Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland

Clim. Past, 8, 325–335, 2012 www.clim-past.net/8/325/2012/ Climate doi:10.5194/cp-8-325-2012 of the Past © Author(s) 2012. CC Attribution 3.0 License. Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland R. Auchmann1 , S. Br¨onnimann1 , L. Breda2 , M. Buhler ¨ 1 , R. Spadin2 , and A. Stickler1 1 Oeschger Center and Institute of Geography, University of Bern, Bern, Switzerland 2 Institute for Atmospheric and Climate Science, ETH Z¨urich, Z¨urich, Switzerland Correspondence to: R. Auchmann ([email protected]) Received: 3 October 2011 – Published in Clim. Past Discuss.: 8 November 2011 Revised: 18 January 2012 – Accepted: 20 January 2012 – Published: 24 February 2012 Abstract. We analyze weather and climate during the 1 Introduction “Year without Summer” 1816 using sub-daily data from Geneva, Switzerland, representing one of the climatically One of the most severe climatic deviations of the past few most severely affected regions. The record includes twice hundred years in central Europe was the “Year Without Sum- daily measurements and observations of air temperature, mer” (YWS) 1816 (Stothers, 1984; Briffa and Jones, 1992; pressure, cloud cover, wind speed, and wind direction as well Harington, 1992; Robock, 1994, 2000). This event has been as daily measurements of precipitation. Comparing 1816 to a studied extensively both by historians (Skeen, 1981; Pfis- contemporary reference period (1799–1821) reveals that the ter, 1992, 1999) and climate scientists (Briffa and Jones, coldness of the summer of 1816 was most prominent in the 1992; Kington, 1992; Shindell et al., 2004) with respect to afternoon, with a shift of the entire distribution function of causes and consequences. The event is mostly related to temperature anomalies by 3–4 ◦ C. Early morning tempera- the 1815 eruption of Tambora in Indonesia, which injected ture anomalies show a smaller change for the mean, a sig- a huge amount of sulfur into the stratosphere that was capa- nificant decrease in the variability, and no changes in nega- ble of altering global climate (Stommel and Stommel, 1983; tive extremes. Analyzing cloudy and cloud-free conditions Stothers, 1984, 1999; Piervitali et al., 1997; Briffa et al., separately suggests that an increase in the number of cloudy 1998; Chenoweth, 2001; Stendel et al., 2005). In addition, days was to a significant extent responsible for these features. reduced solar activity related to the so-called Dalton mini- A daily weather type classification based on pressure, pres- mum might have played a role (Lean et al., 1995; Mann et sure tendency, and wind direction shows extremely anoma- al., 1998). lous frequencies in summer 1816, with only one day (com- The global scale cooling due to the Tambora eruption is es- pared to 20 in an average summer) classified as high-pressure timated to approximately 0.5 ◦ C. However, in Central Europe situation but a tripling of low-pressure situations. The after- where consequences were devastating both economically and noon temperature anomalies expected from only a change in socially (Hoyt, 1958; Stothers, 1999; Oppenheimer, 2003), weather types was much stronger negative in summer 1816 the cold anomalies were much larger (Trigo et al., 2009; than in any other year. For precipitation, our analysis shows Luterbacher et al., 2004), calling for additional or amplify- that the 80 % increase in summer precipitation compared to ing mechanisms. the reference period can be explained by 80 % increase in the frequency of precipitation, while no change could be found Most previous studies on the YWS in Europe addressed neither in the average intensity of precipitation nor in the the monthly or seasonal scale (e.g. Self et al., 1980; Trigo frequency distribution of extreme precipitation. In all, the et al., 2009), for which abundant information is available analysis shows that the regional circulation and local cloud from direct measurements as well as climate proxies and cover played a dominant role. It also shows that the sum- documentary data. Daily or even sub-daily data have much mer of 1816 was an example of extreme climate, not extreme more rarely been studied (Baron, 1992; Chenoweth, 2009). weather. This would be important, however, as sub-daily information might potentially give further insights into the underlying Published by Copernicus Publications on behalf of the European Geosciences Union. 326 R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland e au l at P ns ss ai i nt w S ou M ra Ju Geneva ve s lè a lp A .S Mt 0 N 100 km Fig. 1. Approximate location of the weather station of Geneva in 1799–1821 in a contemporary view (top left panel) and map (right panel) of the city. The bottom left panel shows the topography of the region (relief: Eidgen¨ossische Erziehungsdirektorenkonferenz). processes. For the assessment of changes in climate variabil- The paper is organized as follows. In Sect. 2, data (includ- ity and extremes, daily to sub-daily data resolution is indis- ing all aspects of homogeneity) and methods are explained. pensable (Brandsma and K¨onnen, 2006; Alexander, 2006). In Sect. 3 we present and discuss the results. Conclusions are In our study, temporal attribution of the summer 1816 cool- drawn in Sect. 4. ing (e.g. morning or afternoon cooling) as well as the analysis of a combination of parameters (e.g. cloud cover and temper- ature) cannot be conducted by using mean data. In contrast 2 Data and methods to measured values (i.e. the sub-daily data used in this study), they do not represent a state of the atmosphere. Both meth- 2.1 Meteorological measurements and observations ods may contribute to a better understanding of the structure from Geneva of the cooling and provide some new details into processes and mechanisms involved. One main restriction so far was The data from Geneva were digitized from printed sources data availability as many data series were simply not avail- (Journal de Gen`eve, Biblioth`eque britannique/universelle) able in their original form. reaching back to the 1780s. We digitized all elements (tem- In this paper we analyze the YWS 1816 in a recently perature, precipitation, pressure, clouds, wind) at the full digitized record of twice daily measurements performed in temporal resolution and converted the data to current units. Geneva, Switzerland, including air temperature, pressure, No further pressure reduction was necessary as it was already cloud cover, wind speed and direction, and daily precipita- reduced to constant 10◦ R. This is different to today’s stan- tion. Geneva is in the region with the largest negative tem- dard (0 ◦ C). However, this is irrelevant for our study as we perature