Development pressure on biodiversity reserve networks in densely populated countries may lead to the decision to compensate for biotope loss by improving connectivity. Such a decision makes sense, if creation of new biotopes takes too long and if improving population exchange is a conservation target of the reserve network. To explore the impact of such decisions, we analyse four compensation scenarios. The scenarios vary in how strong loss and compensation is locally fixed. The reserve network was modelled as a graph where biotope patches are represented by nodes and connectivity corresponds to edges along which animals migrate from patch to patch. Connectivity improvement was modelled as a reduction of edge lengths. Ecological equivalence is measured by metapopulation capacity as defined by Hanski and Ovaskainen (2000). Localised modifications were analysed with eigenanalysis. Modifications spread over the whole component were analyzed with a linear regression model which uses the total biotope area and the length of the minimal spanning tree as input. Our results show that both general connectivity improvement and clearly localised connectivity improvement can be efficient compensation measures for area loss. Local measures best focus on connectivity improvement between the largest patches. For Switzerland's dry grassland reserve network, we found that in general, for half of the patches it is possible to compensate an area loss of 100 m² by a connectivity improvement equivalent to an edge length reduction of less then 3 m. Our results show that connectivity improvement is a valuable compensation alternative to creation of new patches.