This page has been updated with official data until January 2017. For 2014 wells, decline data for 24 months, starting with the peak production month, comprising 2059 (confidential and non confidential) horizontal wells, is now shown. For the 2015 wells, the shape of the decline curve for the first 12 months is mentioned . For the long term decline rate, the third year y/y decline rate is now based on the partial data of 2014 wells instead of data from older wells.
The decline is measured from the IP30 month (month 1 in the table).
The numbers are for oil, not for boe (oil and gas) and are based on monthly “runs” (volume sales) for each well. On the aggregation level used here, the difference between runs and production data is below 1%. The runs have been chosen because for confidential wells, the publication of runs data is much more timely than the publication of the production numbers.
|% of IP30 (peak Month) Production -oil – 2014 wells
| Year 1
The second year production amounts to 43.2% of first year production. The ramp up phase of a well, preceding the IP30 month, is quite irregular and can last one month, but also more than 6 months. The cumulated average pre IP30 production amounts to 70% of that IP30 month production.
The observed decline rate is steeper than the one shown for pre 2014 wells in D. Hughes’ Drilling Deeper. The Hughes data show 29.1% in month 12 and 19.1% in month 24.
In 2014, tight spacing of wells had become the rule (6 to 8 wells per space unit). Interference could explain part of the difference with the curves based on the older wells shown by Hughes. More intensive fracking, boosting peak production, is most likely also a factor. For the pre 2014 production wells, total first year production amounts to 6 times peak month production. For the 2014 wells, the multiple has fallen to 5.23. That is a significant difference.
Regarding the 2015 wells, the decline curve for the first 12 months, based on all 1614 wells, is very similar to that of the 2014 wells, but partial data for the 2nd year indicate a slightly steeper decline.
To see the possible impact of enhanced completions on decline rates, the performance of EOG wells have been analysed (see here).
Long term average well production decline curve
To be able to make a serious well cost calculation covering the expected well life, a 30 year decline curve has been generated. An extension to 50 years was also tested. The difference in financial results for the two variants was, as expected, insignificant. Some companies seem to assume a 50 year well life for their EUR calculations.
The decline rate for the first 3 years is based on the measured decline of 2014 wells (for the first 2 years, see above). The decline rate of the 3th year is based on still partial 2014 well data. It was previously based on the the y/y decline rate of pre 2014 Bakken wells, which was 20% lower. This has led to a revision of the curve data. EUR of the revised “type” curve is 7% lower than before.
For the year 4 and 5, the relative decline rates of pre 2014 wells are maintained.
For the years 6 to 30, the theoretical decline rate values of a study published by the USGS continue to be used as a basis, although the numbers in that study seem to be a mixed well and field decline approach. In the retained curve, y/y production decline progressively falls to 10% in year 10 and 5% in year 15. From year 15 to year 30, the y/y decline remains at 5%.
If the derived EUR values (oil) for 2016 wells is compared with EUR (oil) estimates of pre 2014 wells, based on lower IP30 values and less steep decline rates, the EUR increase of the 2016 wells seems to be very low.
The 30 year decline curve described here is used in the well cost model as the standard curve.
For different fields and producer specific calculations, the model assumes that the observed differences in decline rates (retained in respective calculations) are limited to the first two years of production. Starting with the 3th year, the y/y decline rates are supposed to be similar in all cases.
The rounded production values for the standard decline curve are shown for selected years in the following table:
|bpd in % IP30