From the first search I tried the first result says:
According to the Second IMO GHG Study 2009, which is the most comprehensive and authoritative assessment of the level of GHG emitted by ships, international shipping was estimated to have emitted 870 million tonnes, or about 2.7% of the global man-made emissions of CO2 in 2007. Exhaust gases are the primary source of GHG emissions from ships and carbon dioxide is the most important GHG, both in terms of quantity and of global warming potential.
There's a comparison here with this graph:

Wikipedia has a page specifically on the "water pollution from sea transport" question that covers several aspects with links to other sources.
The second part of your question is impossible to answer with any certainty unless you're willing to spend a lot of money funding the study. There are a lot of marginal costs that would change and among other complexities it's likely that the total impact of international trade is significantly positive (pdf link). It follows that removing it would reduce economic growth and effectively lower global GDP. The whole "is trade good" question is the subject of considerable research and your question could be seen as "trade is good economically, but is it also good under the triple bottom line approach?"
To take one specific example: there are a very small number of computer chip producing plants, and generally only one or two producing "current generation" chips. Mostly because those plants are extremely expensive to build. So they are built to run the key parts of the process as fast as possible all the time. The resulting chips are shipped all over the world.
There are only a couple of these plants, and they're expensive. De-globalising those plants would mean we'd need more of them, and they don't scale down very well. So rather than one, $10B plant we'd probably need 10, $2B plants to get the same production. Even if it's only $1.5B that's still quite a lot of money, plus of course all the supporting people and supply chains (we'd need 9 more silicon refineries, for example). It's going to cost a lot of money just to regionalise chip production to 10 parts of the globe, and that's not very "regional" to most people.
As well, we have a smaller plant that is probably not able to run at full capacity all the time because it serves a smaller group of people (so random variations in demand are proportionally larger as well as more likely). This pushes the cost up further. The need to finance it locally means that most of those plants will need to pay higher interest rates on the money they borrow to build. And so on.
We also need to consider the demand side of the equation. What those plants produce will now be more expensive. That means they will sell fewer of them (would you buy a new smartphone every two years if it was $5000 instead of $1000? Would everyone?)
This pattern will reproduce itself across everything that's not already locally produced to some extent, and even to some things that are due to reduced competition. Some things will get dramatically more expensive due to the wide range of global labour prices affecting labour-intensive goods. In countries with a high minimum wage and effective prohibitions on slavery (much of Europe, Australia) that shift will be especially pronounced. Australia will also be hit very badly by the loss of its huge food exports (with corresponding loss of population in countries no longer able to import food).
You could probably get a better idea of how that would pan out by reading about the early 20th century when international trade tariffs were high, import quotas low, and many industries completely protected at a national level. As free trade was introduced things changed, usually in the direction of more money. Assuming your unit of "local" is the national-state, of course. If it's more local, like the local food people who often use 100km or 100 miles, you would prevent production of anything electronic that's more complex than a calculator. But to enforce that you'd need unrealistic levels of political power.
edit in response to comment: locally grown is more mixed.
The greenhouse-grown salads in the UK are a classic example where it's better to grow them in Africa and fly them to the UK than grow them locally. But trucking food around inside the USA so that your locally-grown salad might have been shipped 1000 miles to be processed then shipped back? Not so much.
There are centralisation/resilience risks that come into play as well. If the only source of something breaks things can get difficult while it's being fixed. The "cities are three days from starvation" idea is kind of true but it's not a great way to look at it (that link explains why). And food is a bad example - people can go 30 days without food, but only 3 without water.
That's why governments (especially) pay such a huge premium for distributed, redundant supplies of critical things like water, electricity and transport. The grid failure in NE USA in 2003 is a great example of the cost of not having resilience - electricity turns out to be more important than many people thought, and the extra cost of secure supply turns out to lower than the cost of having it fail. But from a pure financial return point of view it makes no sense to spend the extra money now for a problem that probably wont happen this quarter. Which is why governments would ideally own those things and operate on longer timescales.
But that applies to individuals as well. Paying the premium to buy locally will help create resilience in your city. Whether you can afford that, whether the extra is worth while, whether that even works are questions that have local answers. If there's a local farmers market or food co-op those are more likely to achieve that goal than "local" or "organic" labels in the supermarket, simply because supermarkets often make it hard to investigate the sincerity of the label while a co-op should have researchers who do that for you, and at the farmers market you can ask the seller directly.
It might make more sense to invest in food storage and buy in bulk. That way you save money or can afford to shop with other than pure price in mind (for example, may be able to buy local rather than imported), plus you have the security of having a couple of weeks food on hand if there is a disaster.
Similarly, if you can do some gardening you gain a local supply of food that is less likely to be affected by disasters, plus if you're lucky with your location and plan well it may be less environmentally costly than buying that food. Community or guerilla gardening is a cheap option in many places for people who don't have their own land. Plants in boxes on your balcony are unlikely to meet a triple bottom line analysis, unfortunately.