anomaly (e.g. Pfister, 1992; Trigo et al., 2009) in only use standardized pressure anomalies and pressure ten- summer 1816. The aim of the paper is to analyze to what dency. We focus on the sub-period 1799–1821, which can extent the year without summer was characterized by ex- be considered internally homogeneous at least with respect treme climate or by extreme weather (i.e. whether changes to temperature (the homogenization of the full record is on- are largest in the central part of the frequency distribution or going). There were no reported changes in station operation near the extremes). Furthermore, by analyzing sub-daily data during that time. Measurements were made in the old botani- we hope to get more insights into the mechanisms responsi- cal garden, situated on the Bastion St-L´eger in the southwest ble for the YWS. Finally, the paper aims at identifying new of Geneva (Fig. 1). The old botanical garden had an effec- characteristics of the YWS that are testable in or comparable tive area of 1800 m2 (Sigrist and Bungener, 2008) and hence to a modeling framework. the influence of surrounding buildings can be assumed small (Pictet, 1822). Before 1799 observations were made in Gen- thod, after 1822 (until 1825) the observation site was situated Clim. Past, 8, 325–335, 2012 www.clim-past.net/8/325/2012/ R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 327 at the southeastern edge of the town. In 1826 it was relocated and the years 1809 to 1811 (unknown eruption in 1809, Cole- to the new botanical garden, on the northern shore of Lake Dai et al., 2009). The remaining 16 years were used as a ref- Geneva (Schuepp, 1961). erence against which we could compare the YWS 1816. The During the period 1799–1821, Marc-Auguste Pictet was focus in this paper is on the summer of 1816, although we responsible for collecting and publishing the observational also analyzed the summer of 1815 and the winters of 1815/16 data. There are twice daily measurements of air tempera- and 1816/17 (L. Breda, unpublished master thesis). ture and pressure and twice daily observations of cloud cover, First, temperature values were analyzed and compared to wind direction and (sometimes) speed. One observation was the reference because analyzing the occurrences of frost or performed at sunrise (most likely local astronomical sunrise) other environmentally relevant indicators requires an abso- with the aim of capturing the minimum temperature and one lute scale rather than anomalies. In a second step, a mean at 14:00 LT, which is close to the maximum. Because anoma- annual cycle for temperature was formed from the refer- lies from the seasonal cycle are computed later on, the ef- ence period by fitting a seasonal cycle consisting of the first fects of systematic errors related to the time of observation two harmonics. We then subtracted this annual cycle from are small. In addition to the twice daily reports, daily pre- both the reference period and the YWS period in order to cipitation measurements are also reported. Of course, twice study anomalies. Precipitation was analyzed in the form of daily observations cannot capture a complete diurnal cycle, absolute values (not anomalies) because of its skewed and in contrast to, e.g. hourly observations. According to our in- bounded distribution function and because the seasonal cy- formation, there were no changes in instruments, reporting, cle is not well defined. In addition, statistics of precipita- and location during the period 1799–1821. However, notes tion frequency (number of days with >0.1 mm) and 24-h pre- in the station history revealed a trend inhomogeneity in this cipitation intensity (i.e. the amount of precipitation on days period which was corrected. Available calibration informa- with precipitation >0.1 mm) were analyzed. For tempera- tion indicated a drift in temperature, which is a known error ture anomalies and 24-h precipitation intensity, we estimated that is caused by the chemical composition of the glass and probability density functions in order to address the frequen- has been studied in detail for the case of Hohenpeißenberg cies of weather extremes. Also, we assessed the dependence in Germany (see Winkler, 2009, and references therein). As of temperature anomalies upon cloud cover. Note that due the instrument in Geneva was of the same type as that used at to homogeneity reasons (see Sect. 2), all analyses involving Hohenpeißenberg and because the reported shift in the cali- cloud cover used a reduced base period 1812–1821 (without bration (0.6 ◦ C) also was very similar, we corrected it in the the volcanically perturbed years 1815–1817). same way as Winkler (2009) by decreasing temperature by Finally, in order to further analyze the mechanisms, we 0.1 ◦ C per year (from 1796–1801). From 1802 onwards we performed a simple weather type classification for sum- subtracted 0.6 ◦ C. The instrument type used was a mercury mer based on wind direction, pressure, and pressure ten- thermometer with an isolated bulb, divided in 80 parts in unit dency in order to address the effects of changing frequen- degree R´eaumur (Pictet, 1822). cies of weather types and changes within weather types. The Cloud cover observations were recorded qualitatively and weather types were defined based on all summers in the ref- noted verbally in a semi-standardized terminology. In a step- erence period. Then the same criteria were applied to the wise procedure, we categorized the cloud observations into summer of 1816. By retaining only the weather type infor- six cloudiness groups, from clear sky to fully covered. For mation for 1816, randomly sampling from the corresponding our analysis, we isolated days on which both the morning and weather types during the reference period, and then compar- noon observations could clearly be identified as either cloud ing back with the observed anomaly, we estimate the contri- free or fully covered, respectively. By analyzing the distribu- bution of changes in weather types to the anomalies in tem- tion of the cloudiness classes over time, we clearly found the perature, precipitation, or cloud cover. A Monte Carlo ap- cloud cover series to be inhomogeneous. Unreasonable cloud proach (n = 10 000) was applied to obtain a measure of vari- cover distributions and year-to-year variations appeared in ability associated with the sampling, assuming temporal in- the entire period 1799–1811 (which is partly also reflected dependence of the weather types from one day to the next. in the terminology used). However, the period 1812–1821 Although this assumption is certainly not valid, the effect was found to be more homogeneous and relatively reliable. on the result is considered small as other sampling strate- Therefore, any further analysis involving cloud cover used gies (e.g. distinguishing 1st and 2nd days in a sequence or only this (shorter) period. sampling 2-day periods) gave similar results. 