Baffin Island study disappoints: The illusive ‘coup de grace’ on the Medieval Warm Period

Guest essay by Sebastian Lüning

Big news on 4. December 2015 by the Earth Institute of Columbia University. In a press release the institute claimed that climate and human history has to be re-written and climate had no major influence on Viking settlement on Greenland:

Study Undercuts Idea That ‘Medieval Warm Period’ Was Global

Vikings May Not Have Colonized Greenland in Nice Weather

A research team led by the glaciologist Nicolás Young claimed in a new paper in Science Advances that there actually was no Medieval Warm Period (MWP) in Greenland. Instead, they propose that a cold phase occurred 1000-1250 AD in the region which even triggered glacier advance. Interestingly, Young and colleagues base their model on a study of a single local glacier on Baffin Island which they generalize for the entire region. In the press release they state:

“It’s becoming clearer that the Medieval Warm Period was patchy, not global,” said lead author Nicolás Young, a glacial geologist at Columbia University’s Lamont-Doherty Earth Observatory. “The concept is Eurocentric—that’s where the best-known observations were made. Elsewhere, the climate might not have been the same.”

The same press release cites another researcher – independent of the study – who declares that this new paper would finally disprove the global significance of the Medieval Warm Period:

Gifford Miller, a paleoclimatologist at the University of Colorado, called the paper “a coup de grace on the Medieval Warm Period.” Miller said it shows “with great clarity of evidence” that “the idea of a consistently warm Medieval period is certainly an oversimplification and of little utility.”

An interesting, provocative claim. But what did other studies in the region find? Do they support the idea of a Medieval Cold Period in Greenland and neighbouring regions? A good task for the Medieval Warm Period Mapping Project. In this initiative we are currently mapping temperature reconstructions around the globe to better understand the medieval climate change. The ‘MWP Mapping Project’ is supported by crowd funding and has now reached a funding level of nearly 50%. More information here. The interactive map can be accessed here. Click on the dots and summary panels and key figures will appear, together with links to the original peer-reviewed paper. Red dots indicate case studies which yielded evidence for warm medieval climate, blue dots indicate reported cool temperatures (Fig. 1).

Figure 1: Location map of MWP case studies in Arctic Northeast Canada and West Greenland. Numbers refer to description in text. Interactive Google Maps project here.

Already at first sight it becomes clear that there are a large number of studies in the region which indicate warm MWP conditions. Are the cold temperatures suggested in the case study of Young et al. (2015) an exception rather than the rule? Do the locally advancing glaciers really indicate widespread cooling? Most of the studies illustrated on the map in Fig. 1 are not even cited by the Columbia University team. In the following we take a tour through the existing literature which paints a picture that is completely different to the one suggested by Young and colleagues.

(1) This is the location of the study by Young et al. 2015, describing glacier advance 1000-1250 AD interpreted as cooling climate.

(2) Close-by and already different: From the northwestern part of Baffin Island Briner et al. 2009 documented a prominent glacial retreat phase 800-1200 AD (Fig. 2).

Fig. 2: Time-distance diagram for glacier extent on the North Baffin Plateau. From Briner et al. (2009).

(3) Anderson et al. 2008 age-dated vegetation that was buried by advancing glaciers and suggest a warm phase 1000-1250 AD. They report cooling from 1280 AD onwards associated with glacier advance.

(4) This data point represents 94 glacier front locations on Baffin Island, which were studied by Miller et al. 2012a. Result: Warm phase 950-1275 AD (Fig. 3). Cooling of the Little Ice Age and glacier advance commenced abruptly between 1275-1300 AD, which further intensified 1430-1455 AD.

Fig. 3: Ice cap expansion dates base on 94 Arctic Canada sites. From Anderson et al. 2008.

