The Winter and the Summer P50 Peak Load Forecast Methodologies in BPM 103 fail to adequately capture discrete load additions or subtractions of loads in the medium term (as those changes gradually become part of the five years of historical load data used to calculate the monthly P50 Peak Load Forecasts). These discrete load changes could be additions and subtractions, in scenarios where load is existing or new, meaning historical load data exists or does not, respectively. Clear direction and a robust policy for discrete load changes is important for all Participants experiencing discrete load changes, but is of particular consequence when one Participant transfers the responsibility for its load and resources in its entirety to another Participant, as is anticipated by Snohomish PUD’s transition in BPA service contracts, resulting in BPA representation of Snohomish’s load in October 2025. 

• Load Changes when Historical Load Data is Available. As currently described in BPM 103 the P50 Peak Load Forecast methodologies for both Summer and Winter begin by determining the peak load for each month of the Season for the last available five Seasons using the Historical Load Data submitted as part of the Advanced Assessment. Note that Historical Load Data is defined in BPM 101 Advanced Assessment as: ”Load data from one or more Years prior to the current Year, such as the previous 10 Years. Historical Load Data is expected to consist of 8,760 hours (or 8784 hours for a leap Year) of data for a Year”. Tariff Section 16.1.1 allows for the P50 Peak Load Forecast in the Business Practice Manuals to include “a base monthly peak derived from a recent historic period that recognizes additions and removals of load during the historic period”. In the scenario where a load change (addition or subtraction) has historical load data available (as would be the case with a Participant-to-Participant load transfer, or a Participant assuming responsibility for a load that has been operational prior to the Participant assuming said responsibility), the affected Participant(s) is(are) responsible for adjusting the peak load for each Month of the Season for any of the last available five Seasons that do not capture the load change, and then calculating the monthly P50 Peak Load Forecasts, until the load change is automatically captured fully.
  • In the case where one Participant (A) completely assumes the load responsibilities for another Participant (B) in October 2025, Participant A will complete the Forward Showing for Winter 2025/26 on March 31, 2025, using load data that combines Participant A’s and B’s historical loads. Participant B will not complete a Forward Showing for Winter 2025/26 (but will participate in Summer 2025 as normal). 
  • In the case where a Participant assumes responsibility for a discrete load (e.g. a paper mill) that has operated previously, the Participant will add this historical load data from the discrete load to their historical load data for every hour. 
• Load Changes when Historical Load Data is Unavailable. In the scenario where a load change does not have historical load data available (e.g. a new large load such as a data center or manufacturing facility) the affected Participant will generate synthetic load data and adjust monthly P50 Peak Loads for any of the last available five Seasons that do not capture the load change, and then calculate the monthly P50 Peak Load Forecasts, until the load change is automatically captured fully in historical load data.