2.2 Analysis methods The period 1799–1821 was chosen as a base period. From this period we removed the volcanically perturbed years, namely the years 1815 to 1817 (which are perturbed by the eruption of Tambora in April 1815, Bradley and Jones, 1995) www.clim-past.net/8/325/2012/ Clim. Past, 8, 325–335, 2012 328 R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 40 Reference 20 1816 30 Frequency (%) Temperature (°C) 10 20 0 10 1816 Reference mean -10 Reference std.dev. 0 Reference min and max -10 -5 0 5 10 -10 -5 0 5 10 Class boundaries (°C) Class boundaries (°C) 30 Fig. 3. Histograms showing the frequencies of occurrence (in percent) of temperature anomalies at sunrise (left panel) or 20 14:00 LT (right panel) in the summer months (June to August) of Temperature (°C) 1816 (black) and in the reference period (red patterned). 10 Both the early morning temperatures and afternoon tem- 0 peratures were clearly and consistently below the reference mean in summer 1816. However, the afternoon temperatures -10 were more anomalous than the early morning temperatures. Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec This appears more clearly in Fig. 3, which shows histograms Fig. 2. Time series of daily values of temperature at sunrise (top of the anomalies with respect to the reference period for the panel) and 14:00 LT (bottom panel) in the year 1816 (thin black reference period (red) and the YWS (black) for the sunrise solid line) as well as for the average of the reference period (red observation (left panel) and the 14:00 LT observation (right dashed line). The blue (solid bold) lines denote ±1 standard devia- panel) for the summer period (June–August). The sunrise tion from the mean; the green (dashed bold) lines give the minima temperature was, on average, about 1.8 ◦ C cooler than the and maxima for the reference period. Note that all annual cycles reference. Interestingly, the distribution is quite different. from the reference period were obtained from the statistics for each Contrary to what one might expect, negative extremes were calendar day. They were then smoothed by fitting the first two har- not more frequent in 1816 than in the reference, but positive monics of the annual cycle. The black dotted vertical line in the up- extremes were much less frequent in 1816. The distribution per panel indicates the last day of morning temperatures in spring is significantly narrower for 1816 than for the reference pe- below 0 ◦ C in 1816. Ticks on the x-axis indicate the mean last day- riod. In the early afternoon, the difference in the mean was of-year with morning temperatures below 0 ◦ C in the reference pe- riod (red dashed tick), ±1 standard deviation (blue solid bold ticks), much larger and amounts to about 3.8 ◦ C cooling for 1816 and the earliest and latest date with negative spring temperatures in relative to the reference period. In this case, the entire dis- the morning (green dashed bold ticks). The black dotted vertical tribution is shifted: cold extremes were more frequent, warm line in the lower panel indicates the date of the annual maximum extremes less frequent. From this analysis we conclude that temperature in 1816. the YWS was mainly an afternoon phenomenon. 3.2 Frequency of cloud-free days and temperature 3 Results and discussion anomalies 3.1 Morning and afternoon temperature and frequency Clouds might explain the different effect found in the sun- distributions rise and the 14:00 LT temperature. Cloud cover is expected to lead to an increase of the downwelling longwave radi- Daily temperatures of the years 1816 are shown in Fig. 2 ation during the nights and to a decrease of the incoming together with the mean annual cycle from the reference pe- shortwave radiation during the day. If nothing else changes, riod as well as the corresponding annual cycle of minima and cloud cover would thus lead to warmer conditions at sun- maxima. The last negative spring temperature in 1816 was rise and colder conditions at 14:00 LT. This is observed in observed at sunrise on 16 April. This corresponds almost ex- the (short) reference period (Table 1), where overcast nights actly to the mean date of the last negative spring temperature were about 2 ◦ C warmer than clear nights, and 14:00 LT tem- in the reference period. The annual maximum temperature in perature were about 6 ◦ C cooler, leading to a 8 ◦ C change in 1816 was 26.9 ◦ C (observed on the afternoon of 14 August). the difference between 4:00 LT and sunrise (which for sim- Although this is a low value for an annual maximum, one plicity we address as diurnal temperature range or DTR). even lower value was observed in the reference period. Clim. Past, 8, 325–335, 2012 www.clim-past.net/8/325/2012/ R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 329 400 Table 1. Effect of clouds on temperature. Averaged summer- time (June–August) temperature anomalies (◦ C) at sunrise and at 40 75 14:00 LT as well as their difference (diurnal temperature range, 30 DTR) for days with clear sky (both sunrise and 14:00 LT) or fully 300 precipitation amount (mm) overcast days (both sunrise and 14:00 LT) during a short reference number of precipitation days 30 period (1812–1814 and 1818–1821, i.e. the period with relatively number of days in 1816 50 reliable cloud observations) and in summer 1816. Row 5 shows the frequency (%) 20 200 average temperature anomaly for the sum of overcast and clear days 20 for 1816 assuming the mean values from the short reference in the first two rows and the number of fully overcast days (21) and fully 25 10 clear days (15) from 1816. Row 6 is the same but using the mean 10 100 values from the year 1816 itself. The lowest row shows the differ- ence between the two. All temperature anomalies are with respect to the long reference period. 0 0 0 0 1800 1810 1820 0 10 20 30 40 50 Class boundaries (mm/d) period sky sunrise 14:00 LT DTR short reference overcast 0.22 −5.06 −5.28 Fig. 4. Left panel: precipitation sums (green line, right scale) and number of rainy days (bars, left scale) in summer (June to August) short reference clear −1.77 1.04 2.81 from 1799–1821. The years that were excluded from the reference 1816 overcast −0.92 −6.96 −6.05 period are shown in light blue, 1816 in black. Right panel: his- 1816 clear −3.40 −1.47 1.93 togram showing the frequency of occurrence (left scale: percent, 1816 estimated overcast + clear −0.61 −2.52 −1.91 right scale: number of days for 1816) of precipitation amounts in from short the summer months (June to August) of 1816 (black) and in the reference reference period (blue patterned). Note that precipitation amounts 1816 observed overcast + clear −1.95 −4.67 −2.72 <0.1 mm were excluded. difference overcast + clear −1.34 −2.16 −0.82 only addresses cloud-free and overcast conditions and that the reference period is shorter. In fact, the summer of 1816 was anomalously cloudy. 