(5) ‘Big Round Lake’ is located close to the Young-et-al. study area. Thomas & Briner 2009 originally described a warm phase 970–1150 AD from here, which however they thought was more than 1°C cooler than the present-day temperature level. Yet, the applied methodology is not straight-forward. Thomas & Briner measured thicknesses of varves, i.e. annually layered lake sediments. For the past few decades, there is a good correlation of thick varves and higher temperatures. But is this empirical relationship really valid for the entire last 1000 years and all locations? Regional correlations seem to indicate that this may not be the case. Previously, thick varves in this case study may have been actually associated with cooler summer temperatures, as Steve McIntyre documents on Climate Audit-Blog. Meanwhile also Balascio et al. 2015 appear to have adopted the new varve model for Big Round Lake and interpret the relatively thin varves 1000-1400 AD as evidence for warm temperatures and a well-developed MWP. The phase 1200-1300 AD forms an exception when varve thicknesses increase slightly, possibly indicating a slightly cooler interlude.

(6) Joynt III et al. 2001 studied diatoms in Fog Lake and documented a warm phase around 1000 AD which was 1°C warmer than the subsequent Little Ice Age (Fig. 4).

Fig. 4: Temperature reconstruction based on diatoms, Fog Lake, Baffin Island. From Joynt III et al. (2001).

(7) Another cold anomaly: Local glacier advance 1000-1300 AD is also known from North Cumberland Peninsula, as reported by Briner et al. 2009 who age-dated moraines. This seems to fit well with the results of Young et al. further north. Nevertheless, glacier advance can have several reasons, e.g. increased amounts of snowfall. Could this be the reason for local glacier expansion in parts of Baffin Island? Was MWP climate wetter and brought more precipitation in some areas? From Greenland we know that snow fall has increased over the past 120 years, in parallel with warming temperatures. Even today there are various glaciers that are actually expanding rather than shrinking.

(8) Donard Lake is located close to the expanding MWP glaciers in North Cumberland Peninsula. Moore et al. 2001 studied varves in the lake for the past 1200 years and suggest a period of warm summer temperatures 1050-1100 and 1200-1375 AD, followed by cooler temperatures of the Little Ice Age 1375-1820 AD (Fig. 5). Did the short cool interval 1100-1200 trigger glacier advance? It would be important to confirm this temperature development with independent non-varve-methodology, given the controversy in Big Round Lake (location 5).

Fig. 5. Temperature reconstruction of Donard Lake, Baffin Island, based on varve thickness. From Moore et al. 2001.

(9) Glaciers in Central Cumberland Peninsula were expanding 1000-1300 AD (Fig. 6), the opposite development to nearby North Cumberland Peninsula (location 7) (Briner et al. 2009). It becomes clear that glacier development is not uniform across Baffin Island, hence extrapolation of results from a single glacier location by Young et al. 2015 does not make sense. It can only be speculated why different glacier areas reacted differently to climate change. While some glaciers may have benefitted from additional snow during warmer temperatures, other valleys may have been in a mountain shadow and received less snow, resulting in MWP melting and glacier reduction.

Fig. 6. Time-distance diagram for glacier extent in Central Cumberland Peninsula. From Briner et al. (2009).

(10) Margreth et al. 2014 studied glaciers in Cumberland Peninsula, too. They found glacier reduction consistent with MWP warming (Fig. 7). In the abstract they write:

“Intensification of ice expansion between 1.9 and 1.1 ka [100 and 900 AD], followed by halt of ice growth, or ice recession during the Medieval Warm Period, and iii) renewed ice expansion after 0.8 ka [1200 AD], in response to cooling related to a combination of large volcanic eruptions and low solar activity. Overall, the observations support a model of near-instantaneous glacial response to regional climate controls and that these responses were synchronous throughout eastern Canadian Arctic and possibly eastern Greenland“.

Fig. 7: Probability distribution of glacier growth, based on vegetation ages buried by glacier ice. From Margreth et al. 2014.

We are now leaving Baffin Island and look at results from other parts of the Northeast Canadian Arctic.

