As European electricity markets move towards greater unification, one important project is the Manually Activated Reserves Initiative (MARI), which implements cross-border market mechanisms for activating manual Frequency Restoration Reserves (mFRRs). In the Nordics, automation of the mFRR Energy Activation Market (mFRR EAM) is a prerequisite for later connecting to MARI. In the implementation, the mFRR bid selection is based on an imbalance forecast which aims to predict the net system imbalance, excluding any active regulations. Both the net system imbalance and the forecast accuracy are of great importance as the selected bids ultimately become a cost distributed among market participants as a part of the balancing market settlement.

As the share of weather-dependent renewable energy sources in Sweden increases, understanding how varying meteorological conditions affect power grid stability and related imbalances becomes increasingly important. This study investigates the relationship between synoptic weather patterns and both system imbalance and imbalance forecast error in Sweden, using data from the implementation phase (before go-live) of automating mFRR EAM. The weather data used for the study are forecasts from the MetCoOp Ensemble Prediction System (MEPS). Net system imbalance and imbalance forecast data are Svenska kraftnät internal data sources.

The imbalance forecast error is based on an imbalance forecast with a 30-minute horizon, targeting the activation lead time of mFRR balancing reserves. Errors are aggregated to an hourly resolution to match the resolution of the weather data. Absolute values of the forecast errors are computed on 15-minute and 60-minute resolutions to align with historic and evolving market time units for calculating the imbalance settlement price. Circulation types are derived per bidding area using the automated Jenkinson-Collison classification scheme, which produces 11 modes consisting of eight cardinal and intercardinal wind directions along with Cyclonic, Anti-Cyclonic, and Low Flow.

The statistical analysis is primarily based on Analysis of Variance (ANOVA) to test the overall and pairwise significance of circulation types in relation to the error metrics. To assess whether circulation types primarily act as proxies for underlying weather parameters, a control study tests the statistical significance of circulation types while sequentially adding individual weather parameters to the model as controls.

The study reveals some association between circulation types and signed forecast errors, with the relationship becoming notably stronger and more statistically significant when the absolute error is considered. In particular, large absolute forecast errors appear more common during Westerly flows compared to Easterly, Cyclonic, and Anti-cyclonic conditions. In the control study, while the model's explanatory power increases with the addition of weather parameters, the influence of circulation types often persists, suggesting they encapsulate broader synoptic conditions not fully represented by individual meteorological variables.