I apologize in advance if this has already been posted on GHS. I think it is important to note from whom the author is getting her info (NSAI).
Lessons Learned From The Haynesville Shale
The Haynesville is a very young play, but information from additional wellbores and production history is defining economic limits and improving costs and returns
Article By Jeannie Stell
Published Apr 17, 2009
The Haynesville shale in Texas and Louisiana is a very young play, but information from additional wellbores and production history is defining economic limits and improving costs and returns, according to a recent report by Houston-based SMH Capital and Netherland, Sewell & Associates Inc. (NSAI)
Lessons learned by the region’s gas producers are based varied initial production (IP) rates; the basic relationship between IPs and estimated ultimate recoveries (EUR); decline curve analysis; new information on the changes in lithology across the play; the importance of production history; and some interesting comparisons to the Barnett Shale.
Not only have IP rates varied across the play, but reporting methodology has been inconsistent across operators, reports the study. Unlike plays such as the Barnett shale and Cotton Valley sands, where EUR:IP has been 1:1, a surprising rule of thumb for the Haynesville is an EUR per 1 million cubic feet equivalent per day (MMcfed) of average 24-hour IP rate equal to 0.37 Bcfe, although the range has been from 0.30 Bcfe to 0.47 Bcfe. For example, a 15 MMcfed IP rate would predict a 5.5 Bcfe well.
The driver of this ratio is the over-pressured reservoir that benefits the relative economics of the play. However, this would also imply a steeper production decline than forecasted. Over time, the average EUR per MMcfe of IP rate might increase toward 0.4 Bcfe to 0.5 Bcfe based on a combination of more production data being obtained and an industrywide adoption of longer laterals and more frac stages.
Using $5.71 per Mcf and yearend costs, a well requires 4.7 Bcfe EUR to provide a 10% return. Based on NSAI’s average, with a conservative decline curve, that would imply an IP rate of around 13 MMcfed. Operational efficiencies, combined with better completion practices, additional production history and a reduction in costs, is likely to improve economics, resulting in lower needed IP rates.
Although NSAI did not define a sweet spot on its maps, the company did provide a detail of publicly available IP rates by area, which show better well areas in northern Louisiana—specifically in the Caddo, Bossier, Red River, and DeSoto parishes. Because the information is based on public data only, there might be other wells that have not been reported. Also, recent wells might have longer laterals with more frac stages than shown.
Two drivers of the play’s success are shale thickness and completion. According to the report, the average thickness is greater in Texas, although the quality of the rock appears to be a bit lower. The play shallows toward the northwest, and shallower means less reservoir pressure impacting IP rates. Drilling shallower wells save about $1 million in drilling and completion costs.
Also, the data show slightly lower porosity and permeabilities in Texas. Variations in rock quality across the play suggest that the productive zone in northern Louisiana tends to be a more siliceous shale, while there appears to be a slightly higher clay component in Texas, with parts of the lower shale being limey closer to the southwest. This could be particularly interesting as new wells are tested in the Haynesville’s Shelby County. However, more wells, core data, and petrophysical data are needed to properly assess all of the variability across the play.
Well spacing is still an open question in the Haynesville. Usually, only one well per 640-acres is drilled. Since the gas-in-place is close to the Barnett Shale (under development on 40-acre spacing), albeit at a much higher pressure, fewer wells should be needed to drain each section. Since Haynesville wells are twice the size of Barnett wells, the wells are likely draining a greater area. It should take an average of eight wells per section to achieve a recovery factor similar to the Barnett’s, although some better areas may only need five to six wells.
Due to the low permeability of the reservoir, the Haynesville’s recovery factor is not better than the Barnett’s. However, unlike the Barnett, which has numerous high-quality sand lenses, the Haynesville is consistent, permitting large and closely-spaced fracs for better reservoir penetration.
The two big decline curve issues in the Haynesville are the rate of change in the curve as pressure and volume decrease, and the ultimate terminal declines rates, which are still largely unknown. Given pressure and reservoir differences, it is likely that the Haynesville may present a variety of decline curves across the play.
After the first year of data collection in the Haynesville unveiled a tremendous over-pressured gas resource, the industry is still defining the economic limits of the play and optimizing per-well economics despite high initial production rates. Similar to older shale plays like the Barnett and the Fayetteville, it might take producers several years to define the sweet spots of the play, improve industry best practices and understand how to optimally develop the spacing and economics of the play.
The report concludes that more infrastructure, better reservoir knowledge, improvements in completion practices, drilling and completion cost savings, and production history will result in higher EURs and better returns over the next few years.