3.3 Precipitation sum, frequency and intensity Not a single cloud-free day was observed in June 1816 and only 15 in the whole summer. This is much less than in the The total amount of rainfall was as much as 80 % higher in (short) reference period 1812–1821 (22 days). Conversely, summer 1816 compared to the reference period. Partitioning the number of fully covered days was much larger in sum- the precipitation sum into frequency and intensity (Fig. 4) mer 1816 (21 days) than in the reference period (11 days). reveals that this can be fully explained by an increase in the Similar as in the reference period, overcast days in sum- frequency of days with >0.1 mm of precipitation of 80 %. mer 1816 were ca. 5.5 ◦ C colder during the day and 2.5 ◦ C The mean rainfall intensity (8 mm per day with >0.1 mm warmer during the night compared to clear days (Table 1), precipitation) was the same in summer 1816 than for the av- again resulting in an 8 ◦ C effect on DTR. However, in sum- erage summer of the reference period (note that within the mer 1816 all temperature anomalies were much more nega- reference period, average summer precipitation intensities tive compared to the short reference period. Fully covered vary between 4 and 12 mm d−1 , hence there is a consider- summer days in 1816 were 1.1 ◦ C and 1.9 ◦ C colder than the able variation). Not only did the mean intensity not change, reference for sunrise and 14:00 LT, respectively. Clear sum- but we also find no evidence for changes in the distribution mer days were even 1.6 ◦ C and 2.5 ◦ C colder than in the ref- of precipitation intensities (Fig. 4, right panel). A χ 2 -test erence period, for sunrise and 14:00 LT, respectively. (4 classes with theoretical frequencies > 5) yields a p-value Can the temperature anomalies be explained by cloud of 0.96. The highest daily precipitation amount in summer cover alone? Table 1 (row 5) shows estimated average 1816 was 47.5 mm, which is slightly higher than the maxi- temperature anomalies for summer 1816 for sunrise and mum in the reference period (43 mm), but even higher val- 14:00 LT (for cloudy plus clear days) assuming mean val- ues occurred in the two excluded years 1810 and 1811. The ues for the corresponding categories in the short reference second highest amount (among 43 rainy days) was 27.2 mm. period. These numbers can then be compared to observed This rainfall amount corresponds to the 96-percentile of the anomalies (for cloudy plus clear days) in 1816. For the reference period, i.e. an excellent match. The fact that ex- 14:00 LT value, the estimated anomaly is −2.16 ◦ C com- treme precipitation events were not more frequent in 1816 pared to an observed −4.67 ◦ C. Hence, about 50 % of the than in “normal” summers contradicts to some extent the afternoon summer temperature anomalies in 1816 can be ex- contemporary Swiss newspaper reports, which often speak plained by a simple cloud metric. Note, however, that this about torrential rain falls and thunderstorms (see Bodenmann www.clim-past.net/8/325/2012/ Clim. Past, 8, 325–335, 2012 330 R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland et al., 2011, although the reports cited therein do not specifi- 4 anomalies of precip. (mm/day) or temperature (°C) cally refer to the region of Geneva). On the contrary to the wet summer 1816, a notably dry 3 summer appears in 1818 (Fig. 4). Both the total annual 2 precipitation sum and also the number of precipitation days were the lowest within the whole period 1799–1821. Van der 1 Schrier et al. (2007) use the self-calibrated Palmer Drought Severity Index (Wells et al., 2004) to identify wet and dry 0 seasons over the Alps. Negative index values (dry condi- tions) within a prolonged period of positive values (1816– -1 1825; wet conditions) for the NW-sector of the Greater Alpine region (including Geneva) for summer 1818 confirm -2 our results. Furthermore, reconstructed SLP anomalies from Luterbacher et al. (2002) for summer 1818 show, in large -3 contrast to the summer 1816, positive SLP anomalies (with -4 the center west of the northern British Isles) covering almost Stationary-SW High-N Low-N Rising-N Falling/Stat.-N Falling-SW Rising-NE Falling/Stat.-NE High-SW Low-SW Rising-SW Front High-Rest Low-Rest Rising-Rest Rest all Europe (Trigo et al., 2009). 3.4 Frequency of weather types For the 14:00 LT observation, we addressed the role of the mesoscale circulation by defining weather types and analyz- Fig. 5. Temperature anomalies (red) and precipitation (blue pat- ing whether those types that typically are cold and rainy were terned) averaged for each weather type for the summer months more frequent in 1816 than in other years. Ideally, the clas- (June to August) in the reference period. Note that for visualization sification should be chosen such that the classes that can be purposes, precipitation is plotted as anomalies from the seasonal interpreted synoptically and distinguished from each other mean whereas the sampling operates with absolute values. (i.e. large distance between classes in terms of mean tem- perature anomalies and precipitation) and at the same time have a small variability of these variables within each class. categories low, medium, or high pressure using the thresh- For forming classes, we confined ourselves to air pressure olds −1 and +0.75 s.d., respectively. The category medium and wind direction because they define the atmospheric cir- pressure was further subdivided into rising, stationary, and culation most directly. Cloud cover is seen as a dependent falling pressure using the thresholds ±0.2 s.d. This results variable and hence was not used to define weather types, es- in a clearly defined and small (41 cases) class representing pecially since the time series is not homogeneous. vigorous frontal passages as well as five large classes. Wind occurrences provided a classification of wind di- Each of these five classes could be subdivided into nine rection into nine classes (calm, N, NE, W, SE, S, SW, W, subclasses according to wind direction. However, this would NW). For pressure we used both the actual values as well as result in a large number of classes, some of which represent the tendency over the past 24 h to separate days into high, very similar situations. We therefore combined subclasses low, rising, falling, and stationary