regions north and south are wetter than normal; and the second where central Africa is wetter than normal and the other regions are drier than normal. He then showed some GCM results from simulated years within these periods that agreed qualitatively with the observations (Semazzi et al., (1994a), and Semazzi et al. (1994b))

Semazzi then discussed the work of Owen et al. (1988) about seasonal forecasts of Sahelian rainfall made with the UKMET GCM, and the work of Semazzi et. al. (1993) with a nested regional model (inside CCM1) in trying to get more detail in historical seasonal simulations of Sahelian rainfall. He described the urgent need for seasonal forecasts since drought in Africa often means famine in the afflicted areas. He also talked about the work of the Drought Monitoring Center in Nairobi, and how mesoscale models might help in decision making with the Qattarra hydroelectric project in Egypt. There is a plan to fill the Qattar depression with water from the Mediterranean via a canal. Electricity would be generated by the flow of water through the canal. This flow would be sustained because of the intense evaporation of water off the newly created lake due to the extremely arid surroundings of the lake and its shallowness. Mesoscale models could be useful in the planning of this project since they could provide estimates of evaporation off the lake under the new mesoscale wind regime. He then talked briefly about how vegetation changes in the Amboseli National Park of Kenya had been affected by a prolonged wet spell, which had raised the water table enough to reach the roots of existing trees. Since the groundwater was highly saline, the trees died out.

Semazzi then discussed the needs of African scientists and institutions if they wished to take advantage of advances in climate and mesoscale modeling. The basic requirements were computing power, maintenance, telecommunications and the associated infrastructure, observational data and qualified personnel. He then went on to discuss the modes for enhancing opportunities in regional climate modeling in developing countries, such as the research mode and a semi-operational mode. This semi-operational mode, he envisioned, would be a clearing house for mature climate modeling research products, and would be useful in developing a coordinated strategy in minimizing overlap between research at the continental, regional and watershed scales.

After the talk there was a brief discussion of the negative impact on scientific research in developing countries arising from the multi-tiered data cost structuring policies now being adopted by some countries.

Hydrological considerations for regional models (Krajewski)

Witold Krajewski (hydrologist from the University of Iowa), the next speaker, talked about hydrologic considerations at continental, regional and finer scales including issues relating to streamflow.

Krajewski started off by mentioning that they were working with NASA under the TRMM validation program. This program has a goal of extracting rainfall rate estimates from both satellite and ground radar returns. Selection of algorithms to be used with the ground radar data is made through an intercomparison of candidate methods. This framework is similar to the proposed intercomparison of the mesoscale model discussed herein. Krajewski then noted that the main characteristic of observations of hydrologic parameters (e.g. rain rates and radar reflectivity) were their tremendous variability, and that the major statistical assumption used in analysis (e.g. developing regression relationships) is stationarity. He also noted that variability in rainfall pales in comparison to the variability in hydraulic conductivity (which is a function of soil-moisture, which is itself highly variable - there are order-of-magnitude differences in conductivity between wet and dry soils, and sandy and clayey soils). Fortunately over time and space, nature integrates out much of this variability and streamflow is a measure of this integral. He then demonstrated how streamflow reflects long term trends in precipitation by showing spring precipitation and summer streamflow autocorrelations. He also noted the differences in variability in streamflows for drought and flood periods.

Krajewski then showed some composite maps of temperature and precipitation correlations with summer soil-moisture levels based on the work of Konstantine Georgakakos and his colleagues at the Hydrologic Research Center in San Diego. Their work demonstrates the importance of soil moisture in the top two metre zone in controlling summer flooding in the Midwestern basins. metre zone He then showed some scattergrams of precipitation versus moisture convergence, evaporation and precipitation recycling for the Midwest, and he pointed out the location of summer of 1993 on these scattergrams. The results indicated that July 93 was an outlier in terms of precipitation and moisture convergence, but not in terms of evaporation and percent recycled moisture. (These results were from Allen Bradley, a hydrologist at the University of Iowa.)

