An assessment of The Long-Term Energy Analysis Program used for the EIA 1978 report to Congress

The Long-Term Energy Analysis Program (LEAP), a computer model of the energy portion of the U.S. economy that was used by the Energy Information Administration (EIA) of the U.S. Department of Energy (DOE) for the 1995–2020 projections in its 1978 Annual Report to Congress, is evaluated here. Essential considerations include 1.(1) the adequacy of its documentation,2.(2) the evaluator's experience in operating the model,3.(3) the adequacy of the numerical techniques used,4.(4) the soundness of the economic and technical foundations of the model equations,5.(5) the degree to which the computer program has been verified. To show which parameters strongly influence the results and to approach the question of whether the model can project important results with sufficient accuracy to support qualitative conclusions, the numerical sensitivities of some important results to model input parameters were obtained using a novel adjoint technique. This technique enabled efficient determination of first derivatives of results of interest with respect to all input data (~ 1600 non-zero sensitivities). Direct recalculation of the model with altered data sets was used to confirm the results and to assess the extent of non-linear effects. The input data are categorized and discussed, and uncertainties are given for some parameters as examples. From this background, an overall evaluation is given of the model's suitability for use by the EIA.