Impact of climatic factors on germination of the rare shrub V. staminosa
Cracking the Niche:
An investigation into the impact of climatic variables on germination of the rare shrub Verticordia staminosa subspecies staminosa (Myrtaceae).
Greg Simpson
Bachelor of Arts in Sustainable Development
Bachelor of Science in Environmental Science
School of Environmental Science
2011
Impact of climatic factors on germination of the rare shrub V. staminosa
Declaration
This thesis is my original work and has not been submitted, in whole or in part, for a degree at this or any other university. Nor does it contain, to the best of my knowledge and belief, any material published or written by another person, except as acknowledged in the text.
......
Gregory Douglas Simpson 20th May 2011
Word count for all parts of the thesis, excluding references and appendices:
______
Page i Impact of climatic factors on germination of the rare shrub V. staminosa
Abstract
The influence of annual rainfall and a Mediterranean climate in structuring the indigenous vegetation in the Southwest Botanical Province of Western Australia has long been recognised, especially in relation to seedling germination during the cooler wet season. However, over the past decade numerous authors have hypothesised that a number of climatic factors, including variability in the timing and intensity of rainfall events, may be drivers of germination. Verticordia staminosa subsp. staminosa (Myrtaceae) is a naturally rare narrow range endemic shrub that occurs on only one granite inselberg near Wongan Hills in the „wheatbelt‟ of southwest Western Australia. Smoke and fruit wall weathering have been claimed to be specific dormancy breakers for the seeds of many Australian genera, including Verticordia. However, I found no evidence of smoke and artificial weathering influencing germination of Verticordia staminosa. Using data from a long-term (12 year) investigation into field germination of Verticordia staminosa, my thesis applies logistical regression techniques to model the impact of climatic variables on germination within the only recorded Verticordia staminosa subsp. staminosa meta-population. My analyses reveal that a complex interplay of the amount of rainfall, number of rainy days, diurnal temperature range, and storms related to tropical cyclones/lows, under a traditional four seasons or six Noongar seasons climate, best describes the observed germination of Verticordia staminosa seedlings.
Page ii Impact of climatic factors on germination of the rare shrub V. staminosa
Table of Contents List of Boxes ...... vi
Table of Figures ...... vii
List of Tables ...... viii
Acknowledgements ...... ix
Chapter 1: Introduction
1.1 Introduction ...... 1
1.2 Organisation of this Thesis ...... 2
Chapter 2: Literature Review
2.1 Southwest Botanical Province of Western Australia ...... 4
2.2 Defining the Regeneration Niche for V. staminosa ...... 5
2.2.1 Defining a Species‟ Niche ...... 5
2.2.2 The Regeneration Niche ...... 6
2.2.3 The Regeneration Niche of V. staminosa ...... 7
2.3 Seed Dormancy and Germination Stimuli ...... 7
2.3.1 Seed Dormancy ...... 7
2.3.2 Germination Stimuli ...... 8
2.3.3 Weathering and Seed Imbibition ...... 8
2.4 Modelling the Impact of Climatic Variables ...... 9
2.4.1 What Climatic Variables to Model? ...... 10
2.4.2 What are the Seasons of South-West Western Australia ...... 11
Chapter 3: Methodology
3.1 Field Survey Site ...... 14
3.2 Study Species ...... 15
3.3 Seed Viability of the Soil Seed Bank ...... 16
3.4 Selection and Preparation of Germination Containers ...... 16
3.5 Longevity of Seed Viability ...... 16
Page iii Impact of climatic factors on germination of the rare shrub V. staminosa
3.6 Impact of Smoke and Weathering on Germination ...... 18
3.7 Germination Field Surveys ...... 19
3.8 Statistical Analysis ...... 20
3.8.1 General Approach to Statistical Analysis ...... 20
3.8.2 Interpretation of Logistical Regression ...... 21
3.8.3 Seed Viability for Fresh Pericarps ...... 22
3.8.4 Longevity of Seed Viability ...... 23
3.8.5 Impact of Smoke and Weathering on Germination ...... 24
3.8.6 Analyses of Climate Data ...... 24
3.8.7 Impact of Climatic Variables on Germination in the Field ...... 25
Chapter 4: Results
4.1 Exploratory Data Analyses ...... 27
4.1.1 Data Sets that Approximate a Normal Distribution ...... 27
4.1.2 Data that Can Not Be Normalised by Transformation ...... 27
4.2 Seed Viability of the Soil Seed Bank ...... 28
4.3 Germination Response of Recovered Pericarps ...... 29
4.4 Impact of Smoke Water and Weathering on Germination ...... 30
4.5 The Seasons of South-West Western Australia ...... 30
4.5.1 Annual Climate Metrics ...... 31
4.5.2 Mediterranean (Dry and Wet Season) Metrics ...... 32
4.5.3 Metrics for the Four Seasons Climate ...... 33
4.5.4 Six Noongar Seasons ...... 33
4.6 Impact of Climatic Variables on Germination in the Field ...... 34
4.6.1 Annual Values ...... 35
4.6.2 Mediterranean Climate ...... 36
4.6.3 The Traditional Fours Seasons ...... 36
4.6.4 Six Noongar Seasons ...... 38
4.6.5 Ranking the Season Models ...... 38
Page iv Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 5: Discussion
5.1 Longevity of Seeds in the Soil Seed Bank ...... 40
5.2 Impact of Smoke Water and Weathering on Germination ...... 40
5.3 Impact of Climatic Variables on Germination ...... 41
Chapter 6: Conclusions and Recommendations
6.1 Conclusions ...... 45
6.2 Recommendations for Future Research ...... 45
References ...... 47
Appendices
A. Exploratory Analyses: Ex-situ Germination Experiments ...... 57
B. Exploratory Analyses: Field Germination ...... 61
C. Exploratory Analyses: Seasons of SW Western Australia ...... 62
D. Outcomes of Least Squares Regression Analyses of Correlation between Climate Metrics ...... 84
E. Correlation Analyses for Noongar Seasons Temperature Metrics ...... 85
Page v Impact of climatic factors on germination of the rare shrub V. staminosa
List of Boxes
Box 1: Calculating the 95% Confidence Interval...... 20
Box 2: Calculating χ2 to compare nested Models...... 21
Box 3: Calculating the Standard Error and the 95% Confidence Interval...... 23
Box 4: Burnham and Anderson‟s (2002) approach to calculating AIC values...... 26
Page vi Impact of climatic factors on germination of the rare shrub V. staminosa
Table of Figures
Figure 1: Location of V. staminosa populations adapted from Gioia 2007 and Geoscience Australia 2009...... 14
Figure 2: The 5 critical values for Box & Whisker Plots used in this study...... 20
Figure 3: An example of a „model web‟ produced from a Logistical Regression analysis...... 22
Figure 4: Mean mass (± 95% & 99% confidence intervals (CI)) of pericarp with and without a viable seed and overall seed viability (± SE & 95% CI)...... 28
Figure 5: Seedling emergence in recovered pericarp replicates ...... 29
Figure 6: Web of models for effect of Tray and Treatment...... 30
Figure 7: Germination models with both Subpopulation and Year as factors...... 34
Page vii Impact of climatic factors on germination of the rare shrub V. staminosa
List of Tables
Table A: The Six Noongar Seasons...... 12
Table B: Long Term Climate Averages for Wongan Hills Meteorological Station 1981-2000...... 14
Table C: Experiment Balance for Recovered Pericarps Experiment...... 17
Table D: Balance of the Glasshouse Experiment with Four Treatments...... 19
Table E: Balance for Field Germination Experiments...... 19
Table F: Calculation of climate metrics shown to impact germination & seedling survival...... 24
Table G: Outcomes of Chi-Square Tests for differences between proportions of seedlings within treatments and between treatments (Buried in Cracks or in Soil Aprons)...... 29
Table H: Outcomes of Chi-Square Tests for difference between annual values and the 20 year (1991-2010) long-term averages for key climate metrics...... 31
Table J: Outcomes of t-Test with uncommon variance of between seasons...... 32
Table K: Variability in the temperature metrics under the Mediterranean Season‟s model...... 32
Table L: Variability in the temperature metrics under the Four Season‟s model...... 33
Table M: Variability in the temperature metrics under the Six Noongar Season‟s model...... 34
Table N: Ranking of logistical models for germination based on Burnham and Anderson‟s (2002) approach to calculating AIC values...... 39
Page viii Impact of climatic factors on germination of the rare shrub V. staminosa
Acknowledgements
I need to thank my family for the part they have payed in the production of this thesis. To my son Luke who provided the motivation for me to complete my BSc (Hons) in Environmental Science and to become a better person. To my Mum and Dad who wondered many times if they would ever see this day, but always stood by me none the less.
While it is always traditional to do so, I wish to genuinely thank my supervisor (the not quiet so long suffering) Dr Philip Ladd for his guidance, patience, and perseverance through both my undergraduate and Honours programs. Your mentoring and support Phil has opened a plethora of opportunities that I would never have experienced otherwise.
Thanks also for the support of my mates Richard and Andrew who have shared the twists and turns of life‟s journey and a wine or three whenever it was needed.
I would like to acknowledge the Noongar People of the Bibbulmun Country as I draw on their 40,000+ years of climatic and ecological knowledge for this thesis. Specifically I would like to acknowledge the Ballardong Noongar People as the traditional custodians of lands in the vicinity of Wongan Hills.
I am grateful to the Men of the Trees and family of Shirley Fyffe for providing the scholarship that put a roof over my head while I completed this research and has set me up for the next stage of my life.
This thesis would not have been possible without the data from the Department of Environment and Conservation‟s long term study of Verticordia staminosa germination provided by Dr Colin Yates and the weather observations for Wongan Hills published on the Australian Bureau of Meteorology‟s website.
Page ix Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 1: Introduction and Thesis Organisation 1.1 Introduction
At the start of the 20th century Fredrick Clements (1905) wrote “the factors of a habitat are water-content, humidity, light, temperatures, soil, wind, precipitation, pressure, altitude, exposure, slope, surface (cover), and animals”. A foundation researcher in the field of Ecology, he went on to say that “length of season” was an important and complex influence on vegetation that should “be treated under its constituents”, that is humidity, light, temperature, wind, and precipitation. While Clements provided this sage advice over 100 years ago, little progress has been made in elucidating the impact of climate variables on recruitment, even for the highly studied vegetation found in Mediterranean climates. In a review of how plant life-history and ecological traits relate to rarity or commonness, Murray et al. (2002) found that “a large proportion of studies” are based on small sample sizes with little replication and that duplicate studies “overwhelmingly … differ from one and other in their findings”. When writing about the impact of climate change on Australian ecosystems and native species Hughes (2003) said:
“Predicting what will happen to species distributions in the future must rely heavily on understanding the [climatic] factors by which they are currently limited. Although the question ‘What determines the distribution and abundance of species?’ is fundamental in ecology, we do not know the answer to it for any more than a handful of Australian taxa.”
Cody (1986) and Cowling et al. (2005) both identify the need to more fully investigate the impact of environmental variables on germination in Mediterranean climates, specifically in terms of rainfall variability in the southwest of Western Australia. More recently, Yates and Ladd (2010) concluded that in relation to vegetation in the fragmented landscape of south-western Australia the “effects of climate change on the persistence of rare species … are now being recognised as an area in need of investigation”. The predicted increase in temperatures and decrease in precipitation over most of Australia over the second half of this century further increases the importance of characterising climatic thresholds for vegetation (Christensen et al. 2007; Fischlin et al. 2007; Hughes 2003; Cochrane and Daws 2008).
Characterising every abiotic and biotic factor that impacts seedlings germination for the naturally rare shrub Verticordia staminosa subsp. staminosa (Gardiner
Page 1 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa and George 1963) to define its realised regeneration niche is clearly beyond the scope of an Honours project. Following MacArthur‟s (1968, cited in Krebs 1994) recommendation to describe a niche in terms of a limited number of quantifiable variables, my research investigates possible relationships between climatic variables, fruit being exposed to smoke and weathering, and the quantity and rate of Verticordia staminosa subsp. staminosa seedling emergence. To illuminate any relationship between these parameters I constructed a number of metrics to explore differing hypothesis by applying multivariate statistical techniques to the analysis of both long term (12 years) germination data maintained by the Department of Environment and Conservation and for data I generated in glasshouse germination trials. Specifically I propose the null hypotheses that germination of Verticordia staminosa subsp. staminosa is not impacted by:
The amount and/or timing of rainfall.
Air temperature.
Non-climate variables, such as the age of fruit (hereafter pericarp), the exposure of the pericarp to smoke, and/or weathering of the pericarp.
The experimental design and statistical analyses I apply in this study allowed me to investigate the appropriateness of both Eurocentric and Noongar seasons to elucidate the impact of climatic variables on the germination of Verticordia staminosa subsp. staminosa (hereafter V. staminosa, unless otherwise specified).
1.2 Organisation of this Thesis
This thesis is structured so that each chapter provides information and builds an understanding of the material that is presented in the following chapter. The structure of my thesis and a brief overview of each chapter appear below.
Chapter 1: Introduction. This chapter provides a rationale for my investigations, defines the research questions that I explore in the study (null hypotheses), outlines how my thesis is structured, and how my understanding of the variables that influence V. staminosa germination developed.
Chapter 2: Literature Review. This chapter provides background to the major topics of my thesis: the biogeographical characteristics of the Southwest Botanical Province (particularly in relation to V. staminosa), the importance of the regeneration niche for persistence of V. staminosa, which climate variables (e.g. amount of rainfall, maximum and minimum temperatures) impact the life
Page 2 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa histories of vegetation, and the character of the climate and structure of the seasons in southwest Western Australia.
Chapter 3: Methods. I start this chapter with an overview of the field study site and the species V. staminosa. Next I outline my experimental approach to gathering the data for my research before moving on to detail my approach to statistically analysing that data.
Chapter 4: Results. Detailing the findings of my project, this chapter progresses from the insights I gained in the exploratory data analyses and through my investigation of the soil seed bank. It provides evidence of the three different patterns of season in southwest Western Australia, and concludes by presenting and ranking a series of logistical models that describe how climatic variables impact V. staminosa germination.
Chapter 5: Discussion. I use this chapter to describe the understandings that I developed during this study and to place the outcomes of my research in context with the findings and hypotheses reported in the published works of other researchers.
Chapter 6: Conclusion and Recommendations. I conclude my thesis by providing a summary of the most significant findings from my research. I revise the research questions specified in the Introduction of my thesis, pose alternate hypotheses informed by my research, and make recommendations for additional research that will build on this study and further illuminate the variables that most significantly impact germination of the indigenous vegetation of the Southwest Botanical Province.
Page 3 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 2: Literature Review 2.1 Southwest Botanical Province of Western Australia
Located in the south western corner of the Australian continent and covering approximately 360,000 km2, the Southwest Botanical Province of Western Australia (SWBP) is bound by the Indian Ocean, the Southern Ocean, and a diagonal line running southeast from the southern end of Shark Bay to approximately Esperance (Israelite Bay) in the south (Beard 1980; Gole 2006) The SWBP is analogous with both the slightly smaller (300,000 km2) Southwest Australian Floristic Region (SWAFR) of Hopper and Gioia (2004) and the Bibbulman Country that the Noongar people have occupied for the past 50,000 years (Gole 2006; South West Aboriginal Land and Sea Council 2011). Somewhat confusingly WWF-Australia groups the SWBP with the more arid “Transitional Zone” to the east, which is different to the “Transitional Rainfall Province” proposed by Hopper and Gioia (2004), to create the slightly larger (490,000 km2) Southwest Australia Ecoregion (Gole 2006).
With a climate of warm dry summers and cooler wetter winters, the SWBP is one the five Mediterranean climate regions spread around the planet, the other four are centred on northern Africa/southern Europe (Mediterranean basin); California; Chile; and the Cape region of southern Africa (Cody 1986; Lamont 2004; Cowling et al. 2005). Covering less than 5% of the Earth‟s surface these five regions contain almost 20% of the known vascular plants species, many of which are rare narrow range endemics (Cowling et al. 1996; Hopper and Gioia 2004).