Krajewski noted that an important feature of mesoscale model validation in terms of the hydrological cycle is that the models must preserve the statistical character of precipitation variability across a range of scales. One statistical feature important for development of parameterizations of the hydrologic components of the climate models is a strong linear relationship between mean areal rainfall and fractional area covered by rain. A similar relationship holds in time. He then showed some scattergrams and linear regressions from GATE I and GATE II showing the relationship between area averaged rain rate and fractional area coverage.

He then talked about the River Forecast Service (RFS) extended (1-3 month) streamflow predictions. He demonstrated, based on work by Georgakakos, quantitative benefits from operating water resources systems if long term streamflow forecasts are available.

References given were Guetter and Georgakakos (1993) and Roads et al. (1994).

Krajewski's talk was the last of the formal presentations. Less formal discussions concerning the objectives of the regional climate model intercomparison, and the design of the initial experiment with consideration of space and time domains, initial and lateral boundary conditions, surface physics, convective and radiation schemes followed. Practical issues relating to verification, what variables ought to be saved, standardization of reporting procedures, and funding sources were also discussed.

Experiment Design

Mission and Action Statements

Dr. Bill Gutowski (Geological and Atmospheric Sciences Department, Iowa State University) then led the discussion of the project mission and a set of action statements.

The mission statement adopted was "to advance the status of regional climate modeling by evaluating the strengths and weaknesses of current models and their component procedures through systematic, comparative simulations". The four action statements were:

  1. for participants in Project to Intercompare Regional Climate Simulations (PIRCS - suggested name for the project) to actually perform the experiments,
  2. to provide straightforward access for developing countries scientists to the datasets produced for the project by the participants,
  3. to develop collaboration between modeling groups and researchers in developing countries who wish greater involvement in regional climate modeling, and
  4. to coordinate its activities with related international programs occurring under the auspices of the WMO and the U.N. Food and Agriculture Organization.

In the general discussion that followed the following points were made (many of the points raised related to the scope of the participation):

There was some discussion as to the name (acronym) of this experiment. It was noted that the current title (Regional-scale climate model intercomparison project) doesn't specify limited area domain models. It was suggested to leave out MIP since this implied connection with AMIP. PIRCS seemed to be the early favorite.

It was suggested that the group encourage non-modelers to utilize the sets of model output generated in the form of subprojects a la AMIP. This would also be consistent with IITAP goals.

Scientific Objectives

The first discussion period was led by Filipo Giorgi (Climate and Global Dynamics division of NCAR) and this involved the setting of scientific objectives of this effort. After some time these were prioritized, as many people felt that the experiment should be as focused as possible, at least initially.

Giorgi listed some general scientific objectives within the regional climate modeling community. These were

Scientific questions that Giorgi thought ought to be answered in this effort were suggested and were prioritized as follows:

The rest in no particular order were

This list prompted much discussion and many points were raised about the first experiment and project in general, some of which were not resolved. Listed below are some of these points.

Time and Space Domain (Hesselbjerg Christensen)

The last discussion session of the Thursday was led by Jens Hesselbjerg Christensen (Danish Meteorological Institute) and during this session it was decided what the domain and simulation period would be.

The first experiment, designated 1a, was decided to be a simulation of the summer 1988 drought over the central United States. This had the advantage of overlapping with the AMIP simulation period. Another potential experiment involving the simulation of the 1993 summer, when flooding occurred over much of the same region, was given the designation 1b.

Other possibilities discussed were

One of the reasons for choosing 1a and 1b over N. America was the availability of observations for validation. But it was argued that in future experiments the same arguments would be used against any experiments with simulation domains in developing countries. Therefore the participants were urged to seriously consider such a prospect for the second experiment.

