Generalized extreme value statistics, physical scaling and forecasts of gas production in the Haynesville shale

Abstract

We present a hybrid, data-driven and physics-based method of forecasting play-wide gas production in the Haynesville Shale, which is currently the second-largest shale gas producer in the US. We first define several statistical well cohorts, one for each reservoir quality and each well completion technology in the Haynesville. For each cohort, we use the Generalized Extreme Value (GEV) statistics to obtain the historical average well prototypes. The cumulative production of each well prototype is matched with a physics-based scaling curve, and its production is then extrapolated for up to two more decades. The resulting well prototypes are exceptionally robust. If we replace individual production rates from all existing wells with their corresponding well prototypes and sum them up, the total rate will match remarkably the past gas field rate, and – in this case – we obtain a base or do nothing forecast. Next, we calculate the number of potential infill wells per square mile and schedule future drilling programs to obtain plausible production forecasts in the Haynesville. Because Haynesville is the most active shale play in North America in terms of refracturing, we also propose a novel approach to identify refracturing candidates among the old Haynesville wells and deliver another forecast scenario of future refracs. We predict that Haynesville will ultimately produce 30 Tscf of natural gas from the 4684 existing wells. Most likely, by drilling 923 new wells in the core (sweet spot) areas by 2023, EUR will increase to 40 Tscf. Additional 5023 wells in the noncore areas are forecasted to be drilled by 2032, increasing EUR to 90 Tscf. We also show that refracturing old Haynesville wells is more cost-effective than drilling new wells in the poor quality reservoir, especially in the era of low oil and gas prices.