pressure (see below). The according to the following rules: no class can have less large annual variation in pressure variations made it nec- than 30 members; classes to be combined must be neighbor- essary to base the classification on standardized anomalies, ing (e.g. NW and W wind direction); and if several options where annual cycles of both the mean and the standard devi- were possible, the one that produced the smallest within- ation (s.d.) were computed as the first two harmonics of the class standard deviation for temperature was chosen. Due to corresponding data for each day-of-year from the reference orographic wind channeling (see Fig. 1), SW and NW winds period. Plotting the standardized data revealed an asymmet- prevailed in all 5 pressure classes. This allows for some syn- ric distribution for pressure, but not for the pressure tendency. optic interpretation. For instance, high pressure with SW Hence, we considered non-symmetric thresholds in the case winds can occur with W or NW gradient wind (and corre- of pressure. spondingly, temperatures are lower than for high pressure Cases with very low pressure or with very fast rising or situation with other wind directions). falling pressure are often associated with the passage of a The resulting classification for the reference period is cold front. These cases show distinct temperature and pre- shown in Table 2. Figure 5 shows the corresponding anoma- cipitation anomalies. We used thresholds of ±2.5 s.d. for the lies of temperature and precipitation in the reference period. pressure tendency and a threshold of −2.5 s.d. for pressure. The final classification had 16 classes, with sizes between 36 All days that crossed one of the three thresholds were consid- (“low pressure, rest”) and 222 (“high pressure, northerly ered as frontal passages. The other days were separated into wind”). Mean temperature anomalies for the classes ranged Clim. Past, 8, 325–335, 2012 www.clim-past.net/8/325/2012/ R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 331 Table 2. Weather type classification. Note that days classified as “fronts” could theoretically also fit in other categories, but were attributed to “front” (in practice this occurred only rarely). The fifth column indicates whether significant (p < 0.05) differences were found within this weather type between 1816 and the reference period according to a Wilcoxon test (y = yes, n = no, – = not enough cases). The sixth column indicates mean absolute frequencies of a weather type in the reference period for an average summer, JJA (annual averages). The last column indicates absolute frequencies of a weather type for summer 1816, JJA, in brackets: frequencies for the entire year 1816. Class p dp/dt wind Sig. abs. freq. in JJA (year) Ref. 1816 Front x < − 2.5 or: x < − 2.5 or x > 2.5 y 2 (7) 7 (12) High pressure, northerly wind x > 0.75 NW, N, NE – 13 (55) 1 (20) High pressure, southwesterly wind x > 0.75 SW – 3 (17) 0 (12) High pressure, rest x > 0.75 Rest – 3 (15) 0 (7) Low pressure, northerly wind −2.5 < x < −1 NW, N, NE y 4 (17) 8 (22) Low pressure, southwesterly wind −2.5 < x < −1 SW y 7 (29) 31 (62) Low pressure, rest −2.5 < x < −1 Rest – 2 (9) 0 (3) Rising pressure, northerly wind −1 < x < 0.75 x >0.2 N – 4 (11) 0 (4) Rising pressure, northeasterly wind −1 < x < 0.75 x >0.2 NE n 6 (23) 14 (41) Rising pressure, southwesterly wind −1 < x < 0.75 x >0.2 SW y 6 (26) 8 (30) Rising pressure, rest −1 < x < 0.75 x > 0.2 Rest – 3 (13) 4 (15) Falling or stationary pressure, northerly wind −1 < x < 0.75 x <0.2 N – 7 (18) 0 (9) Falling or stationary pressure, northeasterly wind −1 < x < 0.75 x < 0.2 NE y 8 (37) 14 (54) Stationary pressure, southwesterly wind −1 < x < 0.75 −0.2 < x < 0.2 SW – 4 (21) 2 (25) Falling pressure, southwesterly wind −1 < x < 0.75 x < −0.2 SW – 9 (34) 2 (35) Falling or stationary pressure, rest −1 < x < 0.75 x < 0.2 E, SE, S, W, NW, calm – 8 (31) 1 (15) from −4.0 ◦ C (“front”) to +2.4 ◦ C (“high pressure, rest”). 35 1799−1821 Precipitation anomalies range from +3.4 mm d−1 (“front”) to 30 1816 −1.8 mm d−1 (“high pressure, northerly wind”). 25 1981−2010 frequency (%) After the classification scheme was accepted, it was ap- 20 plied to the summer of 1816 and to the contemporary pe- 15 riod 1981–2010. A comparison of the obtained distribu- 10 tions of weather types in the reference and the period 1981– 5 2010 serves as a measure of usefulness of the classification methodology. Figure 6 shows the distribution of weather 0 Front High−N High−SW High−Rest Rising−N Rising−NE Rising−SW Rising−Rest Rest Low−N Low−SW Low−Rest Stationary−SW Falling/Stat.−N Falling/Stat.−NE Falling−SW types in summer 1816, in the reference period, and in the pe- riod 1981–2010 as a histogram. In general, markedly similar patterns are obtained in the periods 1799–1821 and 1981– 2010. However, the class “Rest” is more frequent in the period 1981–2010. We therefore claim that our classifica- tion methodology and the thereof derived weather types can Fig. 6. Histogram showing the frequency of occurrence (percent) be readily used for comparing local circulation patterns in of weather types in the summer months (June to August) of 1816 1816 to the reference period. The main differences between (black), in the reference period (red patterned), and in the period 1816 and the reference period are the almost complete ab- 1981–2010 (light grey). sence of the three high pressure situations in summer 1816 (with only one case), which in the reference period together account for 21 % of the days (corresponding to 20 days in an A very large center with negative SLP anomalies appears average summer). In contrast, the class “low pressure, south- over northern France. Geneva (dot in Fig. 7) was to the south westerly winds” was 3.5 times as frequent in 1816 as in the of this center. This is expected from an increase in low pres- reference period. Hence, the “year without summer” of 1816 sure situations and in situations with westerly geostrophic in Geneva can be explained by frequency shifts of only few flow (which in Geneva could be channeled to southwesterly weather types: an absence of high pressure situation and a wind). In addition Fig. 7b shows the reconstructed absolute tripling of low-pressure situations. SLP values for June-to-August. Seasonal SLP reconstructions from K¨uttel et al. (2010) are Casty et al. (2005b) investigated large scale climate consistent with this result. Figure 7 shows the corresponding regimes over the North Atlantic/European sector since 1766. anomalies for June-to-August with respect to the reference. They show that the period 1805–1825, which overlaps to www.clim-past.net/8/325/2012/ Clim. Past, 8, 325–335, 2012 332 