1: \begin{abstract}

2: We propose nonparametric estimators of the occupation measure and the

3: occupation density of the diffusion coefficient (stochastic volatility) of a

4: discretely observed It\^o semimartingale on a fixed interval when the

5: mesh of the observation grid shrinks to zero asymptotically. In a first

6: step we estimate the volatility locally over blocks of shrinking

7: length, and then in a second step we use these estimates to construct a

8: sample analogue of the volatility occupation time and a kernel-based

9: estimator of its density. We prove the consistency of our estimators

10: and further derive bounds for their rates of convergence. We use these

11: results to estimate nonparametrically the quantiles associated with the

12: volatility occupation measure.

13: \end{abstract}