The mass eruption rate feeding a volcanic plume is commonly estimated from its maximum<br/>height. Winds are known to aﬀect the column dynamics causing bending and hence reducing the<br/>maximum plume height for a given mass eruption rate. However, the quantitative predictions including<br/>wind eﬀects on mass eruption rate estimates are not well constrained. To ﬁll this gap, we present a series<br/>of new laboratory experiments on forced plumes rising in a density-stratiﬁed crossﬂow. We identify three<br/>dynamical regimes corresponding to increasing eﬀect of wind on the plume rise. The transition from one<br/>regime to another is governed by two dimensionless velocity scales deﬁned as a function of source and<br/>environmental parameters. The results are found consistent with the conditions of historical eruptions and<br/>provide new empirical relationships to estimate mass eruption rate from plume height in windy conditions,<br/>leading to valuable tools for eruption risk assessment.