1: \begin{abstract}

2: 	Let $M(\T^d)$ be the space of complex bounded Radon measures defined on the $d$-dimensional torus group 

3: 	$(\R/\Z)^d=\T^d$, equipped with the total variation norm $\|\cdot\|$; and let $\hat\mu$ denote the Fourier 

4: 	transform of $\mu\in M(\T^d)$. We address the super-resolution problem: For given spectral (Fourier transform) 

5: 	data defined on a finite set $\Lambda\subset\Z^d$, determine if there is a unique $\mu\in M(\T^d)$ of minimal norm for which 

6: 	$\hat\mu$ equals this data on $\Lambda$. Without additional assumptions on $\mu$ and $\Lambda$, our main 

7: 	theorem shows that the solutions to the super-resolution problem, which we call \emph{minimal extrapolations}, 

8: 	depend crucially on the set $\Gamma\subset\Lambda$, defined in terms of $\mu$ and $\Lambda$. For example, 

9: 	when $\#\Gamma=0$, the minimal extrapolations are  singular measures supported in the zero set of an analytic 

10: 	function, and when $\#\Gamma\geq 2$, the minimal extrapolations are singular measures supported in the 

11: 	intersection of $\#\Gamma\choose 2$ hyperplanes. By theory and example, we show that the case $\#\Gamma=1$ 

12: 	is different from other cases and is deeply connected with the existence of positive minimal extrapolations. 

13: 	This theorem has implications to the possibility \emph{and} impossibility of uniquely recovering $\mu$ from 

14: 	$\Lambda$. We illustrate how to apply our theory to both directions, by computing pertinent analytical examples. 

15: 	These examples are of interest in both super-resolution and deterministic compressed sensing. Our concept of an \emph{admissibility range} fundamentally connects Beurling's theory of minimal extrapolation \cite{beurling1989balayage,beurling1989interpolation} with Cand\`{e}s and Fernandez-Granda's work 

16: 	on super-resolution \cite{candes2014towards}. This connection is exploited to address situations 

17: 	where current algorithms fail to compute a numerical solution to the super-resolution problem.

18: \end{abstract}

19: