RHODORA, No. 118-135, 2003 Vol. 105, 922, pp. FROM NORTHERN LATE-FLOWERING PLANTS NOVA CANADA SCOTIA, Barry Taylor^ and David Garbary R. J. Department of Biology, Francis Xavier University, St. Nova B2G 2W5, Canada Antigonish, Scotia, [email protected] 'e-mail: abstract. Over 90 species of herbaceous dicotyledonous plants, representing Nova 62 genera and 19 families, were recorded in blossom from northern Scotia, Canada, during November and December 2001. Most observations were made during 10-20 November, but 29 species were in flower after 10 December The number of species flowering declined steadily with time, rather than declining precipitously after Nova the first frost. Only 6 of these 92 species are specificafly reported in the floras of Scotia as flowering in November. Examination of plant collections in the primary Nova showed Scotia herbaria that our collections represent the latest flowering Nova records for all but one of these species in Scotia. Accordingly, our collection known extends the flowering times for these species by an average of 45 days. For most species the late flowering interpreted as an extension of normal phenology is warm autumn some in response to unusually weather, but for species Viola (e.g., cucullata. Ranunculus acris, Fragatia virginiana) appears to represent a second it Our flowering period. collections establish a base line of late flowering times that could be used as an indicator of regional climate change. Key Words: climate change, flowering, global warming, Nova phenology Scotia, The reproductive period of flowering plants determined by external is environmental constraints (both biotic and abiotic) and physiological adaptations to those constraints (Daubenmire 1974; Zeevaart 1962). In temperate zones, the growing and flowering season abruptly is terminated autumn by waning in sunlight intensity and falling temper- atures, especially those that produce frost. Given that temperatures, and hence the length of the growing season, vary from year to year, is it natural to expect that the flowering periods of at least some plant species be enough many will flexible to reflect those differences; late-flowering grow species continue to unUl they are killed by Hence, one frost. effect warmer of temperatures or late frost in a given year would be an ex- tension of the flowering season (Salisbury and Ross 1992). Flexibility in plant flowering periods is pertinent to the question of global warming, which is anticipated to increase mean annual tem- 2^' attendant Autumn of 2001 provided a glimpse of what might the future hold; the weather was exceptionally moderate from September through 118 — Nova 2003] Taylor and Garbary ^Late Flowering in Scotia 119 December, few with only a (and mild) frosts late into the season, warm accompanied by temperatures and gentle rainfall. If plants can adapt to longer growing seasons associated with transient or directional we changes in their environment, predict that reproductive periods would be extended concert with a lengthening in frost-free period. we document Here, the apparently extended periods of flowering Nova in the flora of Antigonish County, on the northern mainland of Scotia, Canada. To confirm that the season was unusually warm, we compared weather records for autumn 2001 against climate normals Our (long-term averages) for the area. data provide the latest records of Nova most flowering for of these species in Scotia, and thereby establish which a base line against extensions or contractions of the flowering season in future years can be compared. Our observations began October when in casual observations of meadows many roadsides and revealed late-summer and that plants fall were unexpectedly still in flower. S>noptic observations of the day -to- We day weather suggested that the autumn was unusually warm. began keeping records of species in which flowering persisted as the autumn progressed. From November onward we began early a concerted effort to observe and collect any flowering plants in blossom. These still observations continued until flowering in native and naturalized afl species in the local flora had ceased, in mid-December. MATERIALS AND METHODS Regional climate. Antigonish County occupies the northeast end of mainland Nova Scotia, on the eastern edge of Canada (Figure As 1). Nova Scotia a peninsula protruding into the Atlantic Ocean, there is is a strong maritime influence on the continental climate of the province. mean Annual temperature inland in Antigonish County 5.6'^C is mean 18.rc for 1916-1990; the in July and -6.7°C in January, is the greatest range for any part of the province. The growing season is Museum about 200 days (Nova Scotia 1996). Because of the maritime influence there a powerful moderation of is compared temperatures along the coast with sites inland. In addition, seasonal changes are delayed and extended compared with regions at the A warm same Canada. long autumn, by latitude in central interrupted moving up occasional storms the Atlantic coast, typical of northern is Museum Nova Scotia (Nova Scotia 1996). Nighttime frosts begin inland may in early October in most years but be delayed by two to three weeks Snow along the coast. does not begin to accumulate until late December. 