Ideally, repeated cross sectional study is that you take cross section of given population at different time frame and then define the changes in given parameter for example the weight of 5years old child in year 2015, 2016 and 2017 or mortality in 2015, 2016 and 2017 etc.
If we observe a number of units within the same time period, we construct a cross-section; if we construct different cross-sections in different periods we obtain a data structure named repeated cross-sections.