The combination of bimodal (dry season/wet season) climate, frequent fires due to summer aridity (expect for Chile), and/or isolation of habitats over geological timescales by both natural land forming processes and fragmentation by humans (thorough activities such as agriculture, logging, and more frequent fire regimes) has resulted in these bioregions having exceptionally high alpha (α), beta (β), and gamma (γ) diversity compared to other temperate biomes (Attiwill and Wilson 2004; Hopper 1979; Cody 1986; Cowling et al. 1996; Lamont 2004). Among the most floristically diverse places on the planet, sites within the SWBP have species richness comparable to tropical rainforests and the fynbos vegetation of the southwest Greater Cape Floristic Region in Africa (Groves and Catling 2004; Cowling et al. 1996). A millennia of biological isolation after Australia rifted from Antarctica between the mid-Cretaceous (130 Million Years
Page 4 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Before Present [MYBP]) and the Oligocene (45 MYBP), the mass loss of angiosperms during the K-T extinction event (66 MYBP), and increased aridity during the late Tertiary (11 MYBP) and mid Quaternary (0.025 to 0.015 MYBP) driving the differentiation and speciation has resulted in between 70% and 80% endemism (at least 3,500 species) for vascular flora of the SWBP (Attiwill and Wilson 2004; Hill and Brodribb 2004; Hopper and Gioia 2004; Raven et al. 2005). Further, many of the species, sub-species, and varieties of the SWBP are considered narrow range endemics and may occur as a single population or across a few disjunct meta-populations (Hopper et al. 1997; Hopper 2009; Withers 2000).
While one of the most species rich regions on the planet, since the introduction of European agricultural practices less than 200 years ago over 70% of the indigenous vegetation has been cleared from the SWBP (Myers et al. 2000). As a result at least 351 vascular plant species and the fauna they support are at risk of extinction, giving Australia‟s southwest corner the shameful distinction of being labelled a Global Biodiversity Hotspot (Myers et al. 2000; Gole 2006).
2.2 Defining the Regeneration Niche for V. staminosa 2.2.1 Defining a Species’ Niche
The 3rd edition of the Oxford Dictionary of Ecology notes that the range of a species is both “the entire geographical area over which it occurs” and “the spread of environmental conditions within which it occurs”. This broad meaning encompasses not only the current spatial distribution of a species, but also the abiotic and biotic aspects of its habitat and the temporal variability as ranges expand or contract across a season, a year, a decade, a century, a millennia or even geological time (Kruckeberg and Rabinowitz 1985; Dynesius and Jansson 2000; Attiwill and Wilson 2004).
At the time Clements (1905) was postulating which environmental variables constitute the habitat of a species, Grinnell (1904) wrote that it was not possible for two species (or subspecies) adapted to the same environmental “conditions” and/or “food habits” to exist without competing for resources and “only by adaptations to different sorts of food, or modes of food getting, that more than one species can occupy the same locality.” Developing this hypothesis further, he was among the first Ecologists to suggest segmenting a habitat into “niches” such that no two species occupying the same range could have “precisely the
Page 5 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa same niche relationships” (Grinnell 1917). Many authors credit C.S. Elton‟s 1927 book Animal Ecology (cited in Krebs 1994) with advancing Grinnell‟s concept by defining a niche as a species‟ role within its community. However, Udvardy (1959) thought that Grinnell‟s concept of niche was complementary with Elton‟s hypothesis that “the 'niche' of an animal means its place within the biotic environment” (quoted in Udvardy 1959 and Krebs 1994). Building on the work of Grinnell and Elton, as expanded by Lotka, Volterra, and Gause (Udvardy 1959; Krebs 1994), Hutchinson (1957) defined the “fundamental niche” of a species to be an “n-dimensional hypervolume [N1]…, every point in which corresponds to a state of the environment which would permit the species S1 to exist indefinitely.”
Hutchinson proposed that for a second species (S2) there would a corresponding hypervolume (N2) and as no two species can occupy exactly the same niche he went on to define “the realised niche N’1 of S1 in the presence of S2 as (N1 - N2), if it exists, plus that part of N1N2 as implies survival of S1, and similarly the realised niche N’2 of S2 as (N2 - N1), if it exists, plus that part of N1N2 corresponding to survival of S2 [bolding Hutchinson‟s emphasis, italics are mine]”.
2.2.2 The Regeneration Niche
The above theories arose from studies of competition among animals, primary in relation to food and physical resources (e.g. nesting and roosting sites) within their habitat. Grubb (1977) proposed that for plants the fundamental and realised niches should be further subdivided with the habitat niche, time niche, growth form niche, and regeneration tactics niche being key for defining which species can coexist. Mature plants can generally cope with a relatively wide range of environmental conditions, but species often require very specific conditions for regeneration (New 2004; Cowling et al. 2005; Fenner and Thompson 2005; Merritt et al. 2007). Grubb (1977) wrote that of the four niche types listed above, the “regeneration tactics niche” was the most important. He postulated this is because the death of any plant creates a space in the regeneration niche and whether that space is filled by an individual of the same species or colonised by an entirely different species depends on:
a) Plant processes in terms of reproductive capacity (by seed or vegetatively), dispersal, germination, establishment, and onward growth; and
b) Characteristics of the space such as size, shape, method and timing of its creation, disturbance of the growing medium (soil and litter), climatic conditions, plant/animal species present, and so on.
Page 6 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
It should be noted that while the concept of the niche is generally accepted as a foundation theory of Ecology, not all researchers support this view and a study by Richardson et al. (1995) found lower diversity than that predicted by niche differentiation and no conclusive evidence of a regeneration niche.
2.2.3 The Regeneration Niche for V. staminosa
A combination of climatic variables (predominately rainfall) and the density of crevices (with suitable micro-habitat) in the granite pavement of the inselberg on which V. staminosa occur appear to be the primary determinants of abundance and distribution of the species (Yates and Ladd 2004; Yates et al. 2007). An increase in both germination and survival of V. staminosa seedlings has been observed in a year when the rainfall was substantially greater than the long term annual mean (Yates and Ladd 2004).
2.3 Seed Dormancy and Germination Stimuli 2.3.1 Seed Dormancy
The term „seed‟ is generally taken to mean the embryo plant and the protective seed coat, that is the testa and, where present, the hard outer capsule or pericarp (Leopold 1964). However, when discussing seed dormancy it is more appropriate to apply the definition of Fenner and Thompson (2005) that a seed is just the embryo plant excluding the seed coat.
Dormant dispersed seeds are prevented from responding to environmental germination stimuli by a single or more often a combination of morphological, physical, or physiological mechanisms (Leopold 1964; Fenner and Thompson 2005; Merritt et al. 2007). Overcoming morphological dormancy, which arises when fruit are dispersed before the seed has completely matured, requires additional development of the seed before germination can occur (Pickup et al. 2003; Fenner and Thompson 2005). The most common dormancy state or syndrome, physiological dormancy is alleviated by a change in the seed that overcomes inhibitor chemicals contained within the embryo or generated by the seed coat (Leopold 1964; Pickup et al. 2003; Fenner and Thompson 2005). Physical dormancy is imposed by the seed coat restricting space for the embryo to develop, being impermeable to respiration gases, and/or more commonly being impermeable to moisture (Leopold 1964; Fenner and Thompson 2005; Merritt et al. 2007). While light may play a role, temperature alone has been demonstrated to alleviate both physical and physiological dormancy by cracking
Page 7 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa the seed coat or increasing its permeability and by facilitating chemical changes in an imbibed seed (Leopold 1964; Fenner and Thompson 2005; Merritt et al. 2007). While morphological and physical dormancy cannot be reimposed once alleviated; a seed that has been released from dormancy and imbibed, but not stimulated to germinate, may redry and enter (or return to) a state of secondary physiological dormancy (Fenner and Thompson 2005; Merritt et al. 2007).
A seed‟s state of dormancy can only be assessed by observing if it develops when exposed to germination stimuli such as temperature, moisture, light, and/or chemicals, for example gibberellins, nitrates, or smoke (Pickup et al. 2003; Fenner and Thompson 2005). Determining if a seed is dormant or if it is imbibed but inactive can be a difficult and contentious assessment (Pickup et al. 2003; Fenner and Thompson 2005; Merritt et al. 2007; Cousens et al. 2010).
2.3.2 Germination Stimuli
Growth of a seedling is the outcome of the progression through the separate but closely coupled processes of the seed being released from dormancy, imbibition of the non-dormant seed, and the response of the imbibed embryo to germination stimuli (Pickup et al. 2003; Fenner and Thompson 2005; Merritt et al. 2007). As mentioned previously, temperature is the only environmental factor that has been demonstrated to alleviate dormancy and all other aspects of the seed‟s microenvironment become germination stimuli (Fenner and Thompson 2005; Merritt et al. 2007). This is a marked shift from the historical view of dormancy-breaking by chemical (e.g. gibberellins and smoke), light and photoperiod, moisture (e.g. wetting/drying cycles), and/or temperature (e.g. stratification) treatments (Leopold 1964; Dixon et al. 1995; Hopper et al. 2006). This shift in understanding is illustrated by the change in language from the hypothesis of Dixon et al. (1995) that ammonia/ammonium in smoke “might be one of the agents active in breaking dormancy” to the observation in Merritt et al. (2007) that “the action of smoke as a germination stimulant, rather than a dormancy-breaking agent, is becoming clearer.”
2.3.3 Weathering and Seed Imbibition
Many studies of germination of Australian hard seeded species report the requirement for a period of „dry after-ripening‟ (morphological dormancy) and enhanced rates of germination for soil stored seed as compared to fresh seed and laboratory stored seed banks (Pickup et al. 2003; Merritt et al. 2007; Martyn
Page 8 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa et al. 2009; Cousens et al. 2010). Morphological dormancy in SWBP soil seed banks avoids fresh seed competing with primed seeds (i.e. dormancy alleviated and imbibed) and prevents late winter rain stimulating germination without sufficient ongoing soil moisture for the seedling to establish prior to the summer drought (Tieu and Egerton-Warbuton 2000; Pickup et al. 2003; Martyn et al. 2009; Cousens et al. 2010). As outlined in Section 2.3.1, the enhanced germination of soil stored seeds following a period of dry after-ripening is attributed to temperature and (possibly) moisture cues in alleviating physiological and physical dormancy (Pickup et al. 2003; Merritt et al. 2007; Cousens et al. 2010). Physiological dormancy is alleviated through natural stratification during periods of elevated soil temperatures and increased moisture from (generally) light early season rains from March to May (Baker et al. 2005; Fenner and Thompson 2005; Merritt et al. 2007). Physical Dormancy is alleviated through weathering (temperature and moisture cycling) weakening the seed coat to increase its permeability to water and gases or in the case of hard seed by cracking the pericarp (Tieu and Egerton-Warbuton 2000; Lush et al. 2003; Pickup et al. 2003; Merritt et al. 2007). Although Martyn et al. (2009) and Cousens et al. (2010) have recently challenged the premise that the pericarps of hard coated Australian seeds enforce physical dormancy through imperviousness to moisture that is alleviated by weathering. Both studies found that hard coated seeds rapidly absorb moisture and start to imbibe within 24 hours of being exposed to available soil moisture. Similar to the findings of Lush et al. (2003), these two papers hypothesise that the seed coat reduced desiccation of the seed during dry periods between rain events promoting an incremental or “stop-go” process of imbibition and germination. Despite differences in view about the moisture permeability of pericarps, all three studies found that fluctuations in temperature and moisture (wetting and drying cycles) experienced by a hard coated seed in the soil degrades and cracks the seed coat, alleviates primary and secondary physiological dormancy, and enhances imbibition of the seed increasing the rate of germination.
2.4 Modelling the Impact of Climatic Variables
The exceptionally high diversity, geographical isolation, and climatic similarity of the Mediterranean regions has generated many studies of convergent and divergent evolution (Cody 1986; Cowling et al. 1996; Lamont 2004). It would seem reasonable to assume that a number of these studies would have
Page 9 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa investigated the impact of climate on regeneration. However, the majority of vegetation studies fail to consider the impact of climatic variables or any reported relationships are often based on only a cursory analysis (Cody 1986; Murray et al. 2002; Hughes 2003; Cowling et al. 2005). With increasing concern about the impact of rapid climate change on the range and survival of species, there is growing recognition of the need to predict changes in distribution and likelihood of climate induced extinction (Hughes 2003; Fay 2009; Tarroso and Rebelo 2010; Yates and Ladd 2010).
2.4.1 What ‘Climatic Variables’ to Model?
While Clements (1905) recognised the importance of humidity, light, temperature, wind, and precipitation in defining a plant‟s habitat, what metrics and interactions of these environmental variables are significant for germination?
Annual rainfall has long been recognised as an important determinate of species distribution (Gardner 1959; Hooper 1979). Increasingly authors are also recommending that variations in the seasonality of rainfall, the frequency of rainfall events, and the amount of rainfall and duration (i.e. intensity) of individual events be considered (Cowling et al. 2005; Fay 2009; Robertson et al. 2009).
In addition to these precipitation predictors, Tarroso and Rebelo (2010) also identify the month to month variation in the monthly average temperature, monthly average maximum temperature, monthly average minimum temperature, diurnal temperature range, vapour pressure (i.e. evaporation and/or humidity), cloud cover, and (where appropriate) frequency of frost free days. At a finer spatial and temporal scale, Fontaine et al. (2010) assessed the daily (24 hour) maximum, minimum, diurnal range, and mean air temperatures (1m above soil surface); the mean daytime temperature (7am to 6pm); the mean „heating‟ temperature (9am to 2pm); and the mean cooling temperature (4pm to 9pm) at multiple locations in logging stands post wildfire. This study identified the need to investigate additional climate related variables such as soil temperatures, relative humidity, and wind.
Another factor that warrants investigation is the potential for intense summer deluges, resulting from the passage of cyclones and topical storms, to influence V. staminosa germination (White et al. 1999; Cowling et al. 2005).
Page 10 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Providing a practical example of how such modelling can be applied, Nicholls et al. (2001) were able to develop „skilful predictions‟ of wheat yields using monthly rainfall values and diurnal temperature range as predictors.
2.4.2 What are the Seasons of South-West Western Australia?
The influence of annual precipitation in structuring the vegetation associations of the SWBP has long been reported by botanists, but the classic Mediterranean warm dry summers and cooler wetter winters is also known to be significant (Gardner 1959; Mott 1972; Brown 1989). What evidence is there about the seasonality of climatic metrics in south-west Western Australia?
The Mediterranean nature of weather in south-west Western Australia (hereafter SWWA) has been recognised from the earliest days of the Swan River Colony, however there has been less clarity about which months constitute the „warm dry summer‟ and „cooler wetter winter‟ (Curlewis 1929 and Wells 1884 quoted in Western Australia. Indian Ocean Climate Initiative (IOCI) Panel 2002; Nicholls et al. 1999; Cowling et al. 2005). Wells (1884) wrote “The wet season commences with light showers in April, which continue to increase in number and force throughout May, June, and July, and from that period decrease until they cease altogether, in the month of November, when the dry weather begins”. Based on evidence from rainfall, both Dirnböck et al. (2002) and Cowling et al. (2005) classified April to September as the southern hemisphere winter and consequently summer occurs from October to March. However, W.J. Wright (1997) found that 80% of the SWWA‟s precipitation fell from May to October and the Australian Bureau of Meteorology Research Centre assigns these months as SWWA‟s „winter‟ period (Nicholls et al. 1999). Two major rainfall variability signals have been noted within the bimodal climate of SWWA. In most years more rain falls in early winter with August to October being dryer (Wells 1884; P.B. Wright 1971, cited in Allan and Haylock 1993; P.B. Wright 1974), and occasionally deep penetrating tropical cyclones transport moisture from northern latitudes into the SWWA causing large summer rainfall events (Milewski 1979, cited in Cowling et al. 2005).