120 Rhodora 105 [Vo!. KM 25 Figure 1. Anti (star) stations (*) at CollegeviUe (C) near Lochaber Lake (LL) and Town Point (T) and ^ = = primary collecting Antigonish Town; sites: star Malignant Cove; 2 North 1 ^ = = Grant; 3 =^ roadside, Highway 337; 4 Antigonish Landing; 5 Dunmore Road; 6 = Andrews; St. 7 Frasers Mills. Town The of Anti The local climate of the town warmer therefore is The landscap The varying leys. elevations and id uses and vegetation types farmland) many creates micro throughout the county. exam It is t; experience autumn frosts earlier than upland sites because of the accumulation of cooler air in depressions. We obtained weather data from two standard Environment Canada weather stations Antigonish and at CollegeviUe. Daily records of minimum maximum and temperatures two at the months sites for the — 2003] Taylor and Garbary Late Flowering Nova in Scotia 121 Summary Table 1. of daily temperature data (°C) from a coastal station (Antigonish) and an inland station (CoIIegeville) northern Nova autumn in Scotia, Normals are 85-year averages (1916-1990) for CoIIegeville. 'Daily tem- maximum perature average of (Max.) and minimum is (Min.) for each day. -^Days < minimum with temperature 0°C. "^Units are desree-days. Mean Temperature °C Heat Units'^ Month p^^^^ > < OT Max. Site Min. Daily' no. ofdays^ 5°C September Antigonish 22.8 10.5 16.7 349.8 CoIIegeville 22.8 9.1 15.9 2 328.3 - Normal 18.8 7.8 13.3 249.2 October Antigonish 17.2 5.0 11.1 4 192.1 CoIIegeville 16.1 3.9 10.0 5 163.8 - Normal 13.0 3.3 8.2 111.9 0.3 November Antigonish 8.7 0.4 4.6 17 5 5.5 .2 1 CoIIegeville 8.8 0.3 4.5 14 51.6 13.5 - Normal -0.9 6.9 3.0 30.8 19.8 December -2.6 Antigonish 3.6 0.5 25 12.5 31.7 -3.5 -0.5 CoIIegeville 2.6 27 8.5 58.5 - Normal -7.5 -3.2 1.0 5.1 130.5 of September through December 2001, along with and precipitation qualitative observations of sky conditions, were provided by the Atlantic New Climate Centre of Environment Canada (Fredericton, Brunswick). The Antigonish station located in South Side Harbour, just outside is Town m) the of Antigonish (45°37'N, 61°54'W; 30 and measures the The coastal microclimate. station CoIIegeville (45°29'N, 62''0rW; 76 at km m) located approximately 25 inland (Figure is 1). We used climate normals from the CoIIegeville station, based on weather records from 1916-1990, as a base line against which to compare the weather in autumn 2001 (Table Climate normals 1). for CoIIegeville were downloaded from the webpage of Meteoro- the logical Service of Canada, Environment Canada (www.msc-smc.ec.gc. ca/climate/index_e.cfm). The weather station Antigonish too at recent is (1990 onward) for climate normals to be established. We Plant collecting. )wn of Antigonish. Thes 122 Rhodora 105 [Vol. collections in rural areas of Antigonish County, in particular along the valley of South River and along the roadside Malignant Cove on to Northumberland Most the (Figure species were found waste Strait 1). in ground mixed herbaceous however, were in vegetation; collections made from roadsides, agricultural land, regenerating woodlots, pasture and mixed While we land, forest. did not undertake a systematic survey we of every habitat in the county did make a concerted effort to seek out a variety of habitats and to thoroughly explore places where flowering were most plants persistent. We collected only non-graminoid, herbaceous plants with conspic- uous some flowers. Grasses and herbaceous species with cryptic flowers were excluded because proved difficult to confirm whether they were it No fully in flower at the time of collection. woody plants flower into fall in this region. was Collecting November terminated initially after 11 because of a snowfall late in the day, followed by several days of heavy and frost Minimum subzero temperatures. temperatures 12-15 November for -2 ranged from -6^C -3 to at Antigonish and from -6.5°C Col- to at legeville. Despite apparently many this "killing" frost, plants (17 species) were found in flower at a diversity of November, sites after 15 the after snow we melted. Therefore resumed collecting, and continued long for as as flowering plants could be found. The day was last collecting 15 December, when permanent snow cover began to accumulate. All species flower were in noted during each sampling expedition, along with their habitat and general At condition. one specimen of least each species flower was in collected, pressed, and mounted. For some species, or more later conspicuously flowering specimens were also mounted. In the species (Appendix), we when list have noted dates were