(11) Rolland et al. 2009 studied ‘Lake 4′ on Southampton Island and reconstructed the temperature development based on the chironomid method. They found warm temperatures during the MWP (Fig. 8). From the abstract:

„Higher temperatures were recorded from cal yr AD 1160 to AD 1360, which may correspond to the Medieval Warm Period. Between cal yr AD 1360 and AD 1700, lower temperatures were probably related to a Little Ice Age event.“

Fig. 8. Chironomid-based August air temperatures of ‘Lake 4′ on Southampton Island. From: Rolland et al. 2009

(12) Adams & Finkelstein 2010 and Iamonaco 2011 studied a sediment core from ‘Lake SP02′ on Melville Peninsula and found good evidence for the MWP. Iamonaco 2011 wrote:

“Period of relative warmth between 1300-1000 yr BP, interpreted as evidence for the Medieval Warm Period”. “The main evidence for short-term climate shifts in the SP02 pollen record, which have been interpreted as possible evidence for the MWP, corresponds to pollen Zone 4b, and covers the period 1300 to 1000 yr BP (Figure 6). During this period there are increases in the relative abundances of long distance pollen representing Pinus and Picea, which may indicate increased prevalence of warmer air currents from areas to the south of the study site”

Adams & Finkelstein 2010:

„The … SP02 record shows that the whole of the past 2000 years should be considered in the context of the long-term cooling of the Neoglacial. There are some small changes in the biostratigraphy which could be indicative of warming coincident with the Medieval Warm Period, including a small increase in the abundance of the taxon P. pseudoconstruens, noted elsewhere to track warmer air and water temperatures.”

(13) ‘Lake JR01′ on Boothia Peninsula was investigated by LeBlanc et al. 2004 and Zabenskie & Gajeweski 2007. LeBlanc et al. 2004 found a warm phase 850-1400 AD which shows subtle changes in the floristic composition and increased production of diatoms. A cooling 1400-1850 AD can be inferred from the decrease in diatom composition, corresponding to the ‘Little Ice Age’. Cold water indicator species Fragilaria spp. sensu lato increased in abundance during this time, followed by a significant decline after 1850 AD.

Zabenskie & Gajweski 2007: A short warming, which could be interpreted as the Medieval Warm Period, occurred between [1100-1250 AD] (Fig. 9). The warming is indicated by small increases in the pollen percentages of Alnus, Betula, Salix and Cyperaceae and decreased pollen percentages of Pinus.

Fig. 9: Temperature reconstruction of Lake JR01, Boothia Peninsula. Thin black line represents average, grey lines maximum and minimum temperatures. From Zabenskie & Gajweski 2007

(14) Finkelstein & Gajewski 2007: Lake PW03, Prescott Island. Major biogenic silica peak centred around 1000 AD indicating high productivity associated with warmer temperatures 700-1300 AD. Cooling from 1300 AD onwards as evidenced by declines in warm-water indicators Navicula submuralis and Sellaphora seminulum, marking transition from warm Medieval Warm Period to cold Little Ice Age.

(15) Vare et al. 2009: Core ARC-3, Barrow Strait (Fig. 10). Study of sea ice biomarker IP25, isolated from a marine sediment core. From the abstract: “We also provide evidence for slightly lower and subsequently higher spring sea ice occurrence during the Mediaeval Warm Period and the Little Ice Age respectively.“

Fig. 10. Development of IP25 sea ice indicator in an offshore sediment core in Barrow Strait. From Vare et al. 2009.

(16) Paterson et al. 1977, Wolfe 2002: Ice Cores on the Devon Island ice-cap. Increase in delta18O during 1000-1400 AD marks warm period (Fig. 11).

Fig. 11: Oxygen isotope curve of ice cores on the Devon Island ice-cap. From Wolfe 2002.

(17) Mustaphi & Gajewski 2013: Lake DV09 auf Devon Island. Warm Medieval Climate Anomaly (1000–1300 AD) indicated by an interval of thick varves (Fig. 12).

Fig. 12: Varve thickness, Lake DV09, Devon Island, Nunavut Territory. Thick varves are interpreted as warm temperatures. From Mustaphi & Gajewski 2013.

Let us now move over to Greenland (Fig. 13).

Fig. 13: Location of MWP studies in western Greenland. Numbers refer to description in text. Red points: warm MWP. See interactive Google Maps Proejct here.

(18) Jennings et al. 2014: Core HU2008029-070CC, outer shelf in the Disko Trough system. This offshore sediment core is located half way between Baffin Island and west coast Greenland. Microfossil investigations yield clear evidence for a warm MWP with subsequent Little Ice Age (LIA) cooling (Fig. 14). According to the authors:

“Neoglacial cooling is inferred between 4000 and 1800 cal a BP. A warm interval that encompasses the Roman Warm Period and the Medieval Warm Period is found between 1800 and 700 cal a BP followed by a cooling after 700 cal a BP that relates in time to the LIA.“

Note limited sample density which may not allow resolving the Dark Ages Cold Period between Roman Warm Period and MWP.