Early European settlers also applied the traditional „Four Seasons‟ model of summer (December-February), autumn (March-May), winter (June-August), and spring (September-November) to the weather of SWWA (Nind 1831). While reporting of almost every botanical investigation undertaken in the SWBP/SWWA has a comment about the region‟s Mediterranean climate, only a few publications
Page 11 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa make mention of the four seasons applied by the general public (Rippey 1995; E.A. George 2002; Hussey and Wallace 2003). However, over the past 40 years numerous studies found significant evidence to support the four season model (Gentilli 1971, cited in Western Australia. IOCI Panel 2002; P.B. Wright 1974; Allan and Haylock 1993; Nicholls et al. 2001). These studies investigate climate metrics such as terrestrial air temperatures, precipitation patterns, and sea surface temperatures.
As the first human arrivals in SWWA, the Noongar people have resided in the SWBP for the past 40,000-50,000 years (Flood 1995; Bindon 1997; Gole 2006). Many of the Noongar creations stories begin “Long before the nyitting (sometimes nydiny) …” or “In the nyitting …”, which loosely translates as “Before/In the cold dark times long ago” (Morrison et al. 1993, quoted in Wallace and Huston 1997; Nannup and Deeley 2006). This living oral history is a record of the cooling experienced in SWWA/Bibbulman Country during the last ice age 14,000-20,000 years ago (Australia. Bureau of Metrology 2010). As the last Ice Age abated and temperatures began to rise, the Noongar people moved across the landscape in response to the rhythm of the six seasons (Table A) that emerged in SWWA (Harben and Collard n.d; Wallace and Huston 1997; Australia. Bureau of Meteorology 2010).
Table A: The Six Noongar Season (Adapted from Harben and Collard n.d.; Bindon and Walley 1992; Murdoch University 2008).
Noongar Approximate Character of the Season Season Timing Birak Dec-Jan First Summer. Season of the Young. Reptiles emerge from hibernation. Hot & dry time with easterly winds during the day and cooler south-westerly sea breeze during late afternoon. Bunurua Feb-March Second Summer. Season of Adolescence. Children learn the art of tool-making. Hot easterly and north winds. Fishing & hunting coincide. Large sections of the country were abandoned for lack of water. Djeran April-May The Autumn. Season of Adulthood. Time of marriages & courtship ceremonies. Becoming cooler with winds from southwest. Makuru June-July The First Rains. Fertility Season. Time to travel to inland hunting areas. Cold & wet weather with westerly gales. Djilba Aug-Sept The Second Rains. Season of Conception. Rains fill lakes & waterholes. Becoming warmer. Kambarang Oct-Nov The Wildflower Season. Season of Birth. Carpets of wildflowers. Receding rain. Small family parties‟ link together and larger groups started moving towards the coast.
Page 12 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
The progression of the Noongar seasons from the first rains of Makuru to Djilba and on into Kambarang with its receding rain, aligns with the previously mentioned observations of Wells (1884, quoted in IOCI Panel 2002), P.B. Wright (1974), and Allan and Haylock (1993), all of whom reported that generally most rainfall occurs in the early months of the wet season, with August to October being dryer.
Page 13 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 3: Methodology 3.1 Field Survey Site
Typical of the numerous short range endemics for which the Southwest Botanical Province is renowned, the distribution of Verticordia staminosa subsp. staminosa (Figure 1) is restricted to crevices in the granite pavement on the upper slopes of a single inselberg approximately 10km northeast of the wheatbelt town of Wongan Hills (Withers 2000; Hopper and Gioia 2004; Yates and Ladd 2004).
One of the six regions around the planet which are described as having a classically Mediterranean climate, the SWBP experiences warm dry summers and cool wetter winters (Australia. Bureau of Meteorology (BoM) V. staminosa ssp. staminosa 2011; Cowling et al. 2005; Yates and Ladd 2004). However, the Noongar people of V. staminosa ssp. cylindracea south-western Australia recognise six seasons var. erecta within this region (Australia. Department of V. staminosa Environment and Heritage 2006; Western ssp. cylindracea var. cylindracea Australia. Department of Water 2011). BoM (2011) weather observations from the
Wongan Hills township provide the nearest verified and analysable climate data (Table B) Figure 1: Location of V. staminosa populations adapted from Gioia 2007 for Verticordia staminosa subsp. staminosa. and Geoscience Australia 2009.
Table B: Long Term Climate Averages for Wongan Hills Meteorological Station 1981-2000. *Daily Evaporation is from Wongan Hills Research Station site which is equidistant from Mocardy Hill and compatible with weather data from the Township site (BoM 2011). Monthly Mean Monthly Mean Annual Summer Parameter Winter (May-Oct) Mean (Nov-April) (Mean ± 95% CI) (Mean ± 95% CI) Maximum Temp. 25.7 oC 31.2 ± 2.4 oC 20.1 ± 2.4 oC Minimum Temp. 11.9 oC 15.7 ± 1.7 oC 8.2 ± 1.2 oC Diurnal Variation 13.7 oC 15.5 ± 1.2 oC 11.9 ± 1.5 oC Rainfall 391.4 mm 17.7 ± 2.8 mm 47.4 ± 15 mm Rainy Days 87.6 3.4 ± 0.8 11.2 ± 2.6 Daily Solar 20.1 MJm-2 25.0 ± 4.1 MJm-2 15.1 ± 4.1 MJm-2 Exposure Daily Evaporation* 6.2 mm 9.0 ± 1.7 mm 3.4 ± 1.1 mm
Page 14 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
3.2 Study Species
Verticordia staminosa subsp. staminosa is a naturally rare sclerophyllous Myrtaceae shrub which grows in a single meta-population of approximately 1 300 plants distributed across 49 largely unconnected subpopulations (E.A. George 2002; Yates and Ladd 2004; Hopper 2009). Approximately 400 kilometres (Figure 1) and an estimated 1.0-1.5 million years of evolution separates subspecies staminosa from its closest relatives, two varieties of the equally rare subspecies cylindracea that grow in several disjunct populations on granite outcrops in the Newdegate/Lake King area (A.S. George 1991; E.A. George 2002; Patten et al. 2004; Yates and Ladd 2004).
During the winter wet season (May-Oct) the one to several basal stems of a reproductive Verticordia staminosa subsp. staminosa plant support a dense low (0.3-0.6 metres) spreading (0.6-1.0 metres) bright green canopy of narrow almost stalkless leaves and many yellow-green to reddish-brown pendulous hermaphroditic non-autogamous flowers that hang under the foliage (A.S. George 1991; E.A. George 2002; Yates and Ladd 2004). All three forms of V. staminosa become dormant to avoid the summer drought (E.A. George 2002; A.S. George 2002; Yates et al. 2007). After rain they recover from a brown mostly leafless and, in some cases collapsed state, with the retained foliage rapidly greening and by producing “small, vigorous tufts of new regrowth on old branches” (Patten et al. 2004; Yates and Ladd 2004). Having no lignotuber, Verticordia staminosa is generally considered to be a fire sensitive species (E.A. George 2002; Patten et al. 2004; Yates et al. 2007).
Verticordia staminosa is a surplus flower producer with low seed set, approximately one in three flowers develops a single viable seed (Yates and Ladd 2004), although a few pericarps contain a second viable seed. It has been suggested that this relative low cost strategy provides an ovary reserve as a buffer against flower mortality caused by adverse weather conditions and/or allows such species to increase germination and establishment rates in seasons with especially favourable environmental conditions (Cochrane et al. 2001; Yates and Ladd 2004).
An increase in germination of V. staminosa seedlings was observed in a year when the rainfall was substantially greater than the long term average (Yates and Ladd 2004). Seedlings of all three forms of Verticordia staminosa have been
Page 15 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa observed to germinate in cracks in the pavement, moss mats, and/or shallow pockets of soil (E.A. George 2002; Patten et al. 2004; Yates and Ladd 2004; Yates et al. 2007). However for Verticordia staminosa subsp. staminosa, only those seedlings that germinate in a crevice survive the first summer (Yates and Ladd 2004; Yates et al. 2007).
3.3 Seed Viability of the Soil Seed Bank
During a site visit in winter of 2008, Dr Colin Yates from the Science Division of the Department of Environment and Conservation (formerly within the Department of Conservation and Land Management) and Dr Philip Ladd from the Murdoch University School of Environmental Science collected a random representative sample of pericarps from under reproductive plants within several subpopulations in the vicinity of monitored subpopulations, which they aggregated into a well mixed composite sample. I randomly selected three sets of 40 pericarps from the composite sample and carried out a „Cut Seed Test‟ to determine which pericarp contained a viable seed (Moles et al. 2003; Turner 2005). I removed the sepals and petals from the selected pericarp by brushing them over a stainless-steel 100μm mesh, weighing them on an analytical balance, and then cutting the pericarps in half to determine how many contained a plump white embryo. My method for statistically analysing the outcome of this Cut Test is detailed in Section 3.8.3.
3.4 Selection and Preparation of Germination Containers
For all my germination experiments I randomly selected the most appropriate matched set of punnets, seedling pots, and/or trays from glasshouse stock; thoroughly cleaned the containers in fresh chlorinated water; sterilised them in 1% hypochlorite solution; rinsed the containers in deionised water; and then let them air dry at 65oC for 45 minutes. Under hygienic conditions I marked each pot with a unique code that identified the container and any treatment applied. To simulate the mineral soils of crevices in the granite pavement on Mocardy Hill, I filled each container with washed river sharp sand supplemented with a proprietary slow release fertiliser formulated for Australian native plants.
3.5 Longevity of Seed Viability
The experimental design for soil storage of pericarps to determine the longevity of seed viability was developed and implemented by Dr Yates and Dr Ladd. The detail for actions performed prior to my participation in both the annual
Page 16 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
V. staminosa demographic surveys from 2007 and recovering buried samples of pericarps in late 2009 are based on my personal communications with Dr Ladd and Dr Yates during the period 2007-2011. In 2001 a random representative sample of pericarps collected from under reproductive plants growing in several subpopulations in the vicinity of monitored subpopulations, were aggregated into a well mixed composite sample, and then 54 sets of 100 fruit were randomly selected from the composite sample. Half the sets of pericarps were randomly selected for burial in cracks within a single subpopulation and the other half for burial in soil aprons adjacent to the same subpopulation, which were in part covered by mosses. The sets of pericarps were mixed with washed river sharp sand and stitched into free draining nylon bags of 20cmx2cm for burial in cracks or 10cmx10cm for burial in the soil aprons.
The wider section of three unmonitored cracks in the granite pavement at subpopulation H were excavated to a depth of approximately 5cm, the nylon bags were tamped into position, the crack was backfilled with excavated material, and wire stakes with plastic tags were used to mark the location of the sample bags. In three aprons of soil congruent to subpopulation H, stakes were driven into the ground in 3mx3m matrices and the sample bags buried to a depth of 5cm at 1m spacing within each grid. The original experimental design involved recovering a randomly selected subset of nine sample bags from the cracks and three sample bags from each soil apron at intervals of one, two, and three years after burial to characterise longevity of seeds in the soil seed bank. However this did not occur. In the spring of 2009 Dr Ladd and I recovered all the bagged samples (hereafter replicates) that we could locate.
Using a binocular dissection microscope and tweezers, I removed all complete and fragmented pericarps from the washed river sand and root material contained in each recovered replicate. To demonstrate that none of the seeds remained viable, I planted all the visually desiccated and weathered pericarps (complete and fragmented) into 30 individual matched pots. Without replication, I planted all pericarps in a replicate into an individual pot Table C: Experiment Balance. Soil and watered them daily with deionised water in an Crack Apron
attempt not to produce seedlings. My experimental balance is shown in Table C and my
approach for statistically analysing the unexpected 13 17
Replicates Replicates (Samples) emergence of seedlings appears in Section 3.8.4. Recovered
Page 17 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
3.6 Impact of Smoke and Weathering on Germination
To investigate the impact of smoke on the germination of V. staminosa seeds I performed the following fully randomised and balanced experiment to examine both number of seedlings that emerged and the time taken for seedlings to emerge. For this experiment I decided that the four treatments should be weathering the pericarps, weathering and smoke treating the pericarps, exposing the pericarps to smoke, and a control group. I randomly selected 360 individual seedling pots and ten seedling trays from glasshouse stocks and pre-treated these by the method detailed in Section 3.4.
I randomly selected pericarps from the composite representative sample collected by Dr Yates and Dr Ladd, removed petals and sepals by the method described in Section 3.3, and weighed them to identify pericarps which were likely (α=0.05) to contain a viable seed. From this batch of 500 pericarps I concurrently and randomly selected four sub-samples of 90 pericarps each.
I randomly selected two sub-samples of 90 pericarps for artificial „weathering‟. Adapting techniques from the literature on artificially hydro-priming/weathering fruits/seeds, I performed 30 repetitions of agitating and soaking the sub-samples in deionised water for one hour and then air drying them in a thermostat controlled oven at 30oC (to avoid heat-shocking the seeds) for two hours (Leopold 1964; Mott 1974, Merritt et al. 2007; Cousens et al. 2010). Following the instructions for the aqueous smoke treatment I obtained from Kings Park and Botanic Garden, I randomly selected one of the weathered sub-samples and one of the unweathered sub-samples soaked each in a mixture of the smoke treatment and deionised water for one hour prior to planting. To keep the number of treatments consistent, I also soaked the remaining two sub-samples in deionised water for one hour. After planting all the individual pericarp into the pots, I randomly assigned nine pots of each treatment to each of the ten seedling trays in a 6x6 matrix. The experimental balance is given in Table D.
Conducting this experiment during winter in Perth, which has a Mediterranean climate similar to Wongan Hills, I placed the ten full trays on a single bench in Murdoch University‟s Glasshouse 2. I inspected the experiment each evening and for every new seedling I again recorded the number of days since planting, pot identification, tray number, and treatment until four weeks after the last seeding emerged. Every second day I watered the pots with deionised water and
Page 18 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
randomly repositioned every tray on the bench to minimise the possibility of tray position confounding the outcome of the experiment. My approach to statistically analysing the outcomes of this experiment is detailed in Section 3.8.5.
Table D: Balance of the Glasshouse Experiment with Four Treatments Treatments Total Smoke and Fruit in Smoke Weathering Control Weathering Tray Tray 1 9 9 9 9 36 Tray 2 9 9 9 9 36 Tray 3 9 9 9 9 36 Tray 4 9 9 9 9 36 Tray 5 9 9 9 9 36 Tray 6 9 9 9 9 36 Tray 7 9 9 9 9 36 Tray 8 9 9 9 9 36 Tray 9 9 9 9 9 36 Tray 10 9 9 9 9 36 Total Fruit in 90 90 90 90 360 Treatment
3.7 Germination Field Surveys
As with the longevity of seedling viability discussed above, the experimental design for collecting the field data I analysed was developed and implemented by Dr Yates in collaboration with Dr Ladd.
In winter of the anomalously wet year of 1999 (England et al. 2006; Nicholls, et al. 2001), between nine and fifteen 1mx1m quadrats were established in each of four subpopulations (A, B, G, and W) of V. staminosa (Table E). The quadrats were positioned to be representative of the subpopulation, but all were positioned to contain at least one crack that contained seedlings and where possible they also contained a section of moss mat with seedlings. In spring of subsequent years, except 2004 when Dr Ladd was on sabbatical, the number of seedlings surviving from the previous year and the number of new seedlings has been recorded for cracks and moss mat within each quadrat. I have participated in these seedling surveys since spring 2007.
Table E: Balance for Field Germination Experiments. Subpopulation Subpopulation Subpopulation Subpopulation
A B G W Number of 10 9 13 15 Quadrats
Page 19 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
My consolidation of the field records produced an analysable data set for field germination covering 11 years with 47 replicates distributed across four subpopulations for germination.
3.8 Statistical Analysis
3.8.1 General Approach to Statistical Analysis
Unless I state otherwise, all descriptive statistics in my thesis are reported as the mean ± 95% confidence Highest Value interval for the metric in question. Due to the relatively small sample sizes in my experiments, I calculated 95% 3rd Quartile Median confidence intervals by determining the value of the t- statistic for the appropriate degrees of freedom and
1st Quartile applied the calculation in Box 1 (Berenson et al. 2006).
