To better understand the operating mode of the deep oceans during fundamentally warm climate intervals, we present new Nd isotope data from Deep Sea Drilling Project and Ocean Drilling Program sites in the North Pacific to expand the reconstruction of water mass composition and structure during the early Cenozoic. Fossil fish debris from Sites 192, 464, 465, 883, 884 and 1208 (paleowater depths spanning 900 to 4000 m) were used to reconstruct the water mass composition from ~85 to 30 Ma. The fish debris is shown to not be overprinted as there was no systematic offset between the detrital silicate and the fish debris composition. Cleaned and uncleaned fish debris were both included in the reconstruction of water mass composition as they were found to record the same Nd isotope composition. North Pacific deep water convection occurred from ~67 to 45 Ma, the peak in production is recorded by broadly coincident trends at Sites 192, 464 and 883. Further support for North Pacific deep-water convection during the early Paleogene are the geographic trends in detrital silicate versus fish debris composition, greater separation at the more northerly Emperor Seamount sites, and the location of the most radiogenic detrital values at the Emperor Seamount sites. The Emperor Seamount chain likely played a major role in the flow of the North Pacific deep-water mass as it acted as a physical barrier to flow at deep-water sites compared to shallow depths (albeit still deep-water). εNd values indicate the timing of the cessation of major, deep convection in the North Pacific occurred much earlier, ~52 Ma than the timing obtained from shallower Shatsky Rise sites, ~45 Ma. Convection in the North Pacific likely produced a dense water mass that influenced the deeper sites in this study more than the shallower sites until ~52 Ma when convection was not as intense or the waters were not sufficiently dense to impact the deeper sites. Deep water convection was most intense during the relatively “cool” portion of the Late Cretaceous and Early Paleocene.

Knowing the taxonomy of unicellular dinoflagellate cysts is important in reconstructing paleoenvironments, paleoclimatology and paleobathymetry. Different dinoflagellates tend to prefer different depths, salinity levels and climate. Species of dinocysts only occur during a certain interval in the geological record and are therefore an important tool for dating sediments. Biostratigraphic research uses age determination of sediments depending on first and last occurrences of different species, or co-occurrences of several species. However, it can be rather time consuming and difficult to identify dinoflagellate cysts for people that are not experts in this field. In an attempt to resolve this, a database was created for 145 genera of dinoflagellate cysts existing during Palaeocene and Eocene in the online accessible program Xper3. Morphological criteria or “descriptors” were added along with describing “states”. The number ofmorphological descriptors chosen were 13, with 3 to 15 states for each descriptor.Each genus was described in a matrix and was given one or more unique states for each descriptor, which were defined from the literature by original morphological descriptions of dinocyst genera and photographs of holotypes. A morphological identification key was automatically created in Xper3 from this database, which enabled identification of genera by choosing unique states that were visible for the dinocysts being analysed.A test using photographs with different levels of preservation of dinoflagellate cysts of Eocene age of Aquitaine basin was undertaken in order to assess the reliability of the morphological identification key. The aim was also to see which morphological criteria were more common and how many descriptors and states were needed to reach an end-result of 5 remaining genera, including the correct genus.Errors that were caused in the identification key were immediately reviewed and re-tested. In 38 of 43 tries, a maximum of five genera were remaining, with the correct genus included. This confirmed that the identification key worked relatively well.Another test demonstrated how the identification key worked for identifying dinocysts with an optical microscope in unknown samples, which are photographed and published in the present report, along with data showing how many descriptors and states were used, remaining genera and which states were chosen for each genus.In general, only 1-8 descriptors were necessary and 5 the most common number used. The morphological character (state) “type of ornamentation” was the most widely used followed by “distribution of ornamentation”, “shape of the cyst” and “size of the ornamentation”. However, the archeopyle (excystment opening), was not always visible in the dinocysts, but is still considered a key morphological descriptor. / Dinoflagellater är en typ av mikroskopiska plankton som kan bilda så kallade cystor eller vilosporer för att skydda sig vid kritiska miljöförändringar. Dessa cystor kan ligga vilandes i sediment under en längre tid. Det är också dessa cystor som bevaras som fossil. Eftersom det finns en sådan hög mångfald av dinoflagellatcystor i fossila sediment, liksom att de mikroskopiska, existerar över stora delar av jorden samt att de har en snabb evolution, så används de ofta till att datera sediment inom biostratigrafi.För att kunna datera sediment med hjälp av dinoflagellatcystor, krävs det att man kan identifiera de olika arterna och veta i vilken tidsålder de existerade. Att klassificera dinoflagellatcystor kan därför vara mycket svårt och ta en hel del tid om man inte är någon expert. I denna studie har därför en databas och identifieringsnyckel skapats i ett onlinebaserat dataprogram kallat Xper3. Xper3 är primärt riktat till att hjälpa personer med grundläggande kunskaper i att beskriva utseendemässiga karaktärer hos olika typer av mikrofossil, inklusive dinoflagellatcystor.I denna identifieringsnyckel använde jag mig av 145 olika dinoflagellatsläkten vilka alla existerade någon gång under tidig Paleogen, dvs hela Paleocen och/eller Eocen (66–34 miljoner år sedan). Dessutom användes 13 olika morfologiska beskrivningar ”descriptors” med 3–15 olika svar ”states” per beskrivning.Målet är att få en identifieringsnyckel som unikt kan identifiera olika typer av dinoflagellatcystor på genusnivå, med ett förslag på maximalt fem återstående släkten (genera). Studien undersökte även vilka morfologiska beskrivningar som är mest användbara och hur många olika typer av beskrivningar som behövs för att analysera dinoflagellatcystor.För att testa om identifieringsnyckeln fungerade, användes olika typer av kända och okända dino-flagellatcystor i ett så kallat ”Test 1: Kända dinoflagellatcystor” och ”Test 2: Okända dinoflagellat-cystor.I det första testet, ”Test 1”, grupperades kända dinoflagellatcystor i olika typer av bevarande tillstånd så som ”välbevarade”, ”tveksamt bevarande” och ”väldigt dåligt bevarade” dinoflagellat-cystor. Detta test förde statistik på hur många släkten av förslag som kom upp i slutändan och vilka morfologiska beskrivningar som användes mest och hur många olika beskrivningar som behövdes.Det andra testet, ”Test 2”, undersökte hur denna identifieringsnyckel fungerade att analysera dinoflagellatcystor från okända prover under mikroskop. Fotograferade dinoflagellatcystor presente-rades i denna rapport tillsammans med detaljer på hur de hade identifierats.I de flesta fallen, kunde identifieringsnyckeln ge ett förslag på maximalt fem återstående släkten av dinoflagellatcystor, där det korrekta släktet var ett utav förslaget, vilket tyder på att identifierings-nyckeln fungerar. Den mest använda morfologiska beskrivningen var ”typ av ornament”, och antal morfologiska beskrivningar som behövdes var mellan ett till åtta, där fem var det antal beskrivningar som mest användes i att identifiera dinoflagellatcystor i identifieringsnyckeln.