With a growing global demand of clean renewable energy, offshore wind activities will extend to more harsh environments, including sub-arctic areas like the Baltic Sea. Sea ice can occur here which needs to be taken into consideration in the design of substructures, i.e. jacket substructures, of offshore wind turbines. Several ice mitigating measures exist for jacket substructures, of which one is disregarding braces crossing the waterline. In this way, ice cannot induce loads to the relatively slender braces in the jacket. However, disregarding these braces has disadvantages in terms of structural integrity in comparison to a jacket including these braces. In Part I of this thesis, two types of jackets, one without (Type 1) and one with (Type 2) waterline crossing braces, are implemented in a numerical model and subjected to ice loading in order to conclude whether the use of a Type 1 jacket design can be justified or whether a Type 2 jacket design is also suitable in sub-arctic areas. Since over the past years the modelling techniques regarding ice-structure interaction have been updated and improved significantly, assessing the loads at the jacket braces can be done more accurately, allowing to give a more thorough conclusion on whether to use braces crossing the waterline on jackets in sub-arctic areas. As a result, a design including these braces could become feasible, whereas in the past it would be disregarded. Two types of ice failure are considered in this research, being ice bending failure and ice crushing failure. First, a model is developed to quantify the ice actions occurring at a jacket substructure as function of the approach angle of the ice direction relative to the structure and as function of an introduced 'threshold angle'. It was found that ice failing in bending is equally significantly present as ice failing in crushing. Subsequently, numerical models are introduced that describe the failure behaviour of bending and crushing ice and the force that both ice failure types will induce to the structure. The structure is represented by a 5 [MW] offshore reference turbine supported by a jacket substructure. Using these models, dynamic simulations are performed in order to investigate the local susceptibility to ice induced vibrations (IIV), as well as assessing the Ultimate Limit State (ULS) load at the braces. As a result, local brace IIV regimes could be recognised, however sustained frequency lock-in could not be observed. Furthermore, it was found that a Type 2 jacket could be suitable to use in sub-arctic areas considering the ULS load effects due to ice loading, but care should be taken for the ULS brace load, as that is of great importance for the design. In Part II of this thesis, it is assessed whether coupled dynamic ice-structure interaction can be modelled using a dynamic substructuring modelling technique, also known as Craig-Bampton method, since it is industry practice to use such a method when jackets are involved. This method is based on a model order reduction of the full finite element model representing the structure, resulting in improved computational time and allowing assembly of multiple (reduced) substructures. The method is successfully implemented allowing to solve coupled ice-structure interaction problems and a framework to do so is provided. It is shown that excluding the loaded degrees of freedom (DoFs) in the jacket from the reduction leads to more accurate results and is considered convenient since few DoFs are loaded by ice, maintaining the effectiveness of the reduction method. As a result, IIV can be investigated using this reduction method, however challenges related to the accumulation of numerical accuracy errors arise with truncation of more higher frequency modes and require further development to be resolved.