plants observed in flower when after the date a specimen was col- Francis (STFX) Roland (1969) and Zinck and (1998) collections in the following herbaria Herbarium Roland Herbarium RESULTS Weather confirm warm Based on monthly 1). averages for September through December, mean — 2003] Taylor and Garbary Late Flowering Nova 123 in Scotia maximum, minimum, and daily temperatures CoUegeville at for those months 2001 were warmer in significantly than the long-term average < ^ P minimum (t-tests, 0.05, n 4), although the difference for daily was maximum The barely significant. daily temperature averaged 4''C warmer September, warmer in declining 1.6°C December. Con- to in mean minimum versely, the difference in temperature increased from 0.6°C to 4°C over the same period (October December; Table to 1). Mean daily temperature CoUegeville average of at (the the daily maximum and minimum) was warmer l.S^'C to 2.7''C in 2001 than the long-term average. This is a substantial difference, equivalent in the wanning short term to the degree of expected under projected increases CO2 (Moore in global concentrations et al. 1999). Calculation of heat units above 5°C, taken as the threshold tem- perature for plant growth, illustrates the extent of warmer weather in 2001. Heat units declined steeply from October November to as the mean 30-70% days shortened; nevertheless, heat units 2001 were in greater than the long-term average (Table Similarly, negative heat 1). sum units, calculated as the of degrees below zero for each daily min- imum, summarize the frequency and severity of Negative heat frost. November units were substantially less in and December 2001 than the long-term average, indicating less severe frost (Table 1). Average temperatures at the inland station, CoUegeville, were about November 1°C cooler than near the coast, except in (Table The 1). first occurred on 20 September (— 2°C) and 30 September (— frosts I'^C) at month CoUegeville, but were delayed for a at Antigonish (24 and 28 September was no marked October). After there difference in the frequency of frosts between the two sites, but because of the lower minima inland, frost there was generally more severe. There was a between two substantial difference in heat units the sites as well (Table 1). A Plant collections. total of 92 species of herbaceous plants, representing 19 families, were found in flower in Antigonish County during November-December 2001 (Appendix; Table All were 2). were dicots. One-third of the total (30 species) native, with the re- from mainder (62 species) being introduced elsewhere, chiefly Europe (Zinck 1998). Although the greatest number of species was collected in early November, almost a third of the total (29 species) were still blooming 10 December. All plants in our collection were after ter- we despite regular inspections of aquatic habitats, found no restrial; One November. blooming specimen of Polygonum aquatic plants in ! 124 Rhodora [Vol. 105 f 4 y among Table 2. Distribution families of late-flowering plants from Antigonish, Nova Scotia, in autumn 2001 in order of species richness. ^Ratio of number of number species collected to of species listed in Zinck (1998). i Number Number of of Proportion of Known 1r Species Species Species in hamily Nova i Collected Scotia Collected' s ] + Asteraceae 30 166 0.18 J ± Brassicaceae 12 51 0.24 i Fabaceae 7 41 0.17 t Car>ophyllaceae 7 35 0.20 i i Polygonaceae 7 41 0.17 •i Scrophulariaceae 5 38 i 0.13 Rosaceae 5 97 0.05 ) 8 Lamiaceae 3 28 0.11 Apiaceae 3 30 0.10 ; Onagraceae 2 16 0.13 Malvaceae 2 1 7 0.29 Violaceae 2 15 0.13 J Campanulaceae h 8 0.13 p 1 Chenopodiaceae 23 0.04 ^ f I Clusiaceae 9 0.11 Euphorbiaceae 9 0.11 i Plantaginaceae 7 0.14 Ranunculaceae 24 0.04 I, Solanaceae i 9 0.11 t TOTAL S 92 654 r V t •^^ -_ __ ^ /• 1 i _ -A I bordering Several species were recorded based on single collections of one or J plants. This I culoides, Hcsperis matronalis, Viola Polygonum i cucullata. persi- caria, and P. sagittatum. In several cases only a small population -J rA of was individuals present, growing in highly protected environments ^t Cardamine (e.g., pensylvanica, Centaurea Malva nigra, neglecta, Chaenorrhinum minus, and Dianthus armeria). These soecies cannot be ant a unique event caused by unusual habitat conditions, such as a heat- reflecting wall. Our collections included members of 19 plant families of which were 12 represented by more than one species (Table Greatest diver- 2). occurred sity in Asteraceae (30 species) and Brassicaceae (12 species). These two families, along with the Caryophyllaceae, Fabaceae, and r- — Nova 2003] Taylor and Garbary Late Flowering in Scotia 125 0.4 (0 0) o V 0.3 75 o 0.2 o o t O 0.1 O Q. 2 3 4 5 6 7 8 10 9 11 12 1 Family Figure families 2. :ate proportions in our collections; light bars — Zinck Asteraceae; 2 ^Brassicaceae; 3 Fabaceae; indicate proportions in = = — — = 4 Caryophyllacceeaaee;; 55 Polygonaceae; 6 Scrophulariaceae; 7 Rosaceae; 8 — = = = Lamiaceae; 9 Apiaceae; 10 Onagraceae; Malvaceae; 12 Violaceae. 