Albedo plays an important role in land–atmosphere interactions and local climate. This study presents the impact on simulating regional climate, and the evolution of a drought, when using the default climatological albedo as is usually done in regional climate modelling, or using the actual observed albedo which is rarely done. Here, time-varying satellite derived albedo data is used to update the lower boundary condition of the Weather Research and Forecasting regional climate model in order to investigate the influence of observed albedo on regional climate simulations and also potential changes to land–atmosphere feedback over south-east Australia. During the study period from 2000 to 2008, observations show that albedo increased with an increasingly negative precipitation anomaly, though it lagged precipitation by several months. Compared to in-situ observations, using satellite observed albedo instead of the default climatological albedo provided an improvement in the simulated seasonal mean air temperature. In terms of precipitation, both simulations reproduced the drought that occurred from 2002 through 2006. Using the observed albedo produced a drier simulation overall. During the onset of the 2002 drought, albedo changes enhanced the precipitation reduction by 20 % on average, over locations where it was active. The area experiencing drought increased 6.3 % due to the albedo changes. Two mechanisms for albedo changes to impact land–atmosphere drought feedback are investigated. One accounts for the increased albedo, leading to reduced turbulent heat flux and an associated decrease of moist static energy density in the planetary boundary layer; the other considers that enhanced local radiative heating, due to the drought, favours a deeper planetary boundary layer, subsequently decreasing the moist static energy density through entrainment of the free atmosphere. Analysis shows that drought related large-scale changes in the regional climate favour a strengthening of the second mechanism. That is, the second mechanism is stronger in a drought year compared to a normal year and this difference is larger than for the first mechanism. When both mechanisms are active, the second mechanism tends to dominate across the model domain, particularly during the 2002 drought period. The introduction of observed inter-annual variations in albedo produces an enhancement of the first mechanism and a weakening of the second mechanism during the onset of the drought.