We have noticed problematic performance of the Haynesville LA Short Term Forecast (STF) due to very strange results from the LA state data, and have struggled to make sense of the issues. We have finally resolved those, and are publishing out our revised model.
Starting in this past summer, we noticed very strange results with our STF and its Initial Production Rates (IPRs). Our IPRs are modeled from state well data, and starting in January, and accelerating through the summer, we saw vastly increasing IPRs, leading to a vastly increased STF, when the daily production model was showing none of that dynamism. Greatly complicating this issue is the fact that Louisiana reports gas production at the lease level, and only later do they publish information allowing allocation to specific wells. The chart below shows the result today if we had chosen to update to the information we thought we were getting. Note how high the "TEST" forecast would have gone, and the dip into the end of last year. (Red is the latest report state data, black daily production, green current STF and purple is the STF if we used the problematic data)
With problematic inputs, and other data indicating that trajectory was unrealistic, we chose at the time to stop updating our IPR model and existing history inputs at a much earlier date, leading the current relatively flat curve.
However, the recent Haynesville LA trajectory has again caused us to reevaluate our data. This time we had enough information to resolve the issues we were seeing from the lease data, and they all stem from the Expand merger. The graph below indicates the problem. The red line is a Chesapeake lease. The blue is different lease, with a different number, allocated to Expand. (Remember we still don't have lease to well information).
In hindsight, it is obvious that these are the same lease. Before this discovery, this would result in double counting both for the totals and the effects on the STF - the "old" lease would continue into the future, declining at and the "new" one would be added on top - when that is clearly wrong (it also would have "double counted" the IPR effects in the model).
So, by resolving this, you get the following new STF for Haynesville-LA. (This should look familiar. Purple represents approximately what the new STF will look like tonight. We say approximately as this chart is based on completions from 3 days ago, and tonight's will be slightly updated).
One of the most significant factors to us is that regardless of our choice of "last good date" for total production - i.e. the date from which we "decline" old wells and only use our completion data for new wells - that period from Aug to Nov '24 remained consistently near the LA reported production, giving us far more confidence in the overall trajectory.
However, we do note that the Haynesville - LA daily production is the same or higher than the STF recently. Theoretically, the STF should represent an average upper bound - monthly production shouldn't exceed it by much. This is where we must point to an issue in the daily production model. As mentioned in our update to the Haynesville LA daily production model, we are now counting all LEG and NG3 volumes as incremental to Haynesville LA, but we also know that some of that is coming from the TX side of the border. Until we have enough history to calibrate properly, or other information to indicate a better split, we can't correctly evaluate which side of the border this is coming from.
For this reason, we also look at the total Haynesville, and here we see the relationship is as we expect.
The net effect is to increase production estimates for the next three months in the STF by about 1.3 bcf/d on average.
As usual, reach out to support@synmax.com with any questions.