Sites over threshold
This analysis gives the percentage of surveyed sites that were over a total emission threshold. This document provides the details of this analysis. For information on how to run it see the programs documentation.
Input data
The input data is derived from the Bridger KMZ files and equipment reports. We create a table the has a row for every observation of equipment. If an equipment item is surveyed multiple times, even on the same day, then we have an entry for each of these measurements including non-detections. This table also includes a row for each site where no equipment or emissions were recorded, but that was covered in the survey.
This is more complete than the facility or equipment summaries as we have a finer grained level of detail on each detection and non-detection. However, for this analysis most of this detailed is aggregated into site level statistics.
Method
For each site we compute a total emission rate on the day of the observations. If a site was sampled more than once in one day, then these values are combined into a single value. When an item of equipment has multiple observations in one day, this aggregation is done by either taking the maximum of these or the mean for each item. The mean is of only non-zero values.
For example, if a compressor was emitting at 10 kg/hr and 20 kg/hr on two separate occasions, then the total emission rate for that day would be 20 kg/hr by using the maximum method. If there was also a flare that was only seen emitting once at the same site, at a rate of 30 kg/hr, then the total emission rate for this site would be 50 kg/hr using the maximum method. The mean method would give us 15 kg/hr for the compressor and 30 kg/hr for the flare (since we ignore zeros when taking the mean), giving a total of 45 kg/hr.
The threshold is then applied to this total emission rate. If the total emission rate is over the threshold then the site is considered to be over the threshold.
Finally we compute the count of the sites over the threshold divided by the number of sites surveyed.
If a site is visited on multiple days during the survey period then it can only contribute to the number of sites over the threshold once. We do not double count, but we also do not double in the denominator. This creates a bias towards larger values for the percentage of sites over the threshold - this is by design to create a more conservative estimate.
Comparison with Bridger Facility Summary
The Facility Summary tab in the Bridger reports gives us a way to compare our results. First we must understand how the values in this table are constructed. Let us consider the column Total Facility Emission Alternate (kg/h). This is the sum of the mean emissions for each item of equipment across the entire survey. In our analysis we aggregate by day, and consider site totals for that day. For this value the mean can be across multiple days, which can lead to a lower value. Non-detections are not included in this mean, like in our approach described above.
By comparing with the number of sites with a value of Total Facility Emission Alternate (kg/h) over a given threshold we arrive at a slightly smaller number. For CVX Phase IV 2024 we have 15 and 14 using maximum and mean methods described above, respectively, for sites over 100 kg/h. The Facility Summary gives 13 sites over 100 kg/hr.