Geomagnetic activity at Northern Hemisphere's mid-latitude ground stations: How much can be explained using TS05 model

For the 2007 to 2014 period, we use a statistical approach to evaluate the performance of Tsyganenko and Sitnov [2005] semi-empirical model (TS05) in estimating the magnetospheric transient signal observed at four Northern Hemisphere mid-latitude ground stations: Coimbra, Portugal; Panagyurishte, Bulgary; Novosibirsk, Russia and Boulder, USA. Using hourly mean data, we find that the TS05 performance is clearly better for the X (North-South) than for the Y (East-West) field components and for more geomagnetically active days as determined by local K-indices. In ∼50% (X) and ∼30% (Y) of the total number of geomagnetically active days, correlation values yield r≥0.7. During more quiet conditions, only ∼30% (X) and ∼15% (Y) of the number of analyzed days yield r≥0.7. We compute separate contributions from different magnetospheric currents to data time variability and to signal magnitude. During more active days, all tail, symmetric ring and partial ring currents contribute to the time variability of X while the partial ring and field aligned currents contribute most to the time variability of Y. The tail and symmetric ring currents are main contributors to the magnitude of X. In the best case estimations when r≥0.7, remaining differences between observations and TS05 predictions could be explained by global induction in the Earth's upper layers and crustal magnetization. The closing of field aligned currents through the Earth's center in the TS05 model seems to be mainly affecting the Y magnetospheric field predictions.