CodePudding user response：

Reverse the values and use cummax to get the cumulative maximum:

df["lim_sup"] = df.loc[::-1, "value"].cummax()