Showing papers by "Eugenia Kalnay published in 2000"

Dynamical Seasonal Prediction

Abstract: Dynamical Seasonal Prediction (DSP) is an informally coordinated multi-institution research project to investigate the predictability of seasonal mean atmospheric circulation and rainfall. The basic idea is to test the feasibility of extending the technology of routine numerical weather prediction beyond the inherent limit of deterministic predictability of weather to produce numerical climate predictions using state-of-the-art global atmospheric models. Atmospheric general circulation models (AGCMs) either forced by predicted sea surface temperature (SST) or as part of a coupled forecast system have shown in the past that certain regions of the extratropics, in particular, the Pacific-North America (PNA) region during Northern Hemisphere winter, can be predicted with significant skill especially during years of large tropical SST anomalies. However, there is still a great deal of uncertainty about how much the details of various AGCMs impact conclusions about extratropical seasonal prediction an...

Role of sea surface temperature and soil-moisture feedback in the 1998 Oklahoma–Texas drought

Abstract: The drought that affected the US states of Oklahoma and Texas in the summer of 1998 was strong and persistent, with soil moisture reaching levels comparable to those of the 1930s ‘dust bowl’1,2. Although other effects of the record-strength 1997–98 El Nino were successfully predicted over much of the United States, the Oklahoma–Texas drought was not3. Whereas the response of the tropical atmosphere to strong anomalies in sea surface temperature is quite predictable, the response of the extratropical atmosphere is more variable4,5. Here we present results from mechanistic experiments to clarify the origin and maintenance of this extratropical climate extreme. In addition to global atmospheric models6,7,8,9,10,11, we use a regional model12,13 to isolate regional climate feedbacks. We conclude that during April and May 1998, sea surface temperature anomalies combined with a favourable atmospheric circulation to establish the drought. In June–August, the regional positive feedback associated with lower evaporation and precipitation contributed substantially to the maintenance of the drought. The drought ended in the autumn, when stronger large-scale weather systems were able to penetrate the region and overwhelm the soil-moisture feedback. Our results show the potential for numerical models including appropriate physical processes to make skilful predictions of regional climate.

A New Method of Observed Rainfall Assimilation in Forecast Models

Abstract: A method to assimilate observed rain rates in the Tropics for improving initial fields in forecast models is proposed. It consists of a 6-h integration of a numerical forecast model; the specific humidity at every time step at each grid point is modified (nudged) in such a way that the total model precipitation accumulated during this integration becomes very close to that observed. An increase in the model precipitation is achieved by moistening the lower troposphere above a grid point with prescribed supersaturation; a decrease in the model rainfall is brought about by decreasing the specific humidity in the lower troposphere in proportion to the difference between the model and reference specific humidity profiles. The modified values depend on the difference between the model and target precipitation. The depth of the atmospheric column in which the humidity is changed is proportional to the target rain rate. Quality criteria of a rain assimilation procedure are proposed. The quality of the a...

Application of the Quasi-Inverse Method to Data Assimilation

Abstract: Four-dimensional variational data assimilation (4D-Var) seeks to find an optimal initial field that minimizes a cost function defined as the squared distance between model solutions and observations within an assimilation window For a perfect linear model, Lorenc showed that the 4D-Var forecast at the end of the window coincides with a Kalman filter analysis if two conditions are fulfilled: (a) addition to the cost function of a term that measures the distance to the background at the beginning of the assimilation window, and (b) use of the Kalman filter background error covariance in this term The standard 4D-Var requires minimization algorithms along with adjoint models to compute gradient information needed for the minimization In this study, an alternative method is suggested based on the use of the quasi-inverse model that, for certain applications, may help accelerate the solution of problems close to 4D-Var The quasi-inverse approach for the forecast sensitivity problem is introduced, and then a closely related variational assimilation problem using the quasi-inverse model is formulated (ie, the model is integrated backward but changing the sign of the dissipation terms) It is shown that if the cost function has no background term, and has a complete set of observations (as assumed in many classical 4D-Var papers), the new method solves the 4D-Var-minimization problem efficiently, and is in fact equivalent to the Newton algorithm but without having to compute a Hessian If the background term is included but computed at the end of the interval, allowing the use of observations that are not complete, the minimization can still be carried out very efficiently In this case, however, the method is much closer to a 3D-Var formulation in which the analysis is attained through a model integration For this reason, the method is called ‘‘inverse 3D-Var’’ (I3D-Var) The I3D-Var method was applied to simple models (viscous Burgers’ equation and Lorenz model), and it was found that when the background term is ignored and complete fields of noisy observations are available at multiple times, the inverse 3D-Var method minimizes the same cost function as 4D-Var but converges much faster Tests with the Advanced Regional Prediction System (ARPS) indicate that I3D-Var is about twice as fast as the adjoint Newton method and many times faster than the quasi-Newton LBFGS algorithm, which uses the adjoint model Potential problems (including the growth of random errors during the integration back in time) and possible applications to preconditioning, and to problems such as storm-scale data assimilation and reanalysis are also discussed