Fig. 14. Temperature reconstruction based on microfossils, Disko Trough. From Jennings et al. 2014.

(19) Sha et al. 2014, Andresen et al. 2011: Marine sediment core DA06-139G, Vaigat Strait, Disko Bugt. Sha et al. 2014 studied diatoms and found a decreased sea ice concentration 900-1400 AD (Fig. 15). Sea ice was more common before (Dark Ages Cold Period) and after (Little Ice Age) this warm phase. Andresen et al. 2011 investigated lithology, dinoflagellate cysts and foraminifera and documented a warming phase 1000-1200 AD marked by increase in Atlantic warm water, increased iceberg calving (due to more unstable ice margin) and an increase in meltwater.

Fig. 15: Sea ice and warm water reconstruction based on diatoms from a sediment core in Vaigat Strait, Disko Bugt. From Sha et al. 2014.

(20) Lloyd et al. 2007: Core DA03, Disko Bugt.

Study for last 7000 years uses benthic foraminifera as climate proxy and finds a brief warm episode centred around 1000 AD.

(21) Krawczyk et al. 2010: Core DA00-02, Disko Bugt. This study yields a very interesting result and is a good reminder of the complexity of the climate system and respective proxies. The authors analysed diatoms which indicate colder surface water during the MWP and warmer temperatures during the LIA. At first sight this seems to contradict the MWP concept. At closer inspection, however, this result makes sense. The sediment core has been retrieved from a fjord location and the applied diatom method records surface temperatures. A warmer MWP climate leads to an intensified melting of glaciers which result in greater volumes of cold meltwater in the fjord. During subsequent colder (LIA) climate, the cold meltwater stream is reduced so that warmer Atlantic water leads to warming of the fjord surface layer. This is also corresponds to the interpretation favoured by Krawczyk and colleagues.

Btw: Similar anti-phase temperature relationships are also reported from Newfoundland. Please click on blue data points on the MWP map to access the respective summaries

(22) Ouellet-Bernier et al. 2014, Krawczyk et al. 2013, Perner et al. 2011, Ribeiro et al. 2012: Core MSM343310, outer Disko Bay. Several groups of scientists have studied this core using complementary paleontological techniques. As in neighbouring studies, the MWP is characterized by cooling surface waters (due to increased glacier meltwater) and warming bottom waters:

Ouellet-Bernier et al. 2014: From 1000-1200 AD summer sea surface temperatures (SST) increased to about 10°C, which is much higher than the present day summer SST of 4.4°C at the coring site. In the dinocyst assemblage, I. minutum percentages decreased, whereas the S. elongatus percentages increase. We suggest that this event corresponds to a warming of the West Greenland Current.

Krawczyk et al. 2013: Medieval Warm Period with cold sea surface water conditions, Little Ice Age (LIA) with warm sea surface water conditions. Anti-phase to usual North Atlantic trend. Anti-phases patterns could be linked to local hydrological conditions where increased meltwater flux generated during warm climatic phases cools local sea surface water layer.

Perner et al. 2011: During the time period of the ‘Medieval Climate Anomaly’ we observe only a slight warming of the West Greenland Current (WGC). A progressively more dominant cold water contribution from the East Greenland Current on the WGC is documented by the prominent rise in abundance of agglutinated Arctic water species from 1100 AD onwards. This cooling event culminates at 1700 AD and represents the coldest episode of the ‘Little Ice Age’.

Ribeiro et al. 2012: Warm phase 1050-1250 AD, as reflected by low sea ice concentration. Followed by increase in sea ice during Little Ice Age.

(23) Perner et al. 2013: Core MSM 343300, outer Disko Bay. Reconstruction based on benthic foraminifera shows slight warming between 600-1100 AD, linked to the ‘Medieval Climate Anomaly’ (Fig. 16). Severe cooling of the West Greenland Current after 1100 AD, with cold plateau starting 1300 AD.

