With increasing visitor numbers an understanding of the impacts of human activities on Antarctic soil environments has become an important issue. The overall objective of this thesis was to investigate soil physical impacts, and soil recovery rates, following human disturbance in the Ross Sea region of Antarctica. Visually disturbed and nearby control sites were assessed using a combination of techniques, including field-based visual site assessments, comparative photo-records, a desert pavement recovery assessment method, soil sampling, and soil dry bulk density measurements. An experimental soil disturbance trial was set up near Scott Base and bacterial DNA profiling was used to investigate the response of communities to removal of the top 2 cm of soil. Infra-red track counters were installed on Ross Island walking tracks to record visitor use over a two-year period. Five case studies, from former research stations to field campsites, were investigated to assess the accuracy of the impacts predicted in environmental impact assessments and the observed impacts. In all cases there was a high level of consistency between predicted and observed impacts. It was apparent that the environmental impact assessment process raised environmental awareness of visitors; motivating them to avoid, remedy, or mitigate, their environmental impacts. A field-based method was developed to assess desert pavement recovery and tested on 54 sites. Eleven criteria were used: embeddedness of surface clasts; impressions of removed clasts; degree of clast surface weathering; % overturned clasts; salt on underside of clasts; development of salt coatings; armouring per m2; colour contrast; evidence of subsidence/melt out; accumulation of salt on cut surfaces; and evidence of patterned ground development. Recovery criteria were assigned a severity rating on a scale from zero to four, and the Mean Recovery Index (MRI) of the site was calculated relative to an equivalent control. Five recovery stages were defined as recently or highly disturbed (MRI of 0-24%), through to indistinguishable from control site (MRI = 100%). Fifty of the 54 sites investigated were in an intermediate or higher stage of desert pavement recovery (MRI > 50%), 30 sites were in an advanced stage of recovery (MRI > 75%), and four sites were indistinguishable from adjacent control sites (MRI = 100%). Active surfaces, such as gravel beach deposits, aeolian sand, and alluvial fan deposits, recovered relatively quickly and had higher MRIs, whereas less active sites of higher intensity disturbances, such as bulldozed tracks at Marble Point, had lower MRIs, with only intermediate recovery up to 50 years after disturbance. Following physical impacts such as foot and vehicle traffic the surface recovery recorded here was often greater than that predicted by previous researchers. At one-off campsites, footprints from dispersed trampling were undetectable within five years. At some sites walking tracks remained visible in the landscape 17 years after they were formed (due to surface recontouring and larger clasts concentrating along track margins). For steep slopes and sites where repeated visits occur, use of a single track is recommended. At sites where visually obvious impacts were remediated by replacing larger stones back in their original positions, ensuring that surface stones are placed with the weathered side up, and raking of gravel sand-sized displaced materials, visible evidence of former occupation was almost undetectable. Concentrating activity on young, active, and readily recoverable surfaces, or resilient bedrock, is recommended. There were no significant changes in bacterial community structure in response to experimental removal of the top 2 cm of soil over a 35 day sampling period. Differences in bacterial community structure between samples correlated with differences in soil electricity conductivity (R2 = 0.55) and soil pH (R2 = 0.67), reflecting small scale (< 2 m) soil heterogeneity. Infrared track counters recorded: 5084 passes on the Scott Base to McMurdo Station walking track, 2842 on the Wind Vane Hill walking track, 3561 on the Round Observation Hill walking track, and 10936 on the Up Observation Hill track between January 2009 and January 2011. On the Crater Hill summit walking track there were 693 passes in 2009. Higher counts were recorded on all tracks in the 2010/2011 summer season, compared with the 2009/2010 summer, and the highest frequency of visitors occurred on Sundays in the summer months. Peak daily counts at the Wind Vane Hill track coincided with the arrival of tourist ships. There was no relationship between the number of passes on the track and the measured impacts, indicating that higher usage of a formed track had little cumulative impact. Track width and track incision were related to the slope of the terrain, with tracks traversing flatter areas generally wider (R² = 0.85) and less incised (R² = 0.96) than those traversing steeper hillsides. Soil dry bulk density was higher in the walking tracks compared with the adjacent control areas (p <0.05).