R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 10 000 repetitions) together with the actual values for sum- mer 1816 as well as the mean values for summers in the ref- 12 10 erence period. While for temperature, the distribution is sym- metric, it is slightly asymmetric for precipitation, indicating that the uncertainty is higher towards the high precipitation sums. 10 Out of 10 000 estimated 14:00 LT seasonal mean tempera- 1012 10 10 16 14 18 ture anomalies, 99 % of the values are below zero (and hence below the reference period). The mean value is −1.00 ◦ C. Hence, there is no doubt that unusual weather types affected 1022 1024 temperature in Geneva. However, the expected contribution of the weather types to the actually observed anomaly is only about one fourth. In fact, the actually observed anomaly is 100 20 well below the range of the Monte Carlo simulations. One 10 0 might be tempted to attribute the remainder to direct volcanic 0 1004 1002 08 06 101 10 10 effects. However, there might also be other effects that are −4 −3 −2 −1 0 1 2 not captured by the classification. The weather types capture SLP anomaly (hPa) an (unknown) mesoscale and not necessarily large-scale or small-scale effects. Fig. 7. SLP anomaly for summer (June to August) 1816 with re- To test this, we applied the same procedure (but now sam- spect to the reference period and absolute SLP values (white con- pling only the mean values of a class) to all years in the ref- tours). The dot marks the location of Geneva. erence period (during which no specific forcing was operat- ing) and also to the summers of 1809, 1810, 1811, 1815, and 1817. Results are shown in Fig. 9. For temperature, there a great extent the reference period, is characterized by an is a clear correlation between the artificial and the observed anomalously westerly flow regime. However, the recon- temperature anomalies that reaches 0.72 when including all structed parameters are confined to winter (December-to- years. This indicates that the not-captured part of the cir- February). According to Casty et al. (2005a) analysis of culation effect on temperature correlates well with the cap- Alpine temperatures, the cold summer 1816 was the coldest tured part. In other words, about 50 % (in terms of vari- of a series of below-average summers, which followed a pe- ance) or 25 % (in terms of magnitude) of summer temper- riod with generally warm summers between 1780 and 1810 ature anomalies can be explained by changing frequencies (the warmest being 1807). While winter climate anomaly of weather types. The summer of 1816 lies close to a re- patterns can be analyzed in the framework of large scale cir- gression line drawn through the data points in the reference culation variability, this is more difficult for the summer sea- period (not shown). From this analysis we conclude that the sons both because of weaker relations between climate and temperature anomaly in Geneva in 1816 is not inconsistent circulation indices and the lower reconstruction skill of cir- with only a change in atmospheric circulation, although the culation indices. In fact, the anomaly pattern of SLP (Fig. 7) uncertainty of the analysis is larger, leaving room for sub- does not project strongly onto the NAO pattern but rather re- stantial non circulation-related cooling. At the same time, it sembles the East Atlantic pattern. Whether this pattern is is important to note that the temperature anomaly predicted affected by global volcanic forcing remains to be explored. only from the weather types is more than twice as negative As the winter 1815/1816 was likely an El Ni˜no winter, re- as in any other case in the reference period. mote effects from the tropical Pacific could be considered. We also analyzed whether or not there are significant dif- There is a tendency for an excitement of such a pattern ac- ferences between temperature anomalies in 1816 and the ref- companied by increased precipitation over central Europe in erence period within the same weather type. We performed springs following El Ni˜no winters, which was also found in the Wilcoxon rank sum test for all classes for which at least climate reconstructions (Br¨onnimann et al., 2007). However, seven days were available in 1816. With one exception (“Ris- the relation between El Ni˜no and European climate is weak. ing pressure, northeasterly wind”), all classes exhibited a sig- nificantly lower mean temperature in 1816 compared to the 3.5 Contribution of weather types to temperature reference period (Table 2). This analysis again suggests that anomalies not all of the cooling can be captured by the weather type frequencies, which can be due to an inadequate classification How much of the afternoon cooling in summer 1816 can or a “missing” mechanism. be reproduced from sampling corresponding weather types in the reference period? Figure 8 shows estimated tem- peratures anomalies and precipitation sums (histograms of Clim. Past, 8, 325–335, 2012 www.clim-past.net/8/325/2012/ R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 333 probability density (1/°C) probability density (1/mm) 0.8 0.006 0.6 0.004 0.4 0.002 0.2 0.0 0.000 -4 -3 -2 -1 0 0 100 200 300 400 500 Temperature anomaly (°C) Precipitation sum (mm) Fig. 8. Histogram showing 10 000 artificial temperature anomaly averages (left panel) and precipitation sums (right panel) for the summer of 1816 based on weather types for 1816 and sampling from these weather types in the reference period. The dotted and solid lines denote the observed values for the year 1816 and the reference period, respectively. 350 rref = 0.53; rref, sp = 0.50 obs. precipitation sum (mm) 3 rref = 0.63; rref, sp = 0.49 obs. temperature anomaly (°C) 2 rall = 0.72; rall, sp = 0.51 rall = 0.48; rall, sp = 0.46 300 1 250 0 -1 200 -2 150 -3 -4 100 -1.0 -0.5 0.0 0.5 150 200 250 predicted temperature anomaly (°C) predicted precipitation sum (mm) Fig. 9. Scatter plot showing the observed temperature anomalies (left panel) and precipitation sums (right panel) as a function of their predicted values based only on the weather types information. The years in the reference period are indicated in red and blue (points), for temperature and precipitation, respectively, the years that were excluded from the reference period are shown in grey (triangles), 1816 in black (stars). Correlation (r) and non parametric Spearman correlation (rsp ) coefficients are shown for only the reference period (subscript ref, red number) and all years (subscript all, black number). 