Let $D$ be a $\mathbb{Q}$-divisor in a smooth variety $X$. In Lazarsfeld book "Positivity in Algebraic Geometry 2" I found Proposition 9.5.13 saying that if for any $x\in D$ we have $mult_xD < 1$ then the pair $(X,D)$ is klt.

I am wondering if it possible to check that a pair is klt by compunting the discrepancies of a partial resolution and apply the fact above. I will be more precise.

Assume we have a birational divisor contraction $f:Y\rightarrow X$ such that $f^{-1}D\cup Exc(f)$ is simple normal crossing and the discrepancies of $(X,D)$ with respect any exceptional divisor are gretaer than $-1$. Let $Z = f(Exc(f))\subset X$. If for any $x\in D\setminus (D\cap Z)$ we have $mult_xD< 1$ is it true that $(X,D)$ is klt?


If I interpreted correctly your question I guess the answer is positive. Here is the statement:

Let $f:Y\rightarrow X$ be a proper birational morphism between normal $\mathbb{Q}$-factorial varieties, and let $D$ be an effective $\mathbb{Q}$-divisor in $X$. Let us write $$K_Y = f^{*}(K_X+D)+\sum_{i}a_iE_i-\widetilde{D}$$ where $E_i$ are $f$-exceptional divisors and $\widetilde{D}$ is the strict transform of $D$. If $a_i \leq 0$ for any $i$, and $mult_y(\widetilde{D}-\sum_{i}a_iE_i) < 1$ for any $y\in Y$ , then the pair $(X,D)$ is klt.

To prove this you can proceed as follows. We have $\widetilde{D}-\sum_{i}a_iE_i = \widetilde{D}+\sum_{i}b_iE_i$ with $b_i = -a_i \geq 0$. Then, $\widetilde{D}+\sum_{i}b_iE_i$ is effective. Since $mult_y(\widetilde{D}+\sum_{i}b_iE_i) < 1$ for any $y\in Y$, by Proposition 9.5.13 of "Positivity in Algebraic Geometry 2" we have that the multiplier ideal $\mathcal{I}(\widetilde{D}+\sum_{i}b_iE_i)$ is trivial. Then, by Section 9.3.B of "Positivity in Algebraic Geometry 2" we get that $(Y,\widetilde{D}+\sum_{i}b_iE_i)$ is klt. Now, we may write $$K_Y+\widetilde{D}+\sum_{i}b_iE_i = K_Y+\widetilde{D}-\sum_{i}a_iE_i = f^{*}(K_X+D).$$ Finally, since $(Y,\widetilde{D}+\sum_{i}b_iE_i)$ is klt, by Lemma 3.10 in http://arxiv.org/abs/alg-geom/9601026 the pair $(X,D)$ is klt as well.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.