1 1 70% Polygonaceae (7 species each) constituted almost of all the late- flowering species. Asteraceae alone contributed almost a third of the The dominance of these families could be a consequence of a high total. proportion of late-flowering members, or merely the size of the family. In Figure the number of species in each family is expressed as 2, number Only most a proportion of the total of species collected. the 12 abundant those represented by more than one species in our families, For comparison, collection, were included in the calculation. the total number number of species in each family as a proportion of the of all known from Nova Scotia (Zinck 1998) also shown. species is With the exception of the Rosaceae the proportions of famiUes (7), from in our collection are not very different the proportions in the whole of Nova Scotia (Figure There are proportionately more flora 2), Caryophyllaceae, and Polygonaceae among Asteraceae, Brassicaceae, whole the late-flowering plants than in the flora, but the differences are among Hence, dominance of Asteraceae not the the late-flowering great. number plants appears to be a result of the great of species in this family, not a propensity toward extended flowering; the Asteraceae would be 126 Rliodora 105 [Vol. 100 O) 80 I o 60 S O Q. 40 CO o (D 20 E >1 Nov >10 Nov>20 Nov Dec >10 Dec >1 Range Date Figure umber 3. IN ot species in flower Antigonish in beginning November 2001. 1 expected dominate to collections any in season. In fact, the only family that deviates conspicuously from the expected proportion the is ants members of Rosaceae Nova the bloom in Scotia in spring (Zinck 1998). To examine the temporal we distribution of late-flowering plants, divided November-December the season into five periods of 10 days each (Figure Flowering ended middle 3). in the of on the period, fifth We day that flowers were observed, when regardless of the flowering first specimens were discovered. Hence, a species collected in flower for the time on November was first 18 assumed have been to blooming, undiscovered, before that date. The growing season 2001 in did not end we abruptly, had as expected. number Rather, the of species in flower declmed steadily from one period to the next (Figure until the remaining 3), plants died or were covered in The simple a linear function: — Nova Taylor and Garbary Late Flowering Scotia 127 in 2003J N= 107.2- T 16.0 N — = Where number T of species flower and sequential time in The period. coefficient of determination (r for this regression 0,99 is ) = (n 5). Thus, appears that, rather than truncating the growing season for it plants, bouts of cold weather during each period eliminated groups all By of species that had reached the limits of their cold tolerance. the last period, after 10 December, only the hardiest species, or those populations growing in protected coastal sites, persisted. Nevertheless, these 29 species, nearly one-third of the were blooming within total, still the next five days. There was a major decline in flowering of native species relative to introduced species as the season proceeded. In early November 33% (30 of 92) of the species in flower were native; by mid- December only 15% (4 of 29) of the species were native (Figure 4). DISCUSSION Nova Autumn of 2001 was remarkably mild throughout northern Comparison of Environment Canada weather records against Scotia. warm remained long-term normals confirms that temperatures far longer in the than usual in 2001. While intermittent frosts began at about fall week the same time in 2001 as in other years (usually the first of October were unpub. below-zero temperatures neither as frequent inland, data), nor as severe (fewer negative heat units) as in a normal year. It is how many nevertheless surprising plant species continued to flower and Night in the face of declining temperatures intermittent freezing. temperatures regularly intercepted zero in late October, both at Anti- warm November, gonish and inland. After a brief spell in early night became an irregular but frequent event for the rest of the month. frosts Another brief warm spell began in December, before the final descent Despite the obvious cold, most of the plants into winter temperatures. we were robust and evidently healthy, and some had sub- collected populations. stantial local Our with long-standing conclusions of observations are consistent to and hence the length of the plant ecologists sensitivity frost, tfiat among (Daubenmire growing widely species 1974). season, varies many continued grow and bloom In Antigonish County, species to repeatedly frozen several times each week, despite being (apparently) snow Each and sometimes covered with as well. period of cool weather, especially with nights of deep frost, led to the loss of progressively more