95% Confidence Interval for the Mean
μ ± tcritical x Standard Error Standard Deviation = μ ± t x critical √ Sample Size (n)
Such that P(t ≤ t ) = 0.975 (Upper tail area = 0.05/ = 0.025) 0.05,(n-1) critical 2
Box 1: Calculating the 95% Confidence Interval. Lowest Value
I performed an exploratory analysis for data from all Figure 2: The 5 critical experiments by graphing histograms and Box & Whisker values for Box & Whisker Plots used in this study. Plots (Fig. 2). For data that was normally distributed or could be normalised, I graphed means with 95% confidence intervals and tested hypotheses using parametric tests, such as the t-test and two-way ANOVA (Berenson et al. 2006).
Due to the binomial nature of the field and glasshouse experiments (i.e. seedling present/not present), there were a high number of „0‟ results generated by seedlings not emerging in all experiments. This prevented this data being normalised by any means, including the Log(x+1) transformation. To test hypotheses about relationships for these data sets, including the impact of climatic variables, I used logistical regression techniques. This application of the General Linear Model makes no assumptions about the nature of predictor variables (i.e. categorical/factor or covariate) or how they are distributed (Berenson et al. 2006; Johnson and Wichern 2002; Tabachnick and Fidell 1996). Predictor variables (or the associated residuals) do not have to be normally
Page 20 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa distributed, linearly related, or of equal variance in each group (Tabachnick and Fidell 1996). Also, unlike analyses based on the normal distribution, or approximations to it, logistical regression does not assume that predictor values of less than zero have positive probabilities (Tabachnick and Fidell 1996; Hillborn and Mangel 1997). Juliano (1993) recommends logistical regression for investigating survivorship “under different, naturally occurring conditions”.
I performed all exploratory data analysis and parametric statistical testing using Microsoft® Excel 2003 (Microsoft Corporation 2002) and I used Minitab 15.0 (Minitab 2007) for the logistical regression analyses.
3.8.2 Interpretation of Logistical Regression
In the first instance, I assessed the overall significance of each logistical regression model by examining the Log-Likelihood for a model to determine if it was significantly different from Null Effect or Constant Model, which is applicable if none of the hypothesised predictors significantly affects the outcome of an experiment. I also assessed the quality of the model through the use of Delta Pearson χ2 plots for event probability and the leverage of each factor/covariate pattern (Berenson et al. 2006; Tabachnck and Fidell 1996; Mintab 2007).
In developing „model webs‟ to explore the significance of hypothesised predictor terms and their relationship to other predictors I used the calculation in Box 2 to determine the ranking of nested Models (Tabachnck and Fidell 1996).
χ2 = 2[(log-likelihood for larger model) – (log-likelihood for smaller model)] df = (number of Predictors in larger model) – (number of Predictors in smaller model)
Box 2: Calculating χ2 to compare nested Models.
2 In the example model web shown in Figure 3, where the χ Test value is greater 2 than the χ Critical value the larger model is superior and the term absent from the smaller model is a significant predictor (i.e. P-value <0.05). This is illustrated by the arrow pointing to the superior model. If none of the terms assessed by the logistical regression is a significant predictor of the outcome (i.e. P-values >0.05), then the null or constant model applies. Care needs to be taken when interpreting P-values≈0.05 (0.03-0.06), as illustrated by the double ended arrow in the top left of the Model Web (Berenson et al. 2006). The odds ratio is calculated by dividing the probability of a success by the probability of failure, that is odds ratio = P(Success)/(1-P(Success)). For this example Predictor 2 is significant and the optimum model is Ln(odds ratio) = constant + β1Predictor 2.
Page 21 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Ln (odds ratio) = constant + Predictor 1 + Predictor 2 + Interaction (P1*P2)
χ2 = 2.44 P-value = 0.107
Ln (odds ratio) = constant + Predictor 1 + Predictor 2
χ2 = 4.092 χ2 = 2.40 P-value = 0.043 P-value = 0.121
Ln (odds ratio) = constant + Predictor 1 Ln (odds ratio) = constant + Predictor 2
χ2 = 6.466 P-value = 0.891 χ2 = 133.4 P-value = <<0.001 χ2 = 2.374 P-value = 0.499
Null/Constant Model Figure 3: An example of a „Model Web‟ produced from a logistical regression analysis.
When performing stepwise logistical regression to model the impact of climatic variables on germination and seedling survival in the first year, I also considered the individual Z-scores, P-values, and odds ratios for each predictor. I utilised the Deviance statistic, which measures the fit of a hypothesised model compared to the fit of the saturated model, to assess the predictive ability of models with multiple predictors (Berenson et al. 2006).
3.8.3 Seed Viability of the Soil Seed Bank
Sample sizes for this experiment were relatively small, especially in relation to the number of pericarps that contained a viable seed, however my exploratory data analysis suggests, although slightly skewed, the distribution of pericarps with and without a viable seed is approximately bell-shaped (Appendix A2.1). I therefore applied the Central Limit Theorem and assumed that means and proportions of values determined for each trial approximate either a normal distribution N(μ, σ2) or binomial approximation to the normal distribution N(np, np(1-p)) respectively (Berenson et al. 2006).
I determined the mean mass of pericarp with and without viable seeds, and calculated the 95% and 99% confidence level for both means. From this analysis I determined that selecting cleaned pericarp with a mass of greater than 7.5 mg maximised the likelihood that pericarps used in germination trials contained a
Page 22 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa viable seed (Section 4.2). I used the Chi-Square (χ2) Test for difference between more than two proportions to compare the number of pericarps with viable seeds across the three trials (Berenson et al. 2006). With no difference between the proportions for the individual trials (see Section 4.2), I combined the data for the three trials to calculate the overall Proportion (p) ± t0.05,2 x Standard Error proportion of pericarps that contain a p(1-p) = p ± t x viable seed, the standard error, and the 0.05,2 √ Sample Size (n) 95% confidence interval for the proportion where t = 1.9799 0.05,2 using the calculations shown in Box 3 Box 3: Calculating the Standard Error (Berenson et al. 2006). and the 95% Confidence Interval.
3.8.4 Longevity of Seed Viability
As I expected, exploratory data analyses suggest that the number of seedlings emerging in replicates recovered from cracks or soil aprons follows a Poisson distribution (Appendix A2.2) and cannot be analysed by parametric statistical tests. However, the time taken for seedlings to emerge in recovered replicates loosely approximated a normal distribution. Given its robustness and the sample size involved, I used the t-Test for common variances to test the null hypothesis of no difference in the mean emergence time (Berenson et. al. 2006).
For the purpose of this analysis I assumed the initial seed viability of the buried pericarps was 32.5% (justified in Section 4.2). I multiplied the number of recovered pericarps from each replicate by this proportion to determine the „original viability‟ for each replicate and then calculated the proportion of seedlings that emerged. While the data set was relatively small, my exploratory data analyses suggested this binomially distributed data also provided a reasonable approximation to the normal distribution, that is follow a bell-shaped curve (Berenson et. al. 2005). However, the combination of small sample sizes and an apparent bias toward seeds no longer being viable meant that I could not apply the test of two proportions based on the normal approximation to the binomial distribution, so I used the Chi-Square Test for Proportionality to investigate the null hypothesis that burial location was not a factor that impacted long term seed viability (Moore and McCabe 1993; Berenson et al. 2006). Having obtained indicative results from the χ2 Test, I further investigated the veracity of my findings by applying a statistically more powerful and robust logistical regression analysis (Tabachnck and Fidell 1996).
Page 23 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
3.8.5 Impact of Smoke and Weathering on Germination
My exploratory data analyses show a high degree of variability and skew in the distributions for both the number of seedlings that emerged and the time it took seedlings to emerge. With no viable transforms for normalising the data, I applied logistical regression to determine if any relationships exist in relation to the dormancy breaking treatments (i.e. smoke and/or weathering) and/or the experimental design (i.e. Tray effect). The outcomes of the exploratory data analyses and the Logistical Regression are given in Section 4.4.
3.8.6 Analyses for Climate Data
From the Australian Bureau of Meteorology (2011) Climate Data Online I obtained daily observations from the Wongan Hills Meteorological Station (Station 008137) for precipitation (hereafter rainfall), maximum and minimum temperatures, and solar radiation from 1991 to 2010. From these values I calculated the key metrics for germination (Section 2.4.1) listed in Table F.
Table F: Calculation of climate metrics shown to impact germination & seedling survival. Metric Method of Calculation Rainfall Total rainfall recorded for the calendar year (annual (mm) rainfall) or season. Number of Rainy Days Number of days in the calendar year or season when recorded rainfall was greater than or equal to 1mm. Rainfall Intensity (mm) Usually amount of rainfall per event, but here calculated as Average Rainfall per Rainy Day. Average Monthly Average of the Monthly Averages across the period of Solar Radiation interest (i.e. annual or season). Monthly Averages (MegaJoules per square calculated by averaging the incident solar radiation metre – MJ/m2) for each day of each calendar month. Average Monthly Maximum daily temperatures were averaged within Maximum Temperature each calendar month and means of these Monthly (Celsius – oC) Averages are calculated for the periods of interest. Average Monthly Minimum daily temperatures were averaged within Minimum Temperature each calendar month and the means of these Monthly (Celsius – oC) Averages were calculated for the periods of interest. Average Monthly Daily Average Temperatures were calculated by Temperature averaging the Maximum and Minimum Temperature (Celsius – oC) for each day. These values were averaged within each calendar month and the means of these Monthly Averages were calculated for the periods of interest. Average Monthly Diurnal Each day‟s Minimum Temperature was subtracted Temperature Range from the Maximum. These values were averaged (Celsius – oC) within each calendar month and the means of these Monthly Averages were calculated for each period.
Page 24 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
I use these metrics to investigate relationships for the weather of this 20 year period in terms of the annual cycle and the three types of seasons (Section 2.4.2) that have been observed to operate in the SWBP.
Initially I used histograms and Box & Whisker Plots to investigate variability between annual weather cycles and to look for evidence of the three season models reported to operate in the SWBP, which is a Mediterranean climate (Dry Season/Wet Season); the traditional Four Seasons of summer, autumn, winter, and spring; and the Six Noongar Seasons. My analyses revealed that all the annual climate metrics provided a reasonable approximation to the normal distribution. This allowed me to assess inter-year variability by comparing the annual values of each metric to its long-term (1991-2010) mean using a Chi- Squared Analysis (Berenson et al. 2006). At the finer season scale there was a high degree of variability and highly skewed distributions for the rainfall related metrics, however the solar radiation and temperature based metrics were distributed as a reasonable approximation to the normal distribution (Appendix C). This being the case I utilised t-Tests for uncommon variance, which are sufficiently robust to deal with skewed data, to analyse between season variability for the three rainfall metrics under a Mediterranean climate (Berenson et al. 2006). These analyses revealed significant inter-seasonal variation for all three metrics (Section 4.5.2) and as the distributions for these metrics under the other two season models were similarly skewed, I assume the inter-seasonal variation to be significant for all three types of seasons. With the solar radiation and temperature metrics approximating normal distributions for all season models, I utilised two-way ANOVAs without repetition to investigate the significance of inter-season and inter-year variability for these metrics. I used bi-variate least-squares linear regression to investigate collinearity between those metrics within each season that approximated a normal distribution (Philippi 1993). The outcomes of these analyses are all reported in Section 4.5.
3.8.7 Impact of Climatic Variables on Germination in the Field
Given that V. staminosa has been present on granite outcrops in the Wongan Hills area for at least the past 1.5 million years and the highly restricted range of the extant meta-population, I assumed that competition and edaphic variables were largely constant over the restricted geographic and temporal range within which the data I analyse was collected (Gardner 1959; Hopper and Gioia 2004; Yates and Ladd 2004; Yates et al. 2007; Hopper 2009).
Page 25 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Yet again my exploratory data analyses revealed that parametric techniques are not suitable for analysing field germination. I investigated the presence of any significant relationships by using logistical regression (Section 4.5). I analysed this experiment by developing webs of models and applying Stepwise Logistical Regression, with guidance from the correlation analysis described in Section 3.8.6, to ascertain how various climate metrics may have influenced germination of V. staminosa seedlings in crevices within the granite pavement at Mocardy Hill since 1999 (Tabachnck and Fidell 1996; Ramsey and Schafer 2002).
These logistical regressions produced a suite of models for the influence of climatic factors on the germination of V. staminosa seedlings under a Mediterranean climate, the tradition four seasons, and the six Noongar seasons. I applied Burnham and Anderson‟s (2002) approach to Akaike‟s Information Criterion (1973; cited in Burnham and Anderson 2002) to ranking statistical models. This allowed me to identify a „top model‟ from the suite of germination models, and to explore the strength of evidence provided by lesser models in comparison to the top model. The calculations for Akaike‟s Information Criterion (AIC) are given in Box 4. The outcome of this ranking process is reported in Section 4.6.5
Akaike‟s Information Criterion (AIC)
AIC = - 2 Log-likelihood + 2K Model Probabilities (ωi) Where K is the sum of the number of terms for the intercepts, the variances, and the predictors. exp(- ½ ∆i ) ωi = AIC Differences (∆ ) i ∑ exp(- ½ ∆i ) ∆i = AICi – AICmin
Best model (i.e. lowest Log-likelihood) has ∆i ≡ ∆min ≡ 0. Box 4: Burnham and Anderson‟s (2002) approach to calculating AIC values.
Page 26 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 4: Results
4.1 Exploratory Data Analyses
My exploratory data analyses using Histograms and Box & Whisker Plots revealed that many of my raw data sets were skewed and only a few approximated the “bell-shaped” curve of the normal distribution (Appendices A-C).
4.1.1 Data Sets that Approximate a Normal Distribution
From the field and glasshouse germination experiments the mass of pericarp with and without viable seed, the time taken for seedlings to emerge, and the proportion of „originally viable seeds‟ that germinated from the soil stored pericarps generally approximated the normal distribution or could be normalised. This allowed me to directly analyse those datasets using parametric statistical tests. To compare the mass of pericarps with and without a viable seed, I needed to combine the means of three trials and apply the Central Limit Theorem (Section 3.8.3).
The annual climate metrics all approximated the normal distribution (Appendix C1). However, at the finer seasonal scale only the temperature metrics approximated the normal distribution (Appendices C2 to C4). This outcome is predicted by the Central Limit Theorem as the temperature metrics are determined by sets of sample means.
4.1.2 Data that Can Not Be Normalised by Transformation
Histograms and Box & Whisker Plots for the number of seedlings that emerged in the germination experiments for both the soil buried pericarps and the impact of smoke and weathering on seeds appeared to follow a Poisson distribution rather than the normal distribution (Appendices A2 and A3). Unlike the germination of recovered pericarps experiment, the portions of seedlings that germinated in the smoke and weathering experiment is also highly skewed. Unlike the time taken for seedlings to emerge from the soil stored pericarps, surprisingly the average time taken for seedlings to emerge from the smoke treated and weathered pericarps (i.e. for each treatment in a tray) was also too skewed to analyse directly.
To investigate variability of seedling germination between years I investigated if a survey quadrat contained seedlings (i.e. presence/absence), the proportion of quadrats in a subpopulation that contained seedlings, and the average number of seedlings per quadrat within each subpopulation (Appendix B). These
Page 27 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa exploratory analyses reveal that the distribution of data for all three metrics is highly skewed. This is even the case for the distribution of sample means for the number of seedlings, which I had expected to be normally distributed in accordance with the Central Limit Theorem.
The skew in these data sets is so great that even after Log transformation they do not approximate a normal distribution, so I did not conduct further parametric analyses on these metrics.
4.2 Seed Viability of Soil Seed Bank
After applying the Central Limit Theorem (Section 3.8.3) to smooth and normalise the data from each of the three trails, I calculated the means and associated 95% and 99% confidence intervals for pericarps containing a viable seed and those that did not (Figure 4). From this analysis I decided that selecting cleaned pericarps with a mass of greater than 7.5 mg maximised the likelihood that fruit used in my germination experiments contained a viable seed.
