It is the difference between local and global. The sentence
Let $C$ be a non-singular curve and denote by $t\in \Delta$ the local unifomization variable
is very sloppy and one should refrain from using this language. It should be something like:
Let $C$ be a connected Riemann surface of hyperbolic type (e.g. a compact Riemann surface of genus $\ge 2$). Let $f: \Delta\to C$ be the holomorphic universal covering map, where $\Delta=\{t\in {\mathbb C}: |t|<1\}$ is the unit disk. By abusing the terminology, we regard $t$ as the uniformization variable on $C$.
See this Wikipedia article for more details.
Edit. the answer also depends on the context. For instance, in Algebraic Geometry, local uniformizer is used for a generator of the local ring of an algebraic curve at the given point, see this question. From the view point of complex analysis or differential geometry, such local uniformizer is the same thing as a local chart once restricted to a suitable open subset of the Riemann surface. In many cases, the word "local" is dropped.