Finding nonzero global sections of a line bundle

283 Views Asked by At

Let $V_0 = \{s \neq t\}$ and $V_1 = \{s \neq -t\},$ which together give an open cover of $\mathbb{P}^1.$ Let $L'$ be the line bundle defined by the transition functions $h_{01}: V_0 \cap V_1 \rightarrow GL_1(\mathbb{C})$ is $h_{01}([s:t]) = ((s+t)/(s-t))^2$ and $h_{10} : V_0 \cap V_1 \rightarrow GL_1(\mathbb{C})$ is $h_{10}([s:t]) = ((s-t)/(s+t))^2.$

What would global sections on $L'$ look like? The hint is to find functions $s_0: V_0 \rightarrow \mathbb{C}$ and $s_1:V_1 \rightarrow \mathbb{C}$ which patch together, but I don't understand what it means for functions to patch together.

What would the basis of $\Gamma(\mathbb{P}^1, L')$ look like?

2

There are 2 best solutions below

0
On BEST ANSWER

The other answer has the general theory of functions patching together covered pretty well, but there's an important difference for line bundles specified in terms of transition functions. For a line bundle $L$ on a space $X$ given in terms of a collection of transition functions $h_{ij}$ for an open cover $U_i$ of $X$, a global section is equivalent to a family of sections $s_i\in L(U_i)$ for all $i$ so that $s_i = h_{ij} s_j$ on $U_i\cap U_j$.

In our case, there's only two open sets in our cover and $h_{01}=h_{10}^{-1}$ where both are defined, so we only have one equation to worry about: $s_0 = \frac{(s+t)^2}{(s-t)^2} s_1$ for $s_i\in L(U_i)$. So in order to find global sections we'll need to find sections $s_0,s_1$ satisfying this equation.

First we need to identify $L(U_i)$. As both $U_0$ and $U_1$ are isomorphic to copies of affine space and $L$ is trivial on each of them, we know that we'll get a polynomial algebra in one variable over $\Bbb C$ as $L(U_i)$. For $U_0$, we can take our coordinate for affine space to be $\frac{s+t}{s-t}$, and for $U_1$ we can take our coordinate to be $\frac{s-t}{s+t}$, so we have that $L(U_0)=\Bbb C[\frac{s+t}{s-t}]$ and $L_1=\Bbb C[\frac{s-t}{s+t}]$.

From looking at our equation, we see that $s_0$ should be a polynomial in $\frac{s+t}{s-t}$ of degree at most two and similarly $s_1$ should be a polynomial in $\frac{s-t}{s+t}$ of degree at most two. By writing $s_0=\frac{p(s-t,s+t)}{(s-t)^2}$ and $s_1=\frac{q(s-t,s+t)}{(s+t)^2}$ for homogeneous degree two polynomials $p,q$ , we see that our equation forces $p=q$ and so by specifying $p$ we can think of a global section as a homogeneous degree two polynomial in $s-t$ and $s+t$. So we can think of a basis for the homogeneous polynomials of degree two as a basis for the space of global sections of this line bundle.

0
On

Patch together means that the functions $s_0$ and $s_1$ agree on the intersection $V_0 \cap V_1$, so that they can be "glued" to give a section on all of $\mathbb{P}^1$. Let's compare to a standard topological result:

Fact 1 Let $X$ be a topological space with open sets $U_1, U_2 \subseteq X$ such that $U_1 \cup U_2 = X$. Let $Y$ be any topological space. If $f_1 : U_1 \to Y$ and $f_2 : U_2 \to Y$ are continuous functions such that $f_1|_{U_1 \cap U_2} = f_2|_{U_1 \cap U_2}$, then there is a unique continuous function $f : X \to Y$ such that $f|_{U_1} = f_1$ and $f|_{U_2} = f_2$.

Proof Sketch. Let $$f(x) = \begin{cases} f_1(x) : x \in U_1 \\ f_2(x) : x \in U_2 \end{cases}$$ Show that $f$ is well-defined and continuous, then that is the unique continuous function satisfying the desired property.

More generally, we have the following:

Fact 2 Let $X$ and $Y$ be topological spaces. Let $C$ be the presheaf of sets on $X$ defined by letting $C(U)$ be the set of continuous functions $U \to Y$ (with restriction maps $f \mapsto f|_V : C(U) \to C(V)$ whenever $V \subseteq U$). $C$ is a sheaf of sets on $X$.

The proof of Fact 2 is very similar to the proof of Fact 1 (which just verifies a special case of checking the equalizer condition for this presheaf). And there is another related fact:

Fact 3 Let $f : X \to Y$ be a continuous function between topological spaces. Let $D$ be the presheaf of sets on $Y$ defined by $D(U) = \{g : U \to X \mid g~\text{is continuous and } f \circ g = \operatorname{id}_U\}$. $D$ is a sheaf of sets on $Y$.

(Indeed, Fact 3 implies Fact 2 by the étale space construction for presheaves) In algebraic geometry, we have a very similar situation:

Fact 3' Let $f : X \to Y$ be a morphism of schemes. Let $C$ be the presheaf of sets on $Y$ defined by letting $C(U)$ be the set of morphisms $g : U \to X$ such that $f \circ g = \operatorname{id}_U$. $C$ is a sheaf of sets on $Y$.

The line bundle $L$ can be thought of (or defined as!) as a scheme morphism $f : L \to \mathbb{P}^1$ such that $\Gamma(U,L) = C(U)$ in the notation of Fact 3', which tells us that we're allowed to "glue" local sections of a line bundle together as long as they agree on the intersections of their domains of definition. In other words, $\Gamma(\mathbb{P}^1,L)$ is the equalizer of the maps $\Gamma(V_0,L) \times \Gamma(V_1,L) \rightrightarrows \Gamma(V_0 \cap V_1,L)$, with the first map being given by $(s_0, s_1) \mapsto s_0|_{V_0 \cap V_1}$ and the second map being given by $(s_0, s_1) \mapsto s_1|_{V_0 \cap V_1}$.