Abstract
Improvements in drilling and completion technology have resulted in significant increases in production rates from new horizontal wells in the United States. Observations over many years show that a well’s initial rate often has a predictable relationship to decline trends indicating a rate related bias in decline trends. This paper studies the relationship between initial flow rates and related decline trends and reserves forecasts for many of the major horizontal development plays in the United States and confirms that there is a rate related bias. Early decline trend forecast methods considering rate bias lead to improved reserve estimates and fewer revisions to estimates as wells mature. The study does not provide a methodology for determining peak rates but focuses on bias in decline trends related to known peak rates.
For years, reviews of oil and gas reserves estimates have shown that downward revisions on high rate wells (Lee 2017, Lee 2019) have been more common than upward revisions indicating a rate related bias. Various authors have discussed methods to correct various biases (SPEE 2010, 2016) (Freeborn 2012, 2013) which tend to result in overestimating production forecasts. Our experience is that forecast bias to the high side is more significant for high rate wells than for low rate wells. This rate Figure 1. Twelve areas for which peak month rate to decline trends relationships are presented. dependency raises the question of whether a rate dependent bias can be documented and corrected. This paper focuses on one factor: the relationship between peak month rates and decline trends or rate dependent decline trend bias. Results are presented for twelve areas noted in Figure 1 with an attempt to minimize rate dependent estimation bias.
High rate bias was documented with production data from a group of similar wells which was sorted from high peak rate to low peak rate and binned for analysis. This ranking minimizes time sequence bias which results when the best wells are drilled first with poorer wells later or if the better wells are drilled toward the end of a study period. The peak rate ranked wells are divided into approximately equal bins with each bin analyzed to determine the expected decline trends for the bin. The decline trends of the bins are then compared to the bin’s peak month rate to document the relationship between decline trends and peak rates.