10 33
9
8
7
6 32.5
5 (milligrams)
4
with Viable Seed Viable with MassofPericarps
3 PercentageofPericarps
2 32 Viable No Seed Viable No Seed % % Viable Seed 95%CI Seed 99%CI Viable± ± 95%CI 95%CI 99%CI Std Error
Figure 4: Mean mass (± 95% & 99% confidence intervals (CI)) of pericarp with and without a viable seed and overall seed viability (± SE & 95% CI).
Assuming that the masses of cleaned pericarps with and without a viable seed approximated the normal distributions N(8.0 mg, 0.2 mg) and N(5.5 mg, 0.4 mg) respectively, then the probability that I planted a pericarp without a viable seed P(X ≥ 7.5 mg) was very small (3.0 x 10-7).
With approximately one in three pericarps in each of the three trails containing a viable seed, it was not surprising that the proportions of viable seeds I found in each of the 3 trails were not significantly different from each other (χ2 = 0.926,
Page 28 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa p-value = 0.630, DF = 2). My estimate of 32.5 ± 0.2% (p ± Std Error) for the overall proportion of pericarps that contained a viable seed is in good agreement with the values (mean ± SE) of 32.0 ± 6.1% obtained by Cochrane et al. (2001) and the 3 trials by Yates et al. (2007) which produced values (mean ± SE) of 39.92 ± 2.42%, 35.50 ± 2.64%, and 30.59 ± 2.18%.
4.3 Germination Response of Recovered Pericarp
Applying the statistically robust t-Test (Section 3.8.4) to compare the time for seedlings to emerge from the recovered pericarps provides strong evidence (t = 1.655, P-value = 0.562, DF = 146) that there is no difference in the average time for seedlings to emerge from pericarps that were buried in the cracks (34.0 ± 0.8 days) and those buried in the soil aprons (34.3 ± 0.8 days).
12 My analysis of the proportion of „originally
10 viable‟ seeds that produced seedlings (Section 8 3.8.4), revealed there is not a significant
6 difference in the proportions of seedlings
4 Number ofSeedlings Number 2 emerging from replicates buried in a crack or
0 from those buried in the soil apron (Table G). Buried in Cracks Buried in Aprons There may be a difference between replicates Figure 5: Seedling emergence in recovered pericarp replicates recovered from the cracks and the replicates recovered from the soil apron. However, this result is at odds with the graphical evidence (Figure 5), which is supported by the logistical regression model
Ln(odds ratio) = constant + β1Crack producing a Log-Likelihood of -334.967 (χ2 = 0.997; P-value = 0.318; DF = 1). This provides strong evidence that burial location does not significantly impact seed longevity.
Table G: Outcomes of Chi-Square Tests for differences between proportions of seedlings within treatments and between treatments (Buried in Cracks or in Soil Aprons). Degrees of χ2 Test Statistic P-value Freedom Within Cracks 4.62 x 10-122 ≈ 1 12 Within Soil 5.27 x 10-67 ≈ 1 16 Aprons Between Cracks 8.55 0.003 1 and Soil Aprons The median (and range) for the number and the proportion of originally viable seeds that germinated in each replicate was 3 (0-11) seedlings at 10.2% (0%- 35.5%). Normally distributed by the Central Limit Theorem, the mean of the mean emergence time per replicate was 34.8 ± 0.4 days.
Page 29 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
4.4 Impact of Smoke and Weathering on Germination
Mintab 15.0 allows categorical data such as the four treatments of Smoke & Weathering, Weathering, Smoke, and Control to be grouped together and modelled under single „predictor‟ term (e.g. Treatment). Despite having 90 replicates of each treatment, my data set was too small to analyse the interaction between the Tray that a seedling germinated in and the Treatment it received. However, my logistical regression analyses (Figure 6) provides strong evidence that neither predictor is significant, the Constant Model appears the most valid, and it is unlikely that the interaction term (Treatment*Tray) on its own would constitute the “Perfect Model” (Tabachnick and Fidell 1996).
Ln (odds ratio) = constant + Treatment + Tray + Treatment*Tray
Ln (odds ratio) = constant + Treatment + Tray
χ2 = 4.092 χ2 = 2.40 P-value = 0.043 P-value = 0.121
Ln (odds ratio) = constant + Treatment Ln (odds ratio) = constant + Tray
χ2 = 6.466 P-value = 0.891
χ2 = 2.374 χ2 = 4.065 P-value = 0.499 P-value = 0.907
Null/Constant Model Figure 6: Web of models for effect of Tray and Treatment.
The median (and range) for the number and proportion of seedlings emerging in each tray was 21 (17-24) and 58.3% (47.2%-66.7%). Normally distributed by the Central Limit Theorem, the mean emergence time between treatments were not significantly different, but the overall mean emergence time of 45.6 ± 0.7 days is significantly longer than the 34.8 ± 0.4 days for the pericarps recovered from the cracks and soil apron (t=9.48; P-value<<0.001; DF = 10 & 12 respectively).
4.5 The Seasons of South-West Western Australia
Exploratory data analyses outlined in Section 4.1 provides evidence that the Mediterranean style Dry and Wet seasons; the four seasons of summer, autumn, winter, and spring; and the six Noongar seasons apply in the Wongan Hills area.
Page 30 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
4.5.1 Annual Climate Metrics
Chi-Squared analyses of the annual means of each metric against the 20 year (1991-2010) long-term mean value revealed that annual rainfall and the number of rainy days were the only metrics with significant between year variability (Table H).
Table H: Outcomes of Chi-Square Tests for difference between annual values and the 20 year (1991-2010) long-term averages for key climate metrics.
Metric Statistics Metric Statistics
2 2 χ >> 50 Annual Avg Monthly χ = 0.736 Annual Rainfall P-value << 0.001 Temperature P-value ≈ 1
2 2 χ = 40.75 Annual Avg Monthly χ = 0.736 Annual Rainy Days P-value = 0.002 Max Temperature P-value ≈ 1
2 2 χ = 2.226 Annual Avg Monthly χ = 0.736 Annual Intensity P-value ≈ 1.000 Min Temperature P-value ≈ 1
2 2 Annual Avg Monthly χ = 0.736 Annual Avg Monthly χ = 0.736 Solar Radiation P-value ≈ 1 Diurnal Variation P-value ≈ 1
Not unexpectedly amount of Annual Rainfall and the number of rainy days are very highly correlated (R2 = 61.8%; F = 29.06; P-value = <0.001). Annual intensity of rain is not correlated to the number of rainy days and is only weakly correlated to annual rainfall (R2 = 28.1%; F = 7.046; P-value = 0.016). However I applied Chatterjee and Price‟s (1991, quoted in Philippi 1993) axiom that an F-value of greater than 10 “is … a signal that the data have collinearity problems”, meaning this weak correlation between annual rainfall and Annual intensity is not significant. Also as might be expected, there is a high degree of cross correlation for solar radiation and the four temperature metrics (Appendix D). Somewhat surprisingly annual monthly average maximum temperatures are only collinear with the solar radiation metric.
The high degree of cross correlation of diurnal temperature variation with solar radiation, average monthly temperatures, and minium temperatures and the low cross correlation of the maximum temperature metric, combined with the fact that those two metrics (i.e. diurnal temperature variation and maximum temperature) have been shown to impact germination (Section 2.7.1), I investigated these two metrics and disregarded solar radiation (Section 4.6.1).
Page 31 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
4.5.2 Mediterranean (Dry and Wet Season) Metrics As previously mentioned, the rain metrics for the Dry and Wet seasons are skewed (Appendix C2). Skewed distributions such as these, which can not be assessed by parametric methods, are indicative of outliers and indicate a high degree of variability (Tabachnick & Fidell 1996), in this case inter-year variability. With only two data sets, I was able to compare inter-season variability using the robust t-Test for uncommon variance (Section 3.8.6). I discovered that there is significant difference between seasons for the three rain related metrics (Table J).
Table J: Outcomes of t-Test with uncommon variance of between seasons. Rain Metric t-Statistic P-values Amount of Rainfall 8.848 <<0.001 Number of Rainy Days 14.28 <<0.001 Rain Intensity 2.681 0.014
ANOVA analyses of the temperature and solar radiation metrics (Section 3.8.6) revealed while all the temperature metrics are significantly different between the dry and wet seasons, only diurnal temperature range and solar radiation show significant (P-value <0.05) inter-year variability (Table K). Testing the collinearity of the diurnal temperature range and solar radiation I found a high level of correlation in both the Dry Season (R2 = 38.9%; F = 17.06; P-value = <0.001) and the Wet Season (R2 = 48.7%; F = 11.49; P-value = 0.003). Again, as diurnal temperature range impacts germination I investigated this metric further and disregarded the solar radiation (Section 4.6.2).
Table K: Variability in the temperature metrics under the Mediterranean Season‟s model. Temperature Inter-season Inter-year Metric Variability Variability Solar F = 4726 F = 5.194 Radiation P-value <<0.001 P-value <0.001 Average Monthly F = 2692 F = 0.897 Temperature P-value <<0.001 P-value = 0.593 Average Monthly F = 1253 F = 1.150 Maximum Temperatures P-value <<0.001 P-value = 0.381 Average Monthly F = 2224 F = 1.550 Minimum Temperatures P-value <<0.001 P-value = 0.381 Average Monthly F = 385.0 F = 2.689 Diurnal Range P-value <<0.001 P-value = 0.018
Page 32 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
4.5.3 Metrics for the Four Seasons Climate
As with the Mediterranean Seasons in the previous section, the rainfall metrics were skewed suggesting a high degree of inter-year variability (Appendix C3). Similarly my ANOVA analyses of the temperature metrics show a significant inter-season variability for all the metrics, but only the diurnal temperature range and solar radiation had significant inter-year variability (Table L). This degree of variability, as a result of extreme (outlier) events, is required to successfully model climatic phenomena (England et al. 2006).
Table L: Variability in the temperature metrics under the Four Season‟s model. Temperature Inter-season Inter-year Metric Variability Variability Solar F = 4726 F = 4.226 Radiation P-value <<0.001 P-value < 0.001 Average Monthly F = 1140 F = 0.992 Temperature P-value <<0.001 P-value = 0.484 Average Monthly F = 1013 F = 1.601 Maximum Temperatures P-value <<0.001 P-value = 0.088 Average Monthly F = 751.6 F = 1.392 Minimum Temperatures P-value <<0.001 P-value = 0.168
Average Monthly F = 210.7 F = 3.331 Diurnal Range P-value <<0.001 P-value < 0.001
My analyses of the collinearity of the diurnal temperature range and solar radiation metrics also showed a high degree of correlation for Summer (R2 = 48.7%; F = 40.81; P-value < 0.001), Autumn (R2 = 41.6%; F = 12.84; P-value = 0.002), Winter (R2 = 39.8%; F = 11.92; P-value = 0.003), and Spring (R2 = 34.5%; F = 9.480; P-value = 0.006).
4.5.4 Six Noongar Seasons
Similar to both of the previous season models, the rainfall metrics are skewed for the Noongar seasons, which again suggests a high degree of inter-year variability (Appendix C.4). As with my ANOVA analyses for the Mediterranean climate and the traditional four seasons, there is a significant inter-season variability for all the temperature metrics (Table M). A difference with this analysis is that the maximum temperature metric shows significant inter-year variability in addition to the diurnal temperature range and solar radiation metrics. This is the same outcome I observed in my analyses of the annual climate values (Section 4.5.1).
Page 33 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Table M: Variability in the temperature metrics under the Six Noongar Season‟s model. Temperature Inter-season Inter-year Metric Variability Variability Solar F = 1235 F = 4.303 Radiation P-value <<0.001 P-value <0.001 Average Monthly F = 969.2 F = 1.234 Temperature P-value = <<0.001 P-value = 0.249 Average Monthly F = 916.7 F = 2.115 Maximum Temperatures P-value = <<0.001 P-value = 0.009 Average Monthly F = 593.8 F = 1.614 Minimum Temperatures P-value = <<0.001 P-value = 0.068 Average Monthly F = 189.8 F = 4.288 Diurnal Range P-value = <<0.001 P-value <0.001
Across the six seasons collinearity of the diurnal temperature range, maximum temperature, and solar radiation was highly variable; however there is a significant amount of cross-correlation between these metrics in most seasons (Appendix E). On this basis, I considered the impact of the diurnal temperature range on germination (Section 4.6.4), while disregarding both the maximum temperature and solar radiation metrics with which it is highly correlated.
4.6 Impact of Climatic Variables on Germination in the Field
Because of the relatively small sample size I am unable to determine if there is a difference in germination pattern between quadrats or if the interaction between Subpopulation and Year is significant. However, logistical regression on the presence or absence of seedlings in a quadrat revealed that there is no difference between the four Subpopulations, but year was significant as a factor (Figure 7). Given the strength of the evidence against Subpopulation being a significant predictor, it is reasonable to assume that the interaction term (Subpopulation*Year) is also unlikely to be significant.
Ln (odds ratio) = constant + Subpopulation + Year
χ2 = 133.4 χ2 = 2.454 P-value = <<0.001 P-value = 0.117
Ln (odds ratio) = constant + Subpopn Ln (odds ratio) = constant + Year
χ2 = 137.7 P-value = <<0.001
χ2 = 1.864 χ2 = 135.2 P-value = 0.601 P-value = <<0.001
Null/Constant Model Figure 7: Germination models with Subpopulation and Year as factors.
Page 34 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
As I found no significant variation in germination between the subpopulations and given the persistence of V. staminosa near Wongan Hills for at least the past 1.5 million years, I hypothesise that the difference in germination between years is linked to variability of the weather between seasons. The analyses below explore the impact of climatic variables based on annual values and at the finer scale of two, four, and six seasons.
4.6.1 Annual Values
Initially I analysed the impact of climate metrics that I determined to be important in Section 4.5.1 through a standard stepwise logistical regression. This involves a series of iterations that remove the least significant term until only significant predictors remain (Section 3.8.2) producing the optimum model:
Ln(odds ratio) = 0.650 + 0.001 Rainfall + 0.138 Intensity.
While this model is significantly better than the null effect model (χ2 = 80.49; DF = 2; P-Value < 0.001), the constant term is not significant (P-value = 0.847) and the deviance of values predicted by the model from those observed in the 2 field is significant (H0: Model is good fit to data; χ = 54.75; DF = 8; P-Value <0.001). This means that this model is not a good fit to the data and can not be used to make any predictions about germination of V. staminosa seedlings in the monitored quadrats.
I created a web of models to further investigate models that may allow the germination of seedlings to be explained by annual values of the climate metrics. This produced two possible models. The first suggested that the number of rain days alone influenced germination, while the second model suggested that a combination of the metrics Rainfall, Rainy Days, Intensity, Cyclones (and tropical lows) and Diurnal Temperature Range influence germination. However, the deviance of predicted values from observed germination was excessively high (P-value <0.001) for both models and at the 99% confidence level the larger model may not be any better than the model based just on Rainy Days (χ2 = 16.60; DF = 5; P-value = 0.012). Subpopulation was not a significant predictor in any model; which allows the data to be analysed as a single large set.
These analyses provide evidence that while a complex interplay of climatic variables influences germination, aggregating climatic variables as annual values (totals or averages) is too coarse to elucidate their effect on germination.
Page 35 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
4.6.2 Mediterranean Climate
There were insufficient degrees of freedom (i.e. large enough sample) available for me to investigate differences between subpopulations in this analysis. However as sub-population had not been shown to be significant in either of the previous models above, it was again unlikely to be a significant predictor of germination.
Disregarding Subpopulation for this (and all subsequent) model(s) I performed stepwise logistical regression. I started the analysis with a model that had the four rainfall metrics (Rainfall, Rainy Days, Rain Intensity, and Cyclones) and Diurnal Temperature Range. From this model I removed the least significant predictors (i.e. lowest p-value) one at a time. This produced in the optimum model:
Ln(odds ratio) = 28.0 – 0.01 Dry Season Rainfall + 0.16 Dry Season Intensity – 0.43 Wet Season Intensity – 2.06 Wet Season Average Monthly Diurnal Temperature Range.
