The ''t-model'' for dimensional reduction is applied to the estimation of the rate of decay of solutions of the Burgers equation and of the Euler equations in two and three space dimensions. The model was first derived in a statistical mechanics context, but here we analyze it purely as a numerical tool and prove its convergence. In the Burgers case, the model captures the rate of decay exactly, as was previously shown. For the Euler equations in two space dimensions, the model preserves energy as it should. In three dimensions, we find a power law decay in time and observe a temporal intermittency.dimensional reduction ͉ finite time singularity T he advent of powerful computers has enhanced our ability to study complex systems and understand their dynamics, yet there are many problems that are too complex for computer solution. The task facing the numerical analyst working on such problems is to reduce their complexity to something manageable, yet preserve, maybe only in a statistical sense, their salient features. Turbulence and, more generally, flow at a high Reynolds number provide a prime example of a problem where direct solution is impractical and some simplification is needed. Although a lot of knowledge about this problem has accumulated over the years (1-7), a suitable effective model has not yet emerged.In earlier work (8, 9), we and others have derived methods for reducing the number of variables one needs to solve for in complex systems, using a statistical projection formalims borrowed from statistical mechanics. A special case, the ''t-model'' based on a ''long memory'' assumption (i.e., the assumption that the autocorrelations of the noise decay slowly), seemed to be promising for problems in fluid mechanics (10,11). An earlier application of this model to the Burgers equation (12) yielded remarkably accurate estimates of the rate of decay of solutions.In the present paper, we use the t-model to reduce the number of variables in spectral approximations of the Euler equations and calculate the rate of decay of solutions. We expect the methodology to be of broader interest yet, and to be a step toward the development of reduction methods of general applicability. We first prove the convergence of the model as the number of Fourier components increases. We do not address here the claim implicit in previous work that the t-model may yield acceptable results even when the number of variables remains finite. The results we obtain are, however, surprisingly accurate, and a full analysis may well have to go through some version of the arguments on the basis of which the t-model was originally derived. Note that unlike previous damping methods for allowing spectral calculations to proceed to significant time spans (e.g., refs. 13-19), the t-model equations contain no adjustable parameters and is guaranteed to remain stable.The paper is organized as follows. First we present some properties of systems of equations that have constant energy. We then present the derivation of the t-model. We prove some resul...