Julian Adem

Low resolution thermodynamic grid models

In the Conference on the Physical Basis of Climate and Climate Modelling in Stockholm, the author presented a paper on simple climatic models (Adem, 1975a), in which a review on this subject was given, which included zonally averaged models. In this paper, a review of grid models with low resolution will be presented, with special emphasis on the development that took place after the Stockholm conference. Zonally averaged models will not be included. In the early sixties, the author developed a time averaged model for predicting mean anomalies of the temperature, the precipitation and associated heating functions and circulations for periods of a month or a season. The basic idea behind the model is to use the thermodynamic energy equation applied to the atmosphere--ocean--continent system as the forecasting equation. The other equations are used diagnostically together with parameterizations for the heating functions so that all the variables are expressed as functions of the surface temperature and the mean atmospheric temperature. A second essential assumption is the use of an exchange coefficient of the order of 3 X 101° cm 2 sec -1 for the horizontal turbulent transport of heat in the atmosphere and one of the order of 3 X 108 c m 2 sec i for the transport in the ocean. Recently, Sellers (1976) has adopted the author s approach, but instead of applying the thermodynamic energy equation to an atmospheric layer and to a subsurface layer separately, he applies it to a single layer that includes the surface--atmosphere system. This represents a simplification with respect to the author s model, which restricts the interaction between the surface and the atmosphere, because the difference between the surface temperature and the atmospheric temperature, essential in the interaction, is no t generated in such a model. On the other hand, Sellers uses a global model instead of a hemispheric one, but his computat ions are restricted to simulate the mean month ly condi-

Review of the development and applications of the Adem thermodynamic climate model

Recent advances in the development and applications of the authors Hemispheric Thermodynamic Climate Model are presented. The model has been adapted to simulate the climates from 18 kyr BP to the present, and to study the effect of the ice sheets, the insolation anomalies and the atmospheric CO2 content on such climates. The surface ocean temperature anomaly is also simulated in the model, and comparison with values of CLIMAP (1981) for 18 kyr BP shows some agreement. A long series of numerical experiments have lead to the improvement in prediction of the monthly surface temperature anomalies. Verification of 93 predictions over the contiguous United States of America shows a useful skill in the predictions. The model is being adapted for forecasting in the Mexican Republic. Experiments to improve the skill in prediction of surface ocean temperature anomalies in the Northern Hemisphere have been carried out, and using a fine resolution grid, the model has been used to simulate the annual cycle of the normal sea surface temperatures in the Gulf of Mexico, that agrees well with observations.

Numerical experiments on the prediction of sea surface temperature anomalies in the Gulf of Mexico

Abstract The conservation of thermal energy equation applied to the mixed layer of the ocean, has been used to predict the sea surface temperature anomalies (SSTA) and the month-to-month changes in the Gulf of Mexico. The model includes the horizontal transport of heat by mean ocean currents and by turbulent eddies, as well as the heating by short and long wave radiation, evaporation and sensible heat given off to the atmosphere. A comparative study is carried out on the relative importance of the heating and transport terms. An objective verification of the skill of the predictions is presented for each season and for the whole period from March 1986 to February 1987. The predictions using only the heating terms have some skill over the control predictions (persistence and return to normal). The skill is substantially increased when the horizontal transport of heat by turbulent mixing is included in the model. The incorporation in the model of the Ekman wind drift current anomalies computed from the anomalous surface geostrophic wind improves appreciably the skill of the predictions in winter and fall. The mixed layer depth computed using the Kraus and Turner theory with dissipation, shows that the depths in summer and fall are shallower than in spring and winter. The effect of the shallow mixed layer depth in the model becomes apparent in summer and fall, improving the skill of the predictions in these seasons, with respect to the skill obtained using a constant mixed layer depth of 60 m. The incorporation in the model of the cooling in the mixed layer by turbulent entrainment of colder water from the thermocline, does not improve in an appreciable way the average skill of the predictions.

Numerical experiments on ice age climates

Numerical experiments with a hemispheric thermodynamic model are carried out, using present and ice age conditions. The computed surface temperatures for 18,000 years ago are in good agreement with the CLIMAP values. It is shown that the snow-ice cap that existed in the summer 18,000 years ago created a feedback mechanism which was responsible for perpetuating these ice age conditions. The increase of insolation due to orbital variations was responsible, at least in part, for the shrinkage of the snow-ice cap from 18,000 to 8,000 years ago. It is shown that the effect on the earth-atmosphere system of the changes in insolation due to the orbital variations depends on the preexisting snow-ice cap. It is significant for the ice cap that existed 18,000 years ago and insignificant for present conditions. The numerical experiments suggest that the evolution of climate depends in an important way on the initial snow-ice conditions. Therefore, according to the model the problem of simulating the evolution of climate is not determined if one does not prescribe these snow-ice conditions.

Monthly and seasonal prediction of sea surface temperature anomalies in the Gulf of Mexico

Abstract A thermodynamic model is used to predict the sea surface temperature (SST) anomalies in the Gulf of Mexico for extended periods as long as 3 months and for seasonal prediction. The basic equation of the model is the thermodynamic energy equation applied to the upper mixed layer of the ocean, which includes the horizontal transport of heat by mean ocean currents and by eddy turbulence, as well as heating by short and long-wave radiation, evaporation and sensible heat given off to the atmosphere. An objective verification of the skill of the predictions is presented for the period from March 1986 to February 1987. As an initial condition for the first month of the prediction we used the observed SST anomalies in the previous month, and for the second and third month we used the predicted monthly value in the previous corresponding month. Regarding the atmospheric interaction, the results show that the initial atmospheric forcing, which consists of the observed anomalies of surface air temperature and the surface wind in the month previous to the first month of the prediction, plays an important role in the prediction. The skill of the model is increased in the semi-prediction where we use, as atmospheric forcing, the surface air temperature anomalies and the surface wind anomalies for the current month instead of those belonging at the previous month to the first month of prediction. This result suggests that a coupled model in which were predicted simultaneously the ocean temperature and the atmospheric variables would improve the predictions of the SST anomalies.

Satellite observations needed for monthly and seasonal climate prediction

Abstract Monthly climate prediction for the Northern Hemisphere, using a thermodynamic model has shown an encouraging skill which could be improved by improving the data used. The satellite observations of snow and ice, and of sea surface temperature, are being used as input for the model predictions and to verify the results. The outgoing long-wave radiation, absorbed solar radiation, planetary albedo and cloudiness are used to calibrate the models, verify the predictions and simulations, and improve the parameterizations of the heating functions. A detailed description of the data needed and the accuracy requirements is given.