This model is significantly better than the null effect model (χ2 = 126.267; DF = 4; P-value <0.001), the values predicted are a good fit with the values generated for the saturated statistical model with all predictors included (χ2 = 8.981; DF = 6; P-value = 0.175). This model provides a reasonable prediction of germination in the field, providing evidence that the bimodal climate of SWWA influences germination of V. staminosa.
4.6.3 The Traditional Four Seasons
Again, the problem of insufficient degrees of freedom meant that I could not analyse the model with four rainfall metrics (Rainfall, Rainy Days, Rain Intensity, and Cyclones) and Diurnal Temperature Range with twice as many (i.e. four) seasons. To determine if I could detect any relationships between climatic variables and germination under the Four Seasons climate I initially investigated the influence of rainfall and number of rain days. These metrics show significant inter-year variability for their annual values, had the greatest significance in my analyses of the annual metrics, and are suggested as being the most important in the literature (Cowling et al. 2005; Fay 2009). All seasons were significant when I logistically modelled both Rainfall and Rainy Days in isolation. By combining all the data for both these metrics and performing a standard stepwise logistical regression I generated the germination model:
Page 36 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Ln(odds ratio) = -5.01 – 0.02 Summer Rainfall – 0.01 Winter Rainfall + 0.02 Spring Rainfall + 0.19 Summer Rainy Days + 0.11 Autumn Rainy Days +0.20 Winter Rainy Days.
This model is significantly better than the null effect model (χ2 = 130.8; DF = 6, P-value <0.001) and does not deviate significantly from values predicted by the saturated model (χ2 = 4.349; DF = 4; P-value = 0.361). This model will provide a reasonable prediction of germination in the field.
Having found the most significant season predictors for Rainfall and Rainy Days, I attempted to add Intensity and Cyclones to the model. Again the available degrees of freedom were less than the requirements of this model. I chose not to model Summer Cyclones, which were not significant in the optimum germination model for the Mediterranean climate nor when compared to just Autumn Cyclones under the Four Seasons climate. With Autumn Cyclones included, I created four models each of which had the Rainfall Intensity of a different season left out. I then performed another stepwise regression on the most significant of the four models assessed in terms of the P-values of the predictors, difference from the Null Model as measured by the P-value of the test for all slopes being zero, and the P-value of the Deviance. This produced the following model for germination in the Four Seasons Climate:
Ln(odds ratio) = -0.20 – 0.01 Summer Rainfall + 0.01 Winter Rainfall + 0.01 Spring Rainfall + 0.07 Autumn Rainy Days +0.54 Autumn Cyclones – 0.53 Winter Intensity.
Significantly better than the null effect model (χ2 = 132.2; DF = 6, P-value <0.001) and the values predicted from this model do not deviate significantly from the saturated model (χ2 = 3.014; DF = 4; P-value = 0.555). Hence this model also provides a reasonable prediction of germination in the field.
To determine if this model could be further enhanced I added the diurnal temperature range data and again undertook stepwise logistical regression to produce the optimum germination model:
Ln(odds ratio) = -25.9 + 0.03 Summer Rainfall + 0.30 Autumn Rainy Days +0.87 Autumn Cyclones + 1.19 Autumn Avg Monthly Diurnal Temp Range – 0.88 Winter Avg Monthly Diurnal Temp Range + 0.93 Spring Avg Monthly Diurnal Temp Range.
Page 37 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Again this model is significantly better than the null effect model (χ2 = 134.0; DF = 6, P-value <0.001) and predicted values do not deviate significantly from saturated statistical model (χ2 = 1.208; DF = 4; P-value = 0.877). This model also provides a reasonable prediction of field germination.
4.6.4 Six Noongar Seasons
The available degrees of freedom were taxed even further by the inclusion of the additional two seasons, which introduced an additional eight predictors, required under the six Noongar Seasons. To counter this problem I separately stepwise modelled the impact of the number of Rainy Days and the amount of the Rainfall in each season and then combing these models I again progressively removed the least significant terms. The germination model of Rainfall and Rainy Day predictors produced by this process was not a good fit to the observed data (H0: Model is a good fit to data; χ2 = 12.13; DF = 5; P-value = 0.033). I then added the two seasons with Cyclones and this produced the germination model:
Ln(odds ratio) = -4.53 – 0.49 Bunurua Rainy Days + 0.87 Bunurua Cyclones + 0.37 Djeran Rainy Days – 0.02 Makuru Rainfall + 0.10 Makuru Rainy Days + 0.15 Djilba Rainy Days.
This model is significantly better than the null effect model (χ2 = 133.0; DF = 6; P-value <0.001) and is a good fit to values generated for the saturated model with all predictors included (χ2 = 2.244; DF = 4; P-value = 0.691). As such this model will provide a reasonable prediction of germination in the field under the six Noongar Seasons. With the limited degrees of freedom available this is the optimum model I am able to produce for the Noongar Seasons. With the existing data set Minitab 15.0 could not resolve more complex models that included predictors for the rain intensity and diurnal temperature range.
4.6.5 Ranking the Season Models
Applying Burnham and Anderson‟s (2002) approach to Akaike‟s Information Criterion (1973; cited in Burnham and Anderson 2002) to ranking the logistical germination models reported above (Sections 4.6.2-4.6.4) generated the values in Table O.
This analysis provides evidence that the final model for germination under the Four Seasons (Section 4.6.3), which includes terms for amount of rainfall; number of rainy days; cyclone/tropical lows; and diurnal temperature variation,
Page 38 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa is the „top model‟ and the model for the six Noongar Seasons is next best. At the 95% confidence level the model for the Mediterranean climate has only marginal significance.
Table N: Ranking of logistical models for germination based on Burnham and Anderson‟s (2002) approach to calculating AIC values (Larger the probability the better the model). AIC Model Germination Model AIC Difference Probability Mediterranean Climate 569.92 3.774 0.060 Four Seasons – Rainfall & Rainy Days 569.29 3.142 0.088 Four Seasons – with Intensity added 567.95 1.806 0.172 Four Seasons – with Diurnal Range 566.14 0.000 0.424 Six Noongar Seasons 567.18 1.036 0.252
Page 39 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 5: Discussion
5.1 Longevity of Seeds in Soil Seed Bank
With approximately 10% of the originally viable V. staminosa seeds germinating after being buried for seven years, my study is yet another demonstration of the fact that in Australia pericarp clad hard seeds can persist in the soil seed bank for extended periods of time (Tieu and Egerton-Warburton 2000; Fenner and Thompson 2005; Merritt et al. 2007). For the morphologically similar Verticordia fimbrilepis Yates and Ladd (2005) observed a strong decline in the viability of seeds that had been buried for 30 months and hypothesised that viable seed would remain in the soil seed bank for less than ten years. The results reported in Section 4.3 of this study suggest that V. staminosa seeds may be able to persist in the soil seed bank for longer than ten years, but a more structured long-term study is required to quantify my observations. Being unsure of the persistence of the soil seed bank, the population viability analysis for V. staminosa of Yates et al. (2007) adopted a highly conservative approach by assuming that no soil seed bank was carried over from year to year. My finding of a highly persistent seed bank would further enhance the stability of the Wongan Hills V. staminosa meta-population, supporting the hypothesis of Yates et al. (2007).
My analysis of rainfall over the time that these seeds were buried demonstrated that there were numerous years in which viable seeds could have become imbibed and germinated. The persistence of approximately 10% of the originally viable seeds for seven years after burial is consistent with previous studies of dormancy mechanisms and soil seed banks for hard seeded Australian species (Leishman & Westoby 1998; Fenner and Thompson 2005). It is likely that a combination of physical dormancy imposed by the pericarp and physiological dormancy, whereby imbibed seeds that have not germinated desiccate as the weather become drier and warmer late in the year and return to a state of secondary dormancy, contributes to longevity of V. staminosa seeds in the soil (Fenner and Thompson 2005; Merritt et al. 2007; Cousens et al. 2010).
5.2 Impact of Smoke Water and Weathering on Germination
Starting with Dixon et al. (1995) a great deal of prominence has been given to the role of smoke and solutions of smoke products (i.e. smoke water) in breaking the dormancy of Australian native species, particularly species of the SWBP.
Page 40 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
However as I detailed in Section 2.3.2, in recent years there has been a shift in understanding to the point where smoke products in the soil are viewed as one of the many triggers that promote germination of an imbibed seed (Tieu and Egerton-Warburton 2000; Pickup et al. 2003; Merritt et al. 2007). As previously mentioned (Section 2.3.1), several authors contend that temperature is the primary factor, possibly the only factor, that alleviates seed dormancy (Tieu and Egerton-Warburton 2000; Pickup et al. 2003; Fenner and Thomson 2005; Merritt et al. 2007).
As smoke pre-treatment had been shown to increase germination rates in several species of Verticordia (E.A. George 2002; Yates and Ladd 2004), my soaking of V. staminosa seeds in Kings Park smoke water for one hour was very much related to using smoke to „break the dormancy‟ of V. staminosa seeds. The fact that I found no evidence that smoke treatment alleviates dormancy (Section 4.4) appears consistent with the growing view that smoke does not alleviate dormancy in Australian seeds with a pericarp. Upon reflection Yates and Ladd (2004) applied their smoke water treatment directly to the soil seed bank. This could have facilitated germination of seeds that had their dormancy alleviated and were already imbibed. Additionally, up to a third of seeds in the soil seed bank will germinate, even without the presence of aqueous smoke products (Pickup et al. 2003; Nield et al. 2009).
While not increasing the number of seeds that germinate overall, repeated wetting and drying (hydration) cycles have been demonstrated to increase the rate of seedling emergence (Lush et al. 1984; Pickup et al. 2003; Fenner and Thomson 2005). However, I found no evidence of this effect in my experiment (Section 4.4).
I did observe that the „aged‟ seeds recovered from the cracks and soil aprons on average germinated almost 11 days faster than seeds from the contemporary soil seed bank (Section 4.5). This increased rate of germination could be explained by greater weathering of the seed coat allowing moisture to reach the seed more rapidly, which results in faster imbibition and germination of the non- dormant seed (Lush et al. 1984; Martyn et al. 2009; Cousens et al. 2010)
5.3 Impact of Climatic Variables on Germination
In Section 4.6 I detail a number of models to describe the impact of climatic variables on germination within seasons and for different seasons. All the
Page 41 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa models are valid for the metrics and seasons to which they relate and all provide insight into how a complex interplay of climatic factors may influence germination of V. staminosa seedlings. Using techniques of Stepwise Regression, Sequential Addition, Forward Selection, and Backward Elimination to develop a suite of models that contain different „significant‟ predictors is a common approach in hypothesis forming logistical analyses (Tabachnick and Fidell 1996; Ramesy and Schafer 2002).
Ordering of the models using Burnham and Anderson‟s (2002) approach to Akaike‟s Information Criterion (1973; cited in Burnham and Anderson‟s 2002) provides evidence that a combination of the amount of rainfall, the number of rainy days, the impact of tropical cyclones/lows, and diurnal temperature variation modelled in the traditional Four Seasons climate best describes germination of V. staminosa seedlings (P-value = 0.424). The model for the Six Noongar Seasons was the next best (P-value = 0.252), even though limitations caused by the relatively small size prevented the significance of terms for rainfall intensity and diurnal variation being assessed. Interestingly the optimum Mediterranean Climate model for germination of V. staminosa seedlings, which was developed utilising all the climate variables in a single analysis, has only marginal significance (P-value = 0.060).
Considering a suite of models also goes someway to addressing the issue that predictors that were considered „not significant‟ and removed from a preliminary model may well become significant when included with another predictor added latter in the analysis (Ramesy and Schafer 2002). Also, given the limitations of the sample size and the high leverage (i.e. influence of outliers) generated by modelling extreme events, such as cyclones or extremely wet and dry years, care has to be taken interpreting the models and the predictions about germination that they suggest. The two best models for the traditional Four Seasons and the six Noongar Seasons (Sections 4.6.3 and 4.6.4) provide strong evidence that:
The passage of a tropical cyclone or tropical low across the region creating a pulse of germination.
The more rainy days (>1mm of rainfall) that occur in late autumn the greater the amount of germination that will occur.
Surprisingly, higher winter rainfall reduces germination.
Page 42 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
In relation to the diurnal temperature range, a metric that could not be incorporated in the Noongar Seasons modelling, the model for the traditional Four Seasons provides strong evidence that: The greater the diurnal temperature range in autumn (i.e. hotter the days and/or cooler the nights) the greater the germination.
There is less germination in winter with higher diurnal temperature ranges, which would also equate to higher average daily temperatures.
The two models contradict each other in relation to the effect of summer rainfall. The Four Season model suggests it may result in a very slight increase in germination while the Noongar Season model suggests a very sight decrease. The magnitude of the predicted change in germination either way is marginal, but this provides a good example of why it often pays to consider a suite of logistical models, rather than basing firm predictions about biological systems on a single theoretical model.
The rainfall and diurnal temperate predictors of the best Four Seasons model are consistent with modelling of Nicholls et al. (2001) from the Australian Bureau of Meteorology Research Centre, who developed „skilful‟ models for wheat production in SWWA that utilised monthly rainfall values and diurnal temperature ranges as predictors. Further, both the Four Seasons model and the Noongar Seasons models provide strong evidence in support of Cowling‟s et al. (2005) hypotheses about the role of rainfall variability. My study provides strong evidence that in addition to the amount of rainfall, the seasonality of rainfall, the number of rainfall events (days), and the passage of tropical cyclones/lows can significantly influence the structure of vegetation in SWWA. At a finer scale, Merritt et al. (2007) postulates germination of non-dormant seeds was enhanced by warm stratification in moist soil in April and May. They attributed this effect to early rains moistening the soil while mean daily temperature were ≥15 oC and the mean diurnal temperature range was >5oC. This observation fits with predictions of both models that a higher diurnal temperature range and/or rain in autumn increase germination. The impact of higher temperatures during this period can be attributed to the role of temperature in alleviating dormancy (Fenner and Thomson 2005; Merritt et al. 2007; Counsens et al. 2010).
The prediction of both models that the passage of tropical cyclones/lows during autumn will generate a pulse of germination is consistent with Schwinning‟s et al.
Page 43 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
(2004) hypothesis that a combination of rare large recruitment events (pulses) and low continuos recruitment are important for maintaining populations of long- lived perennial plants growing in semiarid environments. Cyclone induced germination of dicotyledon plants has also been observed in the adjoining Murchison region and a pulse in the germination of V. staminosa seedlings was reported after the passage of two cyclones between February and March 1999 (Mott 1972; Mott 1973; Yates and Ladd 2004; Australian Bureau of Meteorology 2011).
The inclusion of germination data from years with extreme weather conditions, such as cyclones and anomalously wet or dry years, resulted in all models having reasonably high levels of leverage, that is the impact of single or a few outliers strongly influenced the values of coefficients and the significance of predictors (variables) of the models reported in Section 4.6. As a result caution needs to be applied in interpretation of the models and predictions arising from that interpretation (Ramesy and Schafer 2002). However, in modelling climate related phenomena it is only these rare extreme events that produce statistically significant variation in the systems being studied (England et al. 2006). After analysing climate data of 5-6 years duration England et al. (2006) stated “A much longer observational record is required to obtain detailed statistically significant patterns … during wet and dry years over SWWA.” This demonstrates the immense value of the long term study of V. staminosa germination that I utilised to produce the modelling in this study.
Page 44 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Chapter 6: Conclusions and Recommendations 6.1 Conclusions
My investigation found no evidence of either smoke treatment or repeated wetting and drying of ripe fruit alleviating dormancy or increasing the rate of germination of V. staminosa seeds in the soil seed bank. However, my study demonstrates that long term in-situ weathering of soil stored seed significantly increases the rate at which a non-dormant seed will become imbibed and germinate in response to cues or triggers in its microenvironment.