It was suggested that experiments be done using both ECMWF and NMC analyses in order to address the issue of noise in LBC's. Given the computing limitations, it was suggested that the internal domain be 70 (lon) x 60 (lat) x 20 (vertical - max) with an extended grid of 10 grid points along each lateral boundary. The resolution would be 50 or 60 km. Details of the exact boundaries were left to be decided by someone with a map of the proper projection and calculator.

Other considerations

Surface forcings (Hesselbjerg Christensen)

Initial conditions needed are for:

At this point the group disbanded for the day.

Many attended the public lecture in the evening by Dr. Mayor entitled "Culture of Peace: Science, Education, and UNESCO". At the start of the Friday morning session, Thursday afternoon's discussion was briefly summarized.

Initial and Lateral Boundary Conditions (Laprise)

This discussion was led by Rene Laprise of the University of Quebec at Montreal.

One main question was whether the boundary conditions were to be generated at ISU then distributed? (Yes)

Then 4 other questions that had to be asked and answered related to

Verification (Chouinard)

This discussion was led by Clement Chouinard. The first point he made was that thought had to put into our objectives so we could match model output to these. Here it was decided that we wanted to show finer detail than is currently realized by GCMs, and that moisture, heat and energy budgets would be key diagnostic variables.

Output variables could be classified as dynamical (u,v,q,T,p,ql) and physical (fluxes). Averages of these quantities could be sent to ISU/IITAP for analysis, but histories could be saved at each institution. Hourly histories for surface budget calculations could also be sent to ISU/IITAP, where useful statistical diagnostics (such as monthly means, standard deviations, etc) or derived variables such as growing degree days could be computed. Other variables that were suggested as being useful were minimum and maximum temperatures, and vertically integrated moisture.

The levels at which the variables would be saved were suggested as 250, 500, 700, 850 mb for the dynamical variables, as well as at the 10 m and 2 m levels in the surface layer. Precipitation and evaporation was suggested as being saved in mm every 6 hours, and snow depth in mm each day.

The length of integration was decided as being two months but thought was given to a 3 month simulation with a 1 month spin up period for soil moisture.

Miscellaneous

The group the went onto discuss some other miscellaneous but nevertheless important topics. It was suggested that the workshop endorse the idea of encouraging an intense observing period in order to create an African dataset.

The global change course URL was distributed.

Discussions then centered on a timetable for finalizing a design for the first experiment - this was set as the end of January 1995. A workshop summary is to be sent to the participants by January 1, 1995.

Funding sources were then discussed. Possibilities included NSF, the World Bank (which supports countries with specific projects), DOE, and with some possible seed money from IITAP. It was noted that with an emphasis on training developing countries' scientists (via IITAP) and providing them with useful tools (mesoscale models) that their countries can use, this project should generate a lot of interest. A possibility of having a future workshop sponsored by NATO was also suggested.

The workshop then ended.

References

Chouinard, C., J. Mailhot, H. L. Mitchell, A. Staniforth, and R. Hogue, 1994: The Canadian regional data assimilation system: Operational and research applications. Mon. Weath. Rev ., 122, 1306-1325.

Gates, W. L., 1992: An AMS continuing series: Global Change. AMIP: The Atmospheric Model Intercomparison Project. Bull. of the Amer. Meteor. Soc. 73, 1962-1970.

Giorgi, F., and G. T. Bates, 1989: On the climatological skill of a regional model over complex terrain. Mon. Weath. Rev., 117, 2325-2347.

Guetter, A. K., and K. P. Georgakakos, 1993: River outflow of the conterminous United States, 1939-1988. Bull. Amer. Meteor. Soc., 74, 1873-1891.

Nicholson, S. E., 1986: The spatial coherence of African rainfall anomalies: Interhemispheric teleconnections. J. Climate Appl. Meteor., 25, 1365-1381.

Owen, J. A., and C. K. Folland, 1988: Sea-surface temperature and tropical rainfall. In "Recent Climate Change" S. Gregory, Ed. Bellhaven Press., London, pp 41-52.

