1: \begin{abstract} 

2: %

3: Recent studies have shown that the number counts of convergence peaks

4: $N(\kappa)$ in weak lensing (WL) maps, expected from large forthcoming

5: surveys, can be a useful probe of cosmology.  We follow up on this

6: finding, and use a suite of WL convergence maps, obtained from

7: ray-tracing N--body simulations, to study (i) the physical origin of

8: WL peaks with different heights, and (ii) whether the peaks contain

9: information beyond the convergence power spectrum $P_\ell$.  In

10: agreement with earlier work, we find that high peaks (with amplitudes

11: $\gsim$ 3.5$\sigma$, where $\sigma$ is the r.m.s.  of the

12: convergence~$\kappa$) are typically dominated by a single massive

13: halo.  In contrast, medium--height peaks ($\approx0.5-1.5\sigma$)

14: cannot be attributed to a single collapsed dark matter halo, and are

15: instead created by the projection of multiple (typically, 4-8) halos

16: along the line of sight, and by random galaxy shape noise.

17: Nevertheless, these peaks dominate the sensitivity to the cosmological

18: parameters $w, \sigma_8$, and $\Omega_m$.  We find that the peak

19: height distribution and its dependence on cosmology differ

20: significantly from predictions in a Gaussian random field.  We

21: directly compute the marginalized errors on $w, \sigma_8$, and

22: $\Omega_m$ from the $N(\kappa) + P_\ell$ combination, including

23: redshift tomography with source galaxies at $z_s=1$ and $z_s=2$.  We

24: find that the $N(\kappa) + P_\ell$ combination has approximately twice

25: the cosmological sensitivity compared to $P_\ell$ alone.  These

26: results demonstrate that $N(\kappa)$ contains non-Gaussian information

27: complementary to the power spectrum.

28: %

29: \end{abstract}}