In Andrew, Majda- Vorticity and incompressible flow page 93, there is a theorem which is not proved:
Take a smooth incompressible (free divergence) vector field $v$ in $\mathbb{R}^2$. Call $w$ its vorticity (its curl), and suppose it is $L^1(\mathbb{R}^2)$. Then $v$ admits a radial energy decomposition, which means:
There exists a radially symmetric smooth scalar field $\overline{w}$ such that we can write $v = u + b$ where $u$ and $b$ satisfy:
$u$ is divergence free and $u \in L^2(\mathbb{R}^2)$. $b(x)= x^{\perp}|x|^{-2}\int_0^{|x|} s\overline{w}(s)ds $.
In the page before we prove that for a smooth incompressible vector field, using the biot savart law: $v(x) = (K_2 \ast w) (x) $ and making a taylor expansion of order 1 of the kernel $K_2$: for large $x$, $v(x)= x^{\perp}|x|^{-2}\int_{\mathbb{R}^2} w + O(|x|^{-2})$ and it is supposed to be a consequence of that, but I don't see why !!