Phillips, T.J., 1994: A summary documentation of the AMIP models. PCMDI report #18.

Reed, R. J., Y-K. Kuo, and S. Low-Nam, 1994: An adiabatic simulation of the ERICA IOP 4 storm: An example of quasi-ideal frontal cyclone development. Mon. Weath. Rev., 122, 2688-2708.

Roads, J. O., S-C. Chen, A. K. Guetter, and K. P. Georgakakos, 1994: Large-scale aspects of the United States Hydrologic cycle. Bull. Amer. Meteor. Soc., 75, 1589-1610.

Semazzi, H. F. M., N.-H. Lin, Y.-L. Lin, and F. Giorgi, 1993: A CCM1-MM4 nested model study of the influence of seas surface temperature anomalies JGR, 20, 2897-290.

Semazzi, H. F. M., B. Burns, N.-H. Lin, Y.-l. Lin, and J. E. Schemm, 1994a: A GCM study of the teleconnections between the continental climate of Africa and global sea-surface temperature anomalies. J. Climate (Revised).

Semazzi, H. F. M., N.-H. Lin, B. Burns, Y.-l. Lin, and J. E. Schemm, 1994b: Teleconnections between the climate of Africa and global sea-surface temperature anomalies. J. Climate (Revised).

Sousounis, P. J., and J. M. Fritsch, 1994: Lake-aggregate disturbances. Part II: A case study of the effects on regional and synoptic-scale weather systems. Bull. Amer. Meteor. Soc., 75, 1793-1811.

Appendix A

Transcript of Federico Mayor's (Director-General UNESCO) remarks to the workshop.

"I would like to make three points, which for me are very important, .... the first point, speaking as a scientist, we talk often about applied science, I would insist that there is no applied science without science to apply. What I mean by this is that basic science is always extremely important, and that we must try to persuade the financial sources, the government, that basic science is extremely important. Because, again when I was minister of Education and Science in my own country I remember the Minister of Finance was always saying "Yes, but applied research!", I was always repeating "there is no applied science if there is no science to apply!". It was very important because sometimes we have this kind of disposition from the very beginning to apply but not to make fundamental research."

" ... the second point, this time as a politician, as a minister (that I have been) I know at what point it is important to have elements of scientific input in decisions today. There are many decisions which are based on economic and social parameters. We need more science, in such a way that it is feasible for the politicians to realize these elements. We need also to make an effort to present these elements in such a way so they understand and take into account the scientific input as one of the aspects of their decision. It is in this respect, I was telling Professor Vary, that it is our intention with UNEF and other agencies of the United Nations and NGO's and universities to start producing every three months a "state of the planet report". Every three months, because now the reports are made every year, or every 2 years. And for politicians this is too long, they have more immediate problems, they need to have this kind of scientific input in their briefings.

And also we have some problems today that should have more science at the municipal level, at the village level, for example, the utilization of fertilizers is very costly. We have no particular analyses of what kind of fertilizer we must utilize for any particular soil. No guidance to follow for recommended applications. We aren't getting the benefits from the fertilizer and polluting in the process."

...."the third point I would like to make, this time as Director General of UNESCO is ...... we have at this moment thirty thousand Ph.D.'s from south Sahelian Africa working in the most developed countries. Today they choose these countries in order to do their own research. It is for this reason that we are talking of having the possibility of doing the opposite; for professors and scientists from advanced countries to go to the south, not for many months, but for 2 or 3 months and to make a very rapid transfer of knowledge, in order that people can benefit on the spot and they also have a better feeling of sharing, because they realize that scientists from advanced countries are actually coming to their country. We create what they say, a unity between universities and central reserve systems with UNESCO's chairs. Here you have a tailor made design for work. The best scientists go there for 2 and 3 months, each one, then in 9 months you can have a very comprehensive transfer of knowledge".

Copyright/Trademark Legal Notice