3.6 Contribution of weather types to precipitation types far exceeds any other value in the 1799–1821 period. anomalies The Wilcoxon test for within-class differences revealed sig- nificant differences only for one class (“front”). Hence, we The same analyses were performed with respect to precipi- have no evidence to falsify the hypothesis that the precipita- tation. Of the 10 000 estimated precipitation sums for sum- tion anomalies are caused by atmospheric circulation. mer 1816 (Fig. 8), 94 % are higher than the reference pe- riod mean. Hence, there are indications that unusual weather 3.7 Contribution of weather types to cloud cover types contributed to the precipitation surplus in Geneva in anomalies summer 1816, but the uncertainty is much larger than in the case of temperature. The uncertainty range includes the ob- We used the same methodology, albeit with the shorter served value. calibration period, to estimate the fraction of cloud-free Applying the procedure to all years (Fig. 9), we also find days (not shown). This variable could be well reproduced correlations between observed and predicted precipitation, from weather types (estimated for 1816: 19.3 %, observed: but compared to the analysis for temperature correlations 16.3 %, reference period: 25.2 %). According to the Monte these are much lower and only significant for the reference Carlo simulation, the estimated fraction deviates signifi- period (not for the full period). Only 25 % of the interan- cantly (p < 0.01) from the reference but not from the ob- nual precipitation variability can be explained by changed served value. weather types. The summer of 1816 falls within the relation found in the reference period and hence is not inconsistent with only a change in atmospheric circulation. However, as for temperature, it is important to note that the precipitation amount predicted for summer 1816 only from the weather www.clim-past.net/8/325/2012/ Clim. Past, 8, 325–335, 2012 334 R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 4 Discussion and conclusions to persistent regional circulation and cloud cover anomalies over Western Europe in summer. We have digitized sub-daily meteorological data from the sta- Supplementary material related to this tion Geneva in order to analyze the “Year Without Summer” article is available online at: of 1816. We chose a sub-period of the record which was http://www.clim-past.net/8/325/2012/ found to be homogeneous after the correction of one (known) cp-8-325-2012-supplement.zip. error. Comparing the YWS 1816 to the remainder of that pe- riod (1799–1821) reveals that the coldness of the summer 1816 was most prominent in the afternoon. The entire tem- Acknowledgements. The project was supported by the Swiss perature distribution function was shifted by almost 4 ◦ C to- National Science Foundation via the NCCR Climate project wards lower values with no apparent change in variability. In PALVAREX, by the cogito foundation, by the Department of contrast, early morning temperatures show a significant de- Environmental Sciences, ETH Zurich and by the COST Action HOME. Data for this study were digitized within the framework of crease in the variability, but no changes in negative extremes the cooperative DigiHom project of MeteoSwiss and ETH Zurich. and a smaller change for the mean. We would like to thank Michael Begert, Christine F¨ullemann, and Possible causes for changes between 1816 and the ref- Mischa Croci-Maspoli (all MeteoSwiss) for providing metadata erence period include direct radiative changes (due to vol- and helpful discussions. Thanks also to St´ephane Fischer (Mus´ee canic aerosols or clouds), changes in the mesoscale circu- d’Histoire des Sciences, Geneva) for information on historic lation (e.g. the frequency of weather types), and changes in instruments. the large-scale climate system (e.g. a cooling of the oceans). The mechanisms are not independent and hence hardly sep- Edited by: P. Brohan arable. Although an attribution of the causes is not possible using only observational data, they still provide interesting References insights which, if well characterized, can be used in model comparisons. Alexander, L. V.: Global observed changes in daily climate ex- Analyzing cloud information suggests that an increase in tremes of temperature and precipitation, J. Geophys. Res., 111, cloud cover was mainly responsible for time-of-day charac- D05109, doi:10.1029/2005JD006290, 2006. teristics of the temperature change. But what changed the Baron, W. R.: 1816 in perspective: the view from the Northeastern clouds? A simple weather type classification (according to United States, in: The year without a summer?: world climate in 1816, edited by: Harrington, C. R., Canadian Museum of Nature, pressure, pressure tendency, and wind direction) well ex- Ottawa, 124–144, 1992. plains the changes in cloud cover. About one fourth of the Bodenmann, T., Hirsch Hadorn, G., Br¨onnimann, S., Kr¨uger, T., deviation in the seasonal mean afternoon temperature can be and Weissert, H.: Perceiving, understanding, and observing cli- explained by changes in weather type (i.e. the mesoscale cir- matic changes: An historical case study of the “year without culation). Although this may seem a small fraction, it cor- summer” 1816, Meteorol. Z., 20, 577–587, doi:10.1127/0941- responds exactly to the expected fraction of explained vari- 2948/2011/0288, 2011. ance for any anomalous year. Importantly, the amount of Bradley, R. and Jones, P. D.: Records of explosive volcanic erup- weather-type-related cooling was much larger than in other tions over the last 500 years, in: Climate Since A.D. 1500, edited years and the within-type differences of afternoon temper- by: Bradley, R. S. and Jones, P. D., Routledge, London, 1995. ature were significant in most cases. For precipitation, our Brandsma, T. and K¨onnen, G. P.: Application of nearest analysis shows that the 80 % increase in summer precipita- neighbor resampling for homogenizing temperature records on a daily to sub-daily level, Int. J. Climatol., 26, 75–89, tion can be explained solely by a higher frequency of precip- doi:10.1002/joc.1236, 2006. itation, while no change could be found in the 24-h intensity Briffa, K. and Jones, P. D.: the climate of Europe during the 1810s of precipitation (including the occurrence of extremes). with special reference to 1816, in: The year without a summer?: In all, the results show that the year without summer world climate in 1816, edited by: Harrington, C. R., Canadian was not characterized by extreme weather (the tails of the Museum of Nature, Ottawa, 372–391, 1992. distributions did not change much except for cold after- Briffa, K., Jones, P. D., Schweingruber, F., and Osborn, T.: In- noons), but extreme climate (i.e. the statistics of weather fluence of volcanic eruptions on Northern Hemisphere Summer types changed). The analysis shows that the local-to-regional