Topic:"Forecasting Methodology" in M12786

, Customer 2,112,088 2,094,094 while 2025 brought upwards of 40 mm of heavy rain that Hours of tapered off before temperatures fell below freezing. Interruption • During the December 19–20, 2025 storm, crews worked Customers 98.93 99.88 Re...

AI summary The 2025 storm caused significant outages due to heavy rain and freezing temperatures, challenging NS Power's restoration efforts. Crews faced obstacles like broken poles and poor visibility, leading to slower response times. Lessons include improved prediction models and reliance on ground patrols during adverse weather.

2025 Storm Restoration Cost Report (NSEB M12786) NSPI Responses to Consumer Advocate Information Requests NON-CONFIDENTIAL 1 Request IR-3: 2 3 Reference: Figure 10 4 5 The actual customer impact of the Level 3 event of December 19/20 was m...

AI summary NSPI attributes the overestimation of customer impact during the December 19/20 Level 3 storm to unanticipated wind speeds exceeding forecasts (e.g., 118 km/h vs. 100 km/h in Halifax). The response cites forecasting methodology limitations and references continuous improvement efforts in post-event reviews, though it states general prediction approaches remain unchanged across storm levels.

2025 Storm Restoration Cost Report (NSEB M12786) NSPI Responses to IG Information Requests NON-CONFIDENTIAL 1 undertaken to assess forecasting accuracy and whether any corrective action(s) were 2 identified or implemented as a result. 3 4...

AI summary NSPI's response to an information request regarding the 2025 storm restoration includes a comparison of predicted vs. actual customer impact and restoration durations, highlighting significant discrepancies. The report notes the activation of Level 3 EOC based on predicted restoration times and contractor mobilization efforts.

uty Officer, and reviewed by the Storm Leads to generate any required 29 action items. The Post-Event Critique summary document is then stored on an internal 30 ESRP SharePoint site. Date Filed: May 21, 2026 NSPI (IG) IR-5 Page 3 of 4 2025...

AI summary NSPI reviewed forecast accuracy post-storm, finding deviations within expected margins. No corrective actions were identified, and NSPI noted it evaluates forecast accuracy on a case-by-case basis. The response relates to NSEB M12786 and references IG IR-09.

1 Request IR-9: 2 3 Preamble: Regarding the December 5, 2025 event, NSPI acknowledges the over-forecast, 4 but as no Storm Report was filed, ratepayers have no insight into lessons learned from this 5 event. 6 7 (a) Did NSPI undertake an i...

AI summary NSPI acknowledges an over-forecast in the December 5, 2025 event but notes no Storm Report was filed. NSPI conducted a Post-Event Critique (PEC) identifying lessons on EOC rosters and communication. They emphasize reliance on third-party forecasts for pre-storm planning accuracy.

scale of customer impact, is dependant on the accuracy of the forecasts received by NS 30 Power from third party forecasters, including Environment Canada. However, NS Power’s Date Re-Filed: May 25, 2026 NSPI (IG) IR-9 Page 1 of 2 2025 Sto...

AI summary NS Power relies on third-party forecasts, including Environment Canada, to plan for storm impacts, acknowledging forecast limitations. It uses updated forecasts and calibrates its Damage Prediction Model with historical data, including the December 5, 2025 event, to refine future parameters.

with DA team o Tighten up per diem process – some employees not in system Customer & ETR Accuracy Reporting: Done Well: o ETR strategy development has enhanced and simplified at the same time. Do Better: o N/A RPP - Modeling/Forecasting: D...

AI summary The report highlights improvements in ETR strategy and team coordination during storm restoration, but notes challenges with damage prediction accuracy for smaller storms and limited access to advanced storm tools. Recommendations include updating models, enhancing training, and improving roster management for RPP.

-5- 1 (i) EOC activation; 2 (ii) Contractor mobilization; and 3 (iii) Staging of external resources. 4 (c) Please describe the process undertaken by NSPI through its Post-Event 5 Critique process, after each level 3 or 4 storm (whether ant...

AI summary The text includes questions about NSPI's Post-Event Critique process for storms, forecasting accuracy analysis, and clarification on its storm classification metrics (customer impact and restoration timelines). It seeks details on documentation, process variations based on storm realization, and evaluation of forecasting biases.

Date Filed: April 30, 2026 CA (NSPI) Page 1 of 3 1 Request IR-1: 2 3 Reference: Figure 9 – Comparison of 2025 Actual Operating Storm Costs to Base Rates by Level 4 5 Question: 6 7 a) Please confirm the 2025 underspending on level 3 and lev...

AI summary The document contains three regulatory information requests (IR-1, IR-2, IR-3) related to storm cost recovery, contractor performance management, and discrepancies in customer impact predictions. NSP is asked to clarify underspending on storm costs, provide mid-storm reporting examples, and explain forecast inaccuracies for a Level 3 event.

3. Forecasting Accuracy The two Level 3 events in 2025 present a contrast in forecasting outcomes. For December 5, NSPI predicted 28,308 customers affected; actual impact was 7,222, i.e., an over-prediction of approximately 74%. For Decemb...

AI summary This section discusses forecasting accuracy in the context of two Level 3 events in 2025, where NSPI significantly over- and under-predicted customer impact and restoration times. NSPI admits it does not conduct cross-event analysis of forecast accuracy and plans to update its Damage Prediction Model (DPM) with 2025 data in Q3 2026. The Industrial Group (IG) argues that the absence of cross-event analysis limits the ability to assess the DPM's reliability and the effectiveness of the annual calibration process.

NON-CONFIDENTIAL 1 2.0 REPLY TO CONSUMER ADVOCATE (CA) SUBMISSIONS 19 including Post-Event Critiques following Level 3 and 4 events, including instances when impacts 20 are ultimately smaller than forecast, to identify opportunities for op...

AI summary NS Power argues that its current processes and reporting frameworks, such as the Storm Restoration Cost Report, are sufficient for continuous improvement and do not require additional reporting mechanisms, even in the context of Post-Event Critiques and forecasting improvements.