In the past, I have made the statement on DSN that there can be no such thing as a sustainable deep-sea fishery. My reasons for this are that

1. Deep-sea fish are slow growing and long lived due to the cold temperatures of deep water. 
   
2. This results in low turnover, or replacement, of commercially important, large individuals in the population. In other words, we harvest fishy grandmothers and grandfathers and we have to wait awhile  before a new batch of grandparents comes around.  Hey grandparents just don't grow on trees!
3. While we could theoretically have a healthy fishery the turnover rates are so low that it could never be commerically viable.

This week in the Proceedings of the National Academy of Science, Thresher and others publish a paper indicating the situation may be more dire. 

The authors looked at 8 exploited marine fish from three different depth groups:

1. <250m-Banded Morwong, Redfish, Jackass Morwong
2. 250-1000m-Black Oreo, Spiky Oreo
3. >1000m-Orange Roughy, Smooth Oreo, Warty Oreo
   

Using otoliths, fish ear bones, they were able to determine growth rates for 555 individuals. Like tree rings, the thickness of a ring in a otolith indicates growth of the fish for that year. The data tell an interesting story. Shallow-water fish (1) show an increase in growth rate from as early as ~1910 to ~2000. In one species, growth rates of individuals were 28.5% faster than they were just 50 years ago. Mid-depth fish (2) show no change. In contrast, growth rates of deep fish (3) decreased over time as much as 27.9% since the late 1880's.

The authors demonstrate that these long term changes in growth rates reflect the dramatic changes in ocean temperatures occurring in the last century. Because growth rate is physiologically tied to temperature the continued warming of the ocean's surface has quickened growth, while the continued cooling of deep water has slowed growth.

As an aside I really, really wanted to reproduce the fantastic figures from the paper here. But given that one journal has already harassed a fellow Sb'ling over doing the same in an obvious case of fair use practices, I decided to decline. However, head over to PNAS and download the paper for yourself.