Fig. 16: Climate reconstruction in outer Disko Bay based on benthic foraminifera (cooling down, warming up). From Perner et al. 2013.

(24) Sha et al. 2016: Sediment core GA306-GC4, Holsteinsborg Dyb basin off Kangerlussuaq. Warm phase 900 AD (start of dataset) til 1350 AD as evidenced by decreased sea ice concentration in study location (Fig. 17). Subsequent increase of sea ice towards Little Ice Age. Here the abstract of this important new study which appeared in January 2016 in the Quaternary Science Reviews:

Solar forcing as an important trigger for West Greenland sea-ice variability over the last millennium

Arctic sea ice represents an important component of the climate system, and the present reduction of sea ice in the Arctic is of major concern. Despite its importance, little is known about past changes in sea-ice cover and the underlying forcing mechanisms. Here, we use diatom assemblages from a marine sediment core collected from the West Greenland shelf to reconstruct changes in sea-ice cover over the last millennium. The proxy-based reconstruction demonstrates a generally strong link between changes in sea-ice cover and solar variability during the last millennium. Weaker (or stronger) solar forcing may result in the increase (or decrease) in sea-ice cover west of Greenland. In addition, model simulations show that variations in solar activity not only affect local sea-ice formation, but also control the sea-ice transport from the Arctic Ocean through a sea-ice–ocean–atmosphere feedback mechanism. The role of solar forcing, however, appears to have been more ambiguous during an interval around AD 1500, after the transition from the Medieval Climate Anomaly to the Little Ice Age, likely to be driven by a range of factors.

Fig. 17: Reconstruction of sea ice concentration in Holsteinsborg Dyb basin off Kangerlussuaq. From Sha et al. 2016.

(25) D’Andrea et al. 2011: Lake Sø and Lake E. Alkenone-based paleotemperatures (UK37). Warming from ca. 900-1150 AD, corresponding to Medieval Warm Period (Fig. 18, from Appendix of paper).

Fig. 18. Temperature reconstruction Lake Sø and Lake E. From D’Andrea et al. 2011.

(26) Willemse & Törnqvist 1999: Kangerlussuaq region. High amounts of organic matter 800-1100 AD indicating high productivity due to warmer temperatures. Subsequent abrupt reduction in organic matter marks cold conditions of Little Ice Age.

Summary

A detailed analysis of the regional literature in Arctic Northeast Canada and west coast Greenland West yields clear evidence for a warm MWP in the region. The claimed widespread cooling during the MWP by Young et al. 2015 is unsupported and unsustainable. It appears that the authors have generalized a local glacier anomaly from a restricted area on Baffin Island and erroneously interpreted it as a regional phenomenon. Notably, there may be alternative interpretations for local MWP glacier advances such as increased snowfall. Furthermore the review demonstrates that surface waters in fjords may have locally cooled during MWP times due to increased influx of cold melt water, displacing warmer Atlantic water. In contrast, fjord bottom waters often show MWP warming.

While the vast majority of the presented studies show a clearly developed warming during MWP times, the detailed correlation of onset, termination and cold interludes may differ. Differences may be caused by local climate variations, different age resolutions and time resolutions of the studies, a limited number of radiometric age dates which in addition may be subject to errors, as well as varying depositional rates and gaps in the sedimentary record which cause problems in the employed age models. Detailed correlations of the various studies are planned as part of the ‘Medieval Warm Period Mapping Project’, once the data screening and mapping phase has been completed.

It is unfortunate that the reviewers of Science Advances have missed the shortcomings of the Young et al. paper and have not demanded a more thorough regional integration of the results into the existing literature. It is even more unfortunate that Young et al. chose to promote their questionable regional interpretation in a widely distributed press release that on top is written in an inadequate sensational style. Historians of Viking research are advised to check the facts thoroughly before using the disputable conclusions proposed by Young and colleagues.

The author of this review contacted Nicolás Young on 11^th January 2016 to discuss the discrepancy between the regional literature and the study’s claim. As of 15 January 2016, no reply was received. In contrast, Gifford Miller who declared in the press release that the study was a “coup de grace on the Medieval Warm Period” answered promptly and engaged in a constructive discussion, for which he is thanked.