The germination models I developed in this study challenge the established view that the amount of annual or „wet season‟ rainfall is the significant driver of germination in the Southwest Botanical Province. These models demonstrate that instead a complex interplay of rainfall characteristics, such as amount and timing of rainfall and number of rainy (≥2mm) days, temperature variables, especially diurnal variation, and extreme climatic events (cyclones/intense tropical lows) provide germination triggers for of V. staminosa.
In terms of the research questions I posed at the commencement of my study, I found strong evidence to reject all three null hypotheses in favour of the alternate hypotheses that:
The amount and timing of rainfall does influence germination of V. staminosa seedlings.
Air temperature, in particular the diurnal temperature range, in combination with the amount of rainfall, the number of rainy days; and/or the rain intensity (amount/event) influences germination of V. staminosa seedlings.
The age of fruit in the soil seed bank influences germination of V. staminosa seedlings.
6.2 Recommendations for Future Research
A number of papers published over the last decade have identified shortcomings in the experimental design and/or challenged the conclusions of earlier and concurrent research (Hughes 2003; Moles and Westoby 2004; Merritt et al. 2007; Cousens et al. 2010). Similar to Cousens‟ et al. (2010) research into the role of the pericarp in dormancy and imbibition of the seed, my research has raised as many questions as it has (possibly) answered. There is clearly a need for better
Page 45 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa designed experiments with larger sample sizes that investigate phenomena of interest at a finer spatial and temporal scale for longer periods (e.g. several years at least) for a range of rare and common conjoiner species across the SWBP. Specifically there need to be additional research into:
Dormancy mechanisms and germination triggers for fresh (current season) Verticordia fruit compared to fruit in the soil seed bank.
The role of the pericarp (if any) in imposing physical and/or physiological dormancy and controlling imbibition and desiccation of Verticordia seeds.
Long term (7+ years) trails of soil seed bank persistence.
Short (months and season) and long term (7+ years) study of the viability and germinability of soil stored fruits of known age.
Seasonal scale investigation of the impact of weather on seedling germination and survival within and between seasons for a minimum of three years.
Establishment and maintenance of long term monitoring of seedling germination and survival, as has been occurring with the Wongan Hills V. staminosa meta-population since 1999.
Page 46 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
References
Allan, R.J., and M.R., Haylock. 1993. “Circulation features associated with winter rainfall decrease in Southwestern Australia.” Journal of Climate 6: 1356-1367.
Attiwill, P. and B. Wilson, eds. 2004. Ecology: An Australian Perspective. Melbourne, Victoria: Oxford University Press.
Attiwill, P. and B. Wilson, 2004. “Ecology in Australia.” In Ecology: An Australian Perspective, edited by P. Attiwill and B. Wilson, 184-194. Melbourne, VIC: Oxford University Press.
Australia. Bureau of Meteorology. 2010. Indigenous Weather Knowledge: Surviving the Ice Age. http://www.bom.gov.au/iwk/climate_culture/ index.shtml.
Australia. Bureau of Meteorology. 2011. Climate Data Online. http://www.bom.gov.au/climate/data/.
Australia. Department of the Environment and Heritage (DEH) 2006. Indigenous Engagement in NRM in WA. http://www.nrm.gov.au/publications/books /indigenous-nrm-engagement.html.
Baker, K.S., K.J. Steadman, J.A. Plummer, D.J. Merrit, and K.W. Dixon. 2005. “The Changing Window of Conditions that Promotes Germination of Two Fire Ephemerals, Actinotus leucocephalus (Apiaceae) and Tersonia cyathiflora (Gyrostemonaceae).” Annals of Botany 96: 1225–1236.
Beard, J. S. 1980. “A New Phytogeographic Map of Western Australia.” Western Australian Herbarium Research Notes 3: 37-58.
Berenson, M. L., D. M. Levine, and T. C. Krehbiel. 2006. International Edition Basic Business Statistics: Concepts and Applications. 10th Edition. New Jersey, USA: Pearson Education Inc.
Bindon, P.R. 1997. “Aboriginal People and Granite Domes”. Journal of the Royal Society of Western Australia 80: 173-179.
Bindon, P. and T. Walley. 1992. “Hunters and gathers.” LANDSCOPE: 8 (1) 28- 35.
Page 47 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Bevill R.L., S.M. Louda and L.M. Stanforth. 1999. “Protection from Enemies in Managing Rare Plant Species.” Conservation Biology 13(6): 1323-1331.
Brown, J. M. 1989. “Regional Variation in Kwongan in the Central Wheatbelt of South-Western Australia.” Australian Journal of Ecology 14: 345-355.
Burnham, K.P. and D.R. Anderson. 2002. Model Selection and Multi Model Inference: A Practical Information – Theoretic Approach. 2nd Edition. New York, NY: Springer Science+Business Media, LLC.
Christensen, J.H., B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. Held, R. Jones, R.K. Kolli, W.-T. Kwon, R. Laprise, V. Magaña Rueda, L. Mearns, C.G. Menéndez, J. Räisänen, A. Rinke, A. Sarr and P. Whetton. 2007. Regional Climate Projections. In Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, ed. S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.B. Averyt, M. Tignor and H.L. Miller, 895- 902. Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press. [NB. IPCC requires both Cambridge and New York cited]
Clements F. E. 1905. Research Methods in Ecology. Lincoln, USA: The University Publishing Company.
Cochrane, A., K. Brown, S. Cunneen and A. Kelly. 2001. “Variation in seed
production and germination in 22 rare and threatened Western Australian
Verticordia (Myrtaceae).” Journal of the Royal Society of Western Australia
84: 103-110.
Cochrane, A. and M.I. Daws. 2008. “How will warming temperatures affect
recruitment of southern Western Australian narrow range mountain
endemics?.” Australasian Plant Conservation 16(4): 16-18.
Cody, M. L. 1986. “Diversity, rarity, and conservation in Mediterranean-climate regions”. In Conservation Biology: The Science of Scarcity and Diversity, edited by M.E Soulé Page Numbers. Sutherland, USA: Sinauer Associates INC.
Cousens, R.D., K. R. Young, and A. Tadayyon. 2010. “The role of the persistent fruit wall in seed water regulation in Raphanus raphanistrum
Page 48 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
(Brassicaceae)”. Annals of Botany 105: 101–108, doi:10.1093/aob/mcp268.
Cowling, R.M., P.W. Rundel, B.B. Lamont, M.K. Arroyo and M. Arianoutsou. 1996. Plant diversity in Mediterranean-climate regions. Tree 11: 362-366.
Cowling, R. M., F. Ojeda, B. B. Lamont, P. W. Rundel, and R. Lecmere-Oertel. 2005. “Rainfall reliability, a neglected factor in explaining convergence and divergence of plant traits in fire-prone Mediterranean-climate ecosystems.” Global Ecology and Biogeography 14: 509-519.
Conservation International. 2007. Biodiversity Hotspots: Southwest Australia. http://www.biodiversityhotspots.org/xp/hotspots/australia/Pages/default.as px
Dirnböck, T., R.J. Hobbs, R.J. Lambeck, and P.A. Caccetta. 2002. “Vegetation distribution in relation to topographically driven processes in south-western Australia.” Applied Vegetation Science 5(1) 147–158. DOI: 10.1111/j.1654-109X.2002.tb00544.x
Dixon, K.W., S. Roche, J.S. Pate. 1995. “The promotive effect of smoke derived from burnt native vegetation on seed germination of Western Australian plants”. Oecologia 101: 185-192.
Dynesius, M. and R. Jansson. 2000. “Evolutionary consequences of changes in
species‟ geographical distribution driven by Milankovitch climate
oscillations.” Proceedings of the National Academy of Sciences of the United
States of America 97(16): 9115-9120.
England M.H., C.C. Ummenhofer, and A. Santoso. 2006. “Interannual Rainfall
Extremes over Southwest Western Australia Linked to Indian Ocean Climate
Variability.” Journal of Climate 19: 1948-1969
Fay, P.A. 2009. “Preciptitation variablty and primary productivity in water-limited
ecosystems: how plants „leverage‟ precipitation to „finance; growth.” New
Phtyologist 181: 5-8.
Fenner, M., and K. Thompson. 2005. The Ecology of Seeds. Cambridge:
Cambridge University Press.
Page 49 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Fischlin, A., G.F. Midgley, J.T. Price, R. Leemans, B. Gopal, C. Turley, M.D.A.
Rounsevell, O.P. Dube, J. Tarazona, A.A. Velichko. 2007. Ecosystems,
Their Properties, Goods, and Services. In Climate Change 2007: Impacts,
Adaptation and Vulnerability. Contribution of Working Group II to the Fourth
Assessment Report of the Intergovernmental Panel on Climate Change. Ed.
M.L. Parry, O.F. Canziani, J.P. Palutikof, P.J. van der Linden and C.E.
Hanson, 211-272. Cambridge, United Kingdom and New York, NY, USA:
Cambridge University Press . [NB. IPCC requires both Cambridge and New
York cited]
Flood, J. 1995. Archaeology of the Dreamtime: The story of prehistoric Australia and its people. Sydney, Australia: Angus and Robertson.
Fontaine, J.B., D.C. Donato, J.L. Campbell, J.G. Martin, B.E. Law. 2010. “Effects of post-fire logging on forest surface air temperature in the Sishiyou Mountains, Oregon, USA.” Forestry Advances Access: 1-6. doi: 10.1093/forestry/cpq030
Gardner, C. A. 1959. Wildflowers of Western Australia. Perth, Western Australian Newspapers Ltd.
George, E. A. 2002. Verticordia: The Turner of Hearts. Perth, Australia: University of Western Australia Press.
George, A.S. 1991. “The maintenance of species richness in plant communities: The importance of the regeneration niche.” Nuytsia 7(3).
Geosciences Australia 2009. Place Name Search: Mocardy Hill. http://www.ga.gov.au/bin/gazd01?rec=274945.
Gioia, P. 2007. Distribution Map Verticordia staminosa. www.florabase.wa.gov.au.
Gole, C. 2006. The Southwest Australia Ecoregion: Jewel of the Australian Continent. Perth, WA: WWF – Australia.
Grinnell. P. 1904. “The Origin and Distribution of the Chestnut-Backed Ckickadee.” The Auk 21: 364-382. Website
Grinnell. P. 1917. “The Niche Relationships of the Californian Thrasher.” The Auk 34: 427-433. Website
Page 50 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Groves, R., and P. Catling. 2004. “Heaths and Low Open-forests.” In Ecology: An Australian Perspective, edited by P. Attiwill and B. Wilson, 184-194. Melbourne, VIC: Oxford University Press.
Grubb, P.J. 1977. “The Maintenance of Species-Richness in Plant Communities: The Importance of the Regeneration Niche.” Biological Reviews 52(1): 107–145. DOI: 10.1111/j.1469-185X.1977.tb01347.x
Harben, S., and L. Collard. n.d. Recording Traditional Knowledge; Our Country...Our Stories…Our People: Literature Review for Avon Basin Noongar Heritage and Cultural Significance of Natural Resources. http://www.wheatbeltnrm.org.au/resources/RTKlitweb.pdf
Hill, R. and T. Broodribb, 2004. “The Evolution of Australia‟s Living Biota.” In Ecology: An Australian Perspective, edited by P. Attiwill and B. Wilson, 184- 194. Melbourne, VIC: Oxford University Press.
Hooper, S.D. 1979. “Biogeographical aspects of speciation in the southwest Australian flora.” Annual Review of Ecology and Systemetics 10: 399-422.
Hopper, S.D. 2009. “OCBIL theory: toward an integrated understanding of evolution, ecology, conservation of biodiversity on old, climatically buffered, infertile landscapes.” Plant Soil 322: 49-86.
Hopper, S.D., A.P. Brown, and N.G. Marchant. 1997. Plants of Western Australian granite outcrops. Journal of the Royal Sovitey of Western Australia 80: 141-158.
Hooper, S.D., and P. Gioia. 2004. “The Southwest Australian Floristic Region: evolution and conservation of a global hotspot of biodiversity.” Annual Review of Ecology and Systemetics 35: 623-650.
Hopper, S.D., K.W. Dixon, and R.S. Hill. 2006. “Australian seeds through time.” In Australian Seeds: A guide to their Collection, Identification, and Biology, edited by L. Sweedman and D. Merritt, 5-10. Collingwood: CSIRO Publishing.
Hughes, L. 2003. “Climate change in Australia: Trends, projections and impacts.” Austral Ecology 28: 423-443.
Hussey, B.M.J., and K.J. Wallace. 2003. Managing Your Bushland. Perth: Western Australia. Department of Conservation and Land Management.
Page 51 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Hutchinson, G.E. 1957. “Concluding Remarks.” Paper presented at Population Studies: Animal Ecology and Demography – Cold Spring Harbor Symposia. In Quantitative Biology 22: 415-427. http://artifex.org/~ecoreaders/lit/ Hutchinson1957.pdf
Krebs, C.J. 1994. Ecology: The experimental analysis of distribution and adundance. 4th ed. New York, NY: Harper Collins College Publishers.
Kruckeberg, A.R. and D. Rabinowitz. 1985. “Biogeographical aspects of endemism in higher plant.” Annual Review of Ecology and Systemetics 16: 447-479.
Johnson, R. A., and D. W. Wichern. 2002. Applied Multivariate Statistical Analysis. 5th Edition. Upper Saddle River, USA: Prentice Hall.
Juliano, S. A. 1993. “Nonlinear curve fitting.” In Design and Analysis of Ecological Experiments, edited by S. M. Scheiner and J. Gurevitch, 179. New York (NY), USA: Chapman and Hall.
Lamont, T. 2004. “Evolutionary convergence and sclerophylly in Mediterranean regions.” In Ecology: An Australian Perspective. Attiwill, P. and B. Wilson, eds. Melbourne, Victoria: Oxford University Press.
Leishman, M.R. and M. Westoby. 1998. “Seed size and shape are not realated to persistence in soil in Ausralia in the same way as in Britain.” Functional Ecology 8: 480-485.
Leopold, A.C. 1964. Plant Growth and Development. New York, NY: McGraw Hill Inc.
Lush W. M., P. E. Kay, and R. H. Groves. 1984. “Germination of Clematis microphylla Seeds Following Weathering and Other Treatments.” Australian Journal of Botany 32: 121-9
Martyn, A.J., L.U. Seed, M.K.J. Ooi, and C.A. Offord. 2009. “Seed fill, viability and germination of NSW species in the family Rutaceae.” Cunninghamia 11(2): 203–212
Merritt D. J., S. R. Turner, S. Clarke, and K. W. Dixon. 2007. Seed dormancy and germination stimulation syndromes for Australian temperate species. Australian Journal of Botany 55: 336–344.
Page 52 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Microsoft Corporation. 2002. Microsoft® Excel 2003. Redmond, Washington: Microsoft Corporation. http://support.microsoft.com/ph/2512
Minitab. 2007. Minitab 15 Statistical Software. State College, Pennsylvania: Minitab Incorporated. http://www.minitab.com.
Moles, A.T., D.I. Warton and M. Westoby. 2003. “Seed size and survival in the soil in arid Australia.” Austral Ecology 28(5): 575-585
Moles, A.T. and M. Westoby 2004. “What do seedlings die from and what are the implications for evolution of seed size.” OIKOS 106: 193-198.
Moore, D.S., and G.P. McCabe. 1993. Introduction to the Practice of Statistics. 2nd Edition. New York, NY: W.H. Freeman and Company.
Mott, J. J. 1974. “Factors Affecting Seed Germination in Three Annual Species from an Arid Region of Western Australia”. Journal of Ecology Vol. 62(3): 699-709. JSTOR Full Text Archive.
Murray, B.R., P.H. Thrall, A.M. Gill and A.B. Nicotra. 2002. “How plant life- history and ecological traits relate to species rarity at varying spatial scales.” Austral Ecology 27: 291-310.
