Following a periodic review of model performance against new state data we have implemented a number of improvements to the Short Term Forecast model for Permian-NM. This update has resulted in a -0.46bcf/d change to the current month's forecast. Details regarding what we changed and why are described below. As always please feel free to reach out with any questions.
What Changed?
-
We made workflow improvements in how candidate wells are selected for use in decline curve construction.
-
We added additional, more recent wells to our training and historical datasets.
-
We improved our decline curve shaping method to better match observed history.
Why it Changed?
With state data now available through February 2024 in New Mexico we observed a growing divergence between our old model’s output and new state production data. The divergence appeared to be entirely coming from our decline curve model for the region, which was underestimating the declines of new wells. We successfully identified and improved our decline curve calculation methods and have back tested both models, old and new, against actualized state production. The results, in the chart below, show the improved performance of our new model when run forward only on data available 2/1/23.
What about other regions?
We are constantly evaluating our models against the latest state data and have not identified any other regions in need of updates.
