In the face of a warming Arctic, ice-rich permafrost landscapes are undergoing rapid changes. Ice-wedge polygonal networks in Arctic lowlands are especially vulnerable and melting ice wedges can induce widespread subsidence and trough formation. The transition from low-centered to high-centered polygons can have important implications on surface hydrology, as the connectivity of the newly forming trough networks determines the rate of drainage for these lowland landscapes. However, quantifying such dynamics can be challenging, as even small-scale changes can have far-reaching implications for the larger scale hydrology of a region. In this talk, we introduce an automated workflow that enables quantification of trough network dynamics in thaw-affected landscapes. We use methods from traditional computer vision to extract (a) the spatial pattern of the trough network and (b) the morphological parameters of trough width and depth from high-resolution digital terrain models. Finally, we (c) incorporate this information into graphs - a mathematical concept used to represent complex networks - and use graph analysis methods to determine progressing subsidence and trough formation. Based on a study area in the Anaktuvuk River Fire scar on the North Slope, Alaska, USA, we present the potentials and benefits of such graph algorithms for quantifying the erosional development of this thaw-affected landscape. In our study region, we observed an increase (+127%) in the number of discernible troughs as well as their connectivity (number of disconnected networks decreased by 89%) over the observed period of ten years. The average width of troughs has increased (+14.5%), while the depth has decreased (-12.5%). With this approach, for the first time, a large-scale analysis of such detailed ground-ice and hydrological surface dynamics is made possible.