Monitoring purposes determine the selection of variables, location of sampling sites, and sampling frequency. The selection should provide the best signal to noise ratio for the parameters of interest. For trend and surveillance monitoring, the deepest point of a lake, where different inputs to the lake are integrated, is frequently selected. However, the representativeness of a single site is often questioned, especially for large lakes. Based on data collected from 10 sampling points during 11 survey expeditions in August 2001-2011 to the large shallow Lake Võrtsjärv, Estonia, we studied the spatial and annual variability of environmental and phytoplankton variables and analysed the representativeness of a permanent sampling station for the whole lake conditions. The two southernmost stations under the influence of the main tributary deviated clearly from the homogeneous group of the other eight stations, which we termed 'Võrtsjärv Proper'. Among the stations of Võrtsjärv Proper, the year-to-year variability dominated strongly over the spatial variability, the latter being almost negligible for most of the variables. Surface water temperature and water level explained approximately half of the total variability in parameters commonly used in ecological status assessment of lakes. This has serious implications for using these variables to detect human impacts in Võrtsjärv. Our study showed that the deep sampling site, which was characterized by the lowest average variability of the parameters measured and was representative of more than 90% of the lake aquatory, possesses all necessary qualities required of a permanent surveillance monitoring station.