How are you quantifying the amount of each species in the past? Or is this just wish-fulfillment hogwash?

For example by looking at historical fishing records. One paper that does this back into the 1750s across mulpile regions and species is this one:

https://doi.org/10.1016/j.jmarsys.2008.12.011

It is behind a pay wall, but I'll quote the methods here:

2. Materials and method

2.1. The pre-industrialized period When it comes to written testimonies of pre-industrial fishing activity, the time frame is in most cases limited to a few hundred years. The starting point of the industrial period is usually considered to be the second half of the 19th century in the major European and North American fisheries. The industrialization of fisheries is characterized by a number of technological changes in fishing techniques, which all contributed to more efficient fishing operations. In the 1860s machine made cotton nets gradually replaced the old heavier hemp nets, and in the following decades steam propulsion and, from the turn of the 20th century, motor propulsion gave extra trawling power and the ability to move independently of prevailing winds. In principal then, historical evidence can be found as far back in time as fishing has taken place. However, the demands for available, consistent and reliable historical data limit the time frame consider-ably. Another limitation is that the historical datasets need to cover a number of years in order to be suitable for testing for climate signals. Therefore, the following discussion of historical data only includes datasets, which span more than c. 50 years.

2.2. Written documents With regard to written documents the oldest known data for fishing are from Europe. During the course of the 14th–16th centuries writing on paper became increasingly common in Europe. This is also the time when the bureaucracy of the emerging modern state bureaucracy as well as larger private enterprises gradually became established. These developments ensured two aspects of fisheries record keeping. First of all, the fiscal interest of the modern state ensured an interest in accurate numbers. Secondly, state interest often lead to an institutionalization of fisheries regulations, whereby a steady, recurring and often quite uniform annual collection took place. In line with this, ancient record keeping deals exclusively with commercially important species. Cod, herring, anchovy, sardine, salmon, various flatfish and tuna therefore are the most prominent in this type of historical material. Thus, along with a bias towards European and Atlantic fisheries, there is an inherent bias in terms of which species feature in historical material. During the last decade several large scale projects have been under way trying to recover archival material for reconstructing historical fish stocks, and this review stands on the shoulders of these efforts, which are producing online free access databases. The History of Marine Animal Populations project of the Census of Marine Life programme (2000–2010) is an umbrella for the research of c. 100 historians, archaeologists and marine scientists trying to assess what lived in the oceans before modern times (http://www/. hmapcoml.org/). Within the INCOFISH Specific Targeted Research Project of the European Community (2005–2008) the recovery of time series for historical fisheries is a means to shift the baseline of understandings of ecosystem functioning (http://www.hull.ac.uk/ incofish/index.htm). The Sea Around Us Project of the University of British Columbia is mainly concerned with fisheries developments since 1950, but also has strong components stretching back hundreds of years (http://www.seaaroundus.org/).

2.3. Long-term environmental time series Comparing long-term changes in fish populations with environ-mental variability is strongly aided by the existence of equally long time series of environmental variability. Records of temperature, wind, air pressure and similar parameters for environmental variability and changes rarely exist for longer than c. 100–200 years back in time. Assessing historical climate reconstructions would be a topic for a paper in its own right, but it should be mentioned that much effort is currently being put into such reconstructions, and the following portals hold valuable collections of such time series: CLIVAR, Climate variability and Predictability (http://www.clivar/. org/). KNMI, The Royal Dutch Meteorological Institute, among other resources, provides access to the global CLIWOC project trying to reconstruct the global weather from 1750–1850 (http://climexp.knmi/. nl/). The NOAA Satellite and Information Service is hosting a large amount of temperature proxies (http://www.ncdc.noaa.gov/paleo/ data.html), while datasets from Greenland ice cores are available from the University of Copenhagen, (http://www.glaciology.gfy.ku.dk/). Finally, a large collection of dataset can be extracted from NASA at (http://gcmd.nasa.gov/index.html). A very large project currently in progress is Millenium, which has as its main goal to reconstruct the climate variability in Europe during the last 1000 years to see whether changes in the last one hundred years are unique in scale (http:// geography.swan.ac.uk/millennium/index.htm).

P.s: Is there good way to share the whole pdf?