temperature over the Past 600 Years, Nature, 393, 450–455, 1998. circulation plays a dominant role in that all climatic anoma- Br¨onnimann, S., Xoplaki, E., Casty, C., Pauling, A., and Luter- lies are statistically consistent with only a change in circula- bacher, J.: ENSO influence on Europe during the last centuries, tion. Historical analyses show that in Switzerland, strong Clim. Dynam., 28, 181–197, doi:10.1007/s00382-006-0175-z, tropical volcanic eruptions are often followed by a “Year 2007. Without Summer” with very similar weather as in the sum- Casty, C., Handorf, D., and Sempf, M.: Combined climate win- mer of 1816 (Pfister, 1999, see also Fischer et al., 2007), ter regimes over the North Atlantic/European sector 1766–2000, suggesting a mechanism linking tropical volcanic forcing Geophys. Res. Lett., 32, 13, doi:10.1029/2005GL022431, 2005a. Clim. Past, 8, 325–335, 2012 www.clim-past.net/8/325/2012/ R. Auchmann et al.: Extreme climate, not extreme weather: the summer of 1816 in Geneva, Switzerland 335 Casty, C., Wanner, H., Luterbacher, J., Esper, J., and B¨ohm, Pfister, C.: Wetternachhersage: 500 Jahre Klimavariationen und R.: Temperature and precipitation variability in the Euro- Naturkatastrophen (1496–1995), Paul Haupt, Bern, 1999. pean Alps since 1500, Int. J. Climatol., 25, 1855–1880, Pictet, M. A.: Biblioth`eque universelle des sciences, belles-lettres doi:10.1002/joc.1216, 2005b. et arts Sciences et arts, Imprimerie de a biblioth`eque universelle Chenoweth, M.: Two major volcanic cooling episodes derived from ed, Gen`eve, 1822. global marine air temperatures, Geophys. Res. Lett., 28, 2963– Piervitali, E., Conte, M., and Colacino, M.: Summer air tempera- 2966, 2001. ture anomalies in europe during the century 1811–1910, Il Nuovo Chenoweth, M.: Daily synoptic weather map analysis of the New Cimento, C20, 195–208, 1997. England cold wave and snowstorms of 5 to 11 June 1816, in: His- Robock, A.: Review of year without a summer? World climate in torical climate variability and impacts in North America, edited 1816, Climatic Change, 26, 105–108, 1994. by: Dupigny-Giroux, L.-A. and Mock, C. J., Springer, Dor- Robock, A.: Volcanic eruptions and climate, Rev. Geophys., 38, drecht, 2009. 191–219, 2000. Cole-Dai, J., Ferris, D., Lanciki, A., Savarino, J., Ba- Schuepp, M.: Lufttemperatur, Beiheft zu den Annalen der SMZ, roni, M., and Thiemens, M. H.: Cold decade (AD 1810– Fotorotar AG, Z¨urich, 1961. 1819) caused by Tambora (1815) and another (1809) strato- Self, S., Rampino, R. R., and Barbera, J. J.: The possible effects spheric volcanic eruption, Geophys. Res. Lett., 36, L22703, of large 19th and 20th century volcanic eruptions on zonal and doi:10.1029/2009GL040882, 2009. hemispheric surface temperatures, J. Volcanol. Geoth.-Res., 11, Fischer, E., Luterbacher, J., Zorita, E., Tett, S., Casty, C., and Wan- 41–60, 1980. ner, H.: European climate response to tropical volcanic eruptions Shindell, D., Schmidt, G., Mann, M., and Faluvegi, G.: Dy- over the last half millennium, Geophys. Res. Lett., 34, L05707, namic winter climate response to large tropical volcanic erup- doi:10.1029/2006GL027992, 2007. tions since 1600, J. Geophys. Res.-Atmos., 109, D05104, Harrington, C. R.: The year without a summer?: world climate in doi:10.1029/2003JD004151, 2004. 1816, Ottawa: Canadian Museum of Nature, 1992. Sigrist, R. and Bungener, P.: The first botanical gardens in geneva Hoyt, J.: The cold summer of 1816, Ann. Assoc. Am. Geogr., 48, (c. 1750–1830): private initiative leading science, St. Hist. Gard. 118–131, 1958. Des. L., 28, 333–350, 2008. Kington, J.: Weather patterns over Europe in 1816, in: The year Skeen, E. K.: The Year without a Summer: A Historical View, J. without a summer?: world climate in 1816, edited by: Harring- Early Republ., 1, 51–67, 1981. ton, C. R., Canadian Museum of Nature, Ottawa, 358–371, 1992. Stendel, M., Mogensen, I., and Christensen, J.: Influence of various K¨uttel, M., Xoplaki, E., Gallego, D., Luterbacher, J., Garc´ıa- forcings on global climate in historical times using a coupled Herrera, R., Allan, R., Barriendos, M., Jones, P. D., Wheeler, D., atmosphere-ocean general circulation model, Clim. Dynam., 26, and Wanner, H.: The importance of ship log data: reconstructing 1–15, 2005. North Atlantic, European and Mediterranean sea level pressure Stommel, H. and Stommel, E.: Volcano Weather: The Story of fields back to 1750, Clim. Dynam., 34, 1115–1128, 2010. 1816, The Year without a Summer, Seven Seas Press ed., New- Lean, J., Beer, J., and Bradley, R.: Reconstruction of solar irradi- port, Rhode Island, 1983. ance since 1610: Implications for climate change, Geophys. Res. Stothers, R.: The great Tambora eruption in 1815 and its aftermath, Lett., 22, 3195–3198, 1995. Science, 224, 1191–1198, 1984. Luterbacher, J., Xoplaki, E., Dietrich, D., Rickli, R., Jacobeit, J., Stothers, R.: Volcanic dry fogs, climate cooling and plague pan- Beck, C., Gyalistras, D., Schmutz, C., and Wanner, H.: Recon- demics in europe and the middle east, Climatic Change, 42, 713– struction of sea-level pressure fields over the eastern North At- 723, 1999. lantic and Europe back to 1500, Clim. Dynam., 18, 545–561, Trigo, R. M., Vaquero, J. M., Alcoforado, M. J., Barriendos, M., doi:10.1007/s00382-001-0196-6, 2002. Taborda, J., Garcia-Herera, R., and Luterbacher, J.: Iberia in Luterbacher, J., Dietrich, D., Xoplaki, E., Grosjean, M., and Wan- 1816, the year without a summer, Int. J. Climatol., 29, 99–115, ner, H.: European seasonal and annual temperature variability, 2009. trends and extremes since 1500 a.d., Science, 303, 1499–1503, van der Schrier, G., Efthymiadis, D., Briffa, K. R., and Jones, P. 2004. D.: European Alpine moisture variability for 1800–2003, Int. J. Mann, M., Bradley, R., and Hughes, M.: Global-scale temperature Climatol, 27, 415–427, doi:10.1002/joc.1411, 2007. patterns and climate forcing over the past six centuries, Nature, Wells, N., Goddard, S., and Hayes, M. J.: A self-calibrating 392, 779–787, 1998. Palmer Drought Severity Index, J. Climate, 17, 2335–2351, Oppenheimer, C.: Climatic, environmental and human conse- doi:10.1175/1520-0442(2004)017<2335:ASPDSI>2.0.CO;2, quences of the largest known historic eruption: Tambora volcano 2004. (Indonesia) 1815, Prog. Phys. Geog., 27, 230–259, 2003. Winkler, P.: Revision and necessary correction of the long-term Pfister, C.: The Years without a Summer in Switzerland: 1628 and temperature series of Hohenpeissenberg, 1781–2006, Theor. 1816, in: The year without a summer?: world climate in 1816, Appl. Climatol., 98, 259–268, 2009. edited by: Harrington, C. R., Canadian Museum of Nature, Ot- tawa, 416–417, 1992. www.clim-past.net/8/325/2012/ Clim. Past, 8, 325–335, 2012