Murdoch University 2008. “OWN: One With Nature. The Nyoongar lifestyle in harmony with nature.” Diversity Matters @ Murdoch, Semester 1, 2008. http://equityaccess.murdoch.edu.au/diversity/DM@M/PublishedEditions/DM @M2008sem01.pdf
Myers, N., R.A. Mittermeier, C.G. Mittermeier, G.A.B. da Fonseca, and J. Kent. 2003. “Biodiversity hot spots for conservation priorities.” Nature 403: 803- 808.
Nannup, N., and D. Deeley. 2006. “Rainfall and Water as Cultural Drivers”. In Proceedings of the 1st National Hydropolis Conference, Burswood Entertainment Complex, 8-11 October 2006. Perth, WA: Storm water Industry Association. http://www.hydropolis.com.au/papers.htm
Navarro, L. and J. Guitián. 2002. Seed germination and seedling survival of two threatened endemic species of the northwest Iberian peninsular. Biological Conservation 109: 313-320.
New, T. 2004. “The Niche.” In Ecology: An Australian Perspective, edited by P. Attiwill and B. Wilson, 184-194. Melbourne, VIC: Oxford University Press.
Page 53 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Nicholls, N., L. Chambers, M. Haylock, C. Frederiksen, D. Jones, and W. Drosdowsky. 1999. Climate Variability and Predictability for South-West Western Australia: Phase 1 Report to the Indian Ocean Climate Initiative. www.ioci.commerce.wa.gov.au/firstphase.htm
Nicholls, N., L. Chambers, and W Drosdowsky. 2001. Climate Variability and Predictability for South-West Western Australia: Report of Second Research Phase to Indian Ocean Climate Initiative. www.ioci.commerce.wa.gov.au/firstphase.htm
Nield, A.P., P.G. Ladd, and C.J. Yates. 2009. “Reproductive biology, posr-fire succession dynamics and population viability analysis of critically endangered Western Australian shrub Calytrix breviseta subsp. breviseta (Myrtaceae).” Australian Journal of Botany 57: 451-464.
Nind, I.S. 1831. Description of the Natives of King George's Sound (Swan River Colony) and adjoining Country. Journal of the Royal Geographical Society of London 1: 21-50. http://en.wikisource.org/wiki/Journal_of_the_Royal _Geographical_Society_of_London/Volume_1
Patten, J., K. Kershaw and B. Loudon. 2004. Interim Recovery Plan No. 161: Granite Featherflower (Verticordia Staminosa subsp. c ylindracea var. cylindracea): Interim Recovery Plan 2004-2009. Wanneroo, Western Australia: Government of Western Australia, Department of Conservation and Land Management.
Philippi, T.E. 1993. “Multiple regression: Herbivory.” In Design and Analysis of Ecological Experiments, edited by S. M. Scheiner and J. Gurevitch, 179. New York (NY), USA: Chapman and Hall.
Pickup, M., K.L. McDougall, and R.J. Whelan. 2003. “Fire and flood: Soil-stored seed bank and germination ecology in the endangered Carrington Falls Grevillea (Grevillea rivularis , Proteaceae).” Austral Ecology 28: 128–136
Ramsey, F.L. and D.W. Schafer. 2002. The Statistical Slueth: A Course in Methods of Data Analysis. Pacific Grove, CA: Thomson Learning Inc.
Raven, P.H., F.F. Evert, S.E. Eichhorn. Biology of Plants. 7th Edition. New York, NY: W.H. Freeman and Company.
Page 54 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Richardson, D.M., R.M. Cowling, B.B. Lamont and H. J. van Hensbergen. 1995. “Coexistence of Banksia species in southwestern Australia: The role of regional and local processes.” The Journal of Vegetation Science 6(3): 329- 342.
Rippery, E. 1995. Plants of the Perth coast and islands. Perth: University of Western Australia Press.
Robertson, T.R., C.W. Bell, J.C. Zak, D.T. Tissue. 2009. “Precipitation timing and magnitude differentially affect aboveground annual net primary productivity in three perennial species in a Chihuahuan Desert grassland”. New Phytologist 181: 230-242.
South West Aboriginal Land and Sea Council. 2011. History of the Noongar People. http://www.noongar.org.au/noongar-people-history.php
Tabachnick, B.G. and L.S. Fidell. 1996. Using multivariate statistics – 3rd Edition. New York, USA: HaperCollins College Publishers.
Tarroso, P., and H. Rebelo. 2010. “E-Clic – easy climate data converter.” Ecograohy 33: 617-620.
Tieu, A., and L.M. Egerton-Warburton. 2000. “Contrasting seed morphology dynamics in relation to the alleviation of dormancy with soil storage.” Canadian Journal of Botany 78: 1187-1198.
Turner, S.R., D.J. Merritt, C.C. Baskin, K.W. Dixon, and J.M. Baskin. 2005. “Physical dormancy in seeds of six genera of Australian Rhamnaceae.” Seed Science Research 15: 51-58.
Udvardy, M.F. D. 1959. “Notes on the Ecological Concepts of Habitat, Biotope and Niche.” Ecology 40(4) 725-728. http://www.jstor.org/stable/1929830.
Wallace, K., and J. Huston. 1997. Exploring Woodlands with Nyoongars. Perth: Western Australia. Department of Conservation and Land Management.
Western Australia. Department of Water. 2011. A Survey of Aboriginal Social Water Requirements for the Southern Blackwood Plateau and the Scott Coastal Plain Southwest, Western Australia.
Western Australia. Indian Ocean Climate Initiative Panel. 2002. Climate variability and change in south west Western Australia. Perth: Department of Environment, Water and Catchment Protection.
Page 55 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
White, D., G. Tupper, and H Mavi. 1999. “Agricultural Climate Research and Services in Australia”. Canberra: Land and Water Australia.
Withers, P.C. 2000. “Overview of granite outcrops in Western Australia.” Journal of the Royal Society of Western Australia 83: 103-108
Wright, P. B. 1974. Seasonal rainfall in southwestern Australia and the general circulation. Monthly Weather Review 102: 219-232. DOI: 10.1175/1520- 0493(1974)102<0219:SRISAA>2.0.CO;2
Wright, W. J., 1997. “Tropical-extratropical cloudbands and Australian rainfall. 1. Climatology.” International Journal of Climatology 17(8): 807-829. DOI:10.1002/(SICI)1097-0088(19970630)17:8<807::AID- JOC162>3.0.CO;2-J
Yates, C.J. and P.G. Ladd. 2004. “Breeding system, population and demography in the rare granite endemic shrub Verticordia staminosa ssp. staminosa in south-west Western Australia.” Austral Ecology 19: 189-200
Yates, C.J. and P.G. Ladd. 2005. “Relative Importance of Reproductive Biology and Establishment Ecologu for Persistence of a Rare Shrub in a Fragmented Landscape.” Conservation Biology 19(1): 239-249.
Yates, C.J. and P.G. Ladd. 2010. “Using population viability analysis to predict the effect of fire on the extinction risk of a endangered shrub Verticordia fimbrilepis subsp. fimbrilepis in a fragmented landscape.” Plant Ecology 211: 305-319.
Yates, C.J., P.G. Ladd, D.J. Coates, and S. McArthur. 2007. “Hierarchies of cause: understanding rarity in an endemic shrub Verticordia staminosa (Myrtaceae) with a highly restricted distribution.” Australian Journal of Botany 55: 194- 205.
Page 56 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Appendices
A. Exploratory Analyses: Ex-situ Germination Experiments
A1 Seed Viability of the Soil Seed Bank
12
10
8
6
milligrams 4
2
0 Trial 1 Trail 2 Trial 3
Figure A1.1: Combined Histograms and the Box & Whisker Plot for pericarps found not contain a seed in the 3 „Cut Seed‟ Trials.
12
10
8
6 milligrams 4
2
0 Trial 1 Trail 2 Trial 3
Figure A1.2: Combined Histograms and the Box-Whisker Plot for pericarps found to contain a seed in the 3 „Cut Seed‟ Trials.
A2 Germination Response of Recovered Pericarp
7 12 Buried in Cracks 6 Buried in Aprons 10 5
4 8
3
Frequency 6 2
1 4
0 Number ofSeedlings Number
>11 2 0 to 1 2 to 3 4 to 5 6 to 7 8 to 9 10 to 11
Number of Seedlings in a Replicate 0 Buried in Cracks Buried in Aprons
Figure A2.1: Combined histograms and Box & Whisker Plots for number of seedlings that emerged in replicates of recovered pericarps.
Page 57 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
18 Buried in Cracks Buried in Aprons Buried in Cracks 12
6 Buried in
Aprons Number ofSeedlings Number
0
25 30 35 40 45 50 25 30 35 40 45 50 Days to Emerge Days to Emerge
Figure A2.2: Combined histograms and Box & Whisker Plots for time taken for seedlings to emerge in replicates of recovered pericarps.
40
35
30
25
20
Percentage 15
10
5
0 Buried in Cracks Buried in Aprons
Figure A2.3: Combined histograms and Box & Whisker Plots for percentage of originally viable seeds that germinated in replicates of recovered pericarps.
A3 Impact of Smoke and Weathering on Germination
Smoke & Weathering 4 Weathering Smoke Control 3
2 Frequency
1
0 ≤3 4 5 6 7 8 ≥9 Number of Seedlings
Figure A3.1: Combined Histograms for number of seedlings to emerge for each germination treatment (Left) and the Log(x) transformed data of the same metric (Right)
Page 58 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
1.00 9
8 0.90
7 0.80
6 0.70 5
0.60
4 Number ofSeedlings Number
3 0.50 Log(Number ofSeedlings) Log(Number
2 0.40 Smoke & Weathering Smoke Control Smoke & Weathering Smoke Control Weathering Weathering
Figure A3.2: Box & Whisker Plots for the number of seedlings that emerged for each germination treatment with raw data on the left and Log-transformed data is on the right.
10 1.00 9 0.90 8
7 0.80 6
5 0.70
4 0.60 3
Numer ofSeedlings Numer 2
0.50 Log(Num ofSeedlings) Log(Num 1
0 0.40
Tray 1 Tray 2 Tray 3 Tray 4 Tray 5 Tray 6 Tray 7 Tray 8 Tray 9 Tray 1 Tray 2 Tray 3 Tray 4 Tray 5 Tray 6 Tray 7 Tray 8 Tray 9 Tray 10 Tray 10
Figure A3.3: Box & Whisker Plots for the number of seedlings with each Tray of replicates. Again raw data is shown on the left and Log-transformed data is on the right.
Figure A3.4: Combined Histograms of raw data (Left) and Log(x) transformed data of the average time taken for seedlings of each treatment to emerge within a Tray.
Page 59 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Smoke & Smoke & Weathering Weathering
Weathering Weathering
Smoke Smoke
Control Control
30 35 40 45 50 55 60 65 1.5 1.55 1.6 1.65 1.7 1.75 1.8 1.85 Average Days from Planting Log(Avgerage Days from Planting)
Figure A3.5: Box & Whisker Plots of average time taken for seedlings with each treatment to emerge within a Tray. Raw data on left and Log-transformed data on right.
Tray 1 Tray 1 Tray 2 Tray 2 Tray 3 Tray 3 Tray 4 Tray 4 Tray 5 Tray 5 Tray 6 Tray 6 Tray 7 Tray 7
Tray 8 Tray 8 Tray 9 Tray 9 Tray 10 Tray 10
30 35 40 45 50 55 60 65 1.50 1.55 1.60 1.65 1.70 1.75 1.80 Average Days from Planting Log(Average Days from Planting)
Figure A3.6: Box & Whisker Plots of average time taken for seedlings to emerge within a Tray. Again the raw data is on the left and Log(x) transformed data is on the right.
1.00 2.10
2.00 0.80
1.90 0.60
1.80
0.40 1.70
0.20
Proportion ofSeedlings Proportion 1.60 Log(Percent of Viable Seed) Log(PercentofViable
0.00 1.50 Smoke & Weathering Smoke Control Smoke & Weathering Smoke Control Weathering Weathering
Figure A3.7: Box & Whisker Plots for the proportion of viable seeds that produced seedlings with raw data is shown on the left and Log(x) transformed data is on the right.
Page 60 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
B. Exploratory Analyses: Field Germination
B1 Seedling Germination in the Field
6 1 S-POPn A S-POPn B S-POPn G S-POPn W
4 0.75 Frequency 2 0.5
Proportion ofQuadrats Proportion 0.25 0
0.8 to 1.0 0.0 to <0.2 0.2 to <0.4 0.4 to <0.6 0.6 to <0.8 0 Proportion of Quadrats S-POPn A S-POPn B S-POPn G S-POPn W
Figure B1.1 Combined Histograms and Box & Whisker Plot showing the between year variability in the proportion of quadrats in each subpopulation that contained seedlings.
10 120
S-POPn A S-POPn B 8 S-POPn G S-POPn W 90
6
4 60 Frequency
2
Number ofSeedlings Number 30 0
0 to 20 21 to 40 41 to 60 61 to 80 81 to 100 101 to 120 0 Numer of Seedlings S-POPn A S-POPn B S-POPn G S-POPn W
Figure B1.2 Combined Histograms and Box & Whisker Plots showing the between year variability in the total number of seedlings in each subpopulation.
8.0 8 S-POPn A S-POPn B S-POPn G S-POPn W 6 6.0
4
4.0 Frequency
2 ofSeedlings Avgerage Number Avgerage 2.0
0
0 to <1 1 to <2 2 to <3 3 to <4 4 to <5 5 to <6 6 to <7 7 to <8 8 to <9 0.0 Average Number of Seedlings S-POPn A S-POPn B S-POPn G S-POPn W
Figure B1.3: Combined Histograms and Box & Whisker Plot showing the between year variability in the average number of seedlings per quadrat by subpopulation.
Page 61 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
C. Exploratory Analyses: Seasons of SW Western Australia C1 Annual Values
650
550
450
350 millimetres
250
150 Annual Rainfall
80
70
60 Days 50
40
30 Rainy Days
9.5
8.5
7.5 millimetres
6.5
5.5 Rainfall Intensity
Figure C1.1: Histograms and Box & Whisker Plots showing between year variability in the Annual Rainfall (Top), the annual number of Rainy Days (more than 1mm of rain in 24 hours), and Annual Rainfall Intensity (Annual Rainfall averaged over the annual number of Rainy Days), for the 20 year period from 1991 to 2010.
22.5
21.8
21.0
20.3 MJ/m2
19.5
18.8
18.0
Figure C1.2: Between year variability in Annual Average Monthly Solar Radiation 1991-2010.
Page 62 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
20.0
19.5
19.0 Celsius
18.5
18.0
Figure C1.3: Between year variability in Annual Average Monthly Temperature 1991-2010.
27.5
27.0
26.5
26.0
Celsius 25.5
25.0
24.5
24.0
Figure C1.4: Between year variability in Annual Average Monthly Maximum Temperature for the 20 year period between 1991 and 2010.
4 12.8
3
12.3
2 Frequency
1 11.8 Celsius
0 11.3
<11.3
10.8 11.3 to <11.511.5 to <11.711.7 to <11.911.9 to <12.112.1 to <12.312.3 to <12.512.5 to <12.7 Celsius
Figure C1.5: Histogram and Box & Whisker Plot showing between year variability in the Annual Average Monthly Minimum Temperature from 1991 to 2010.
16.0
15.0
14.0 Celsius
13.0
12.0
Figure C1.6: Between year variability in the Annual Average Monthly Diurnal Temperature Range for the period 1991 to 2010.
Page 63 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
C2 Mediterranean (Dry/Wet Seasons) Values
6
4
Frequency 2
0
<170
170 to <210210 to <250250 to <290290 to <330330 to <370370 to <410410 to <450 Wet Season (mm)
400
300
200 millimetres
100
0 Dry Season Wet Season
Figure C2.1: Histograms and Box & Whisker Plot showing between year variability in the Dry Season (Nov-May) Rainfall and the Wet Season (Apr-Oct) Rainfall from 1991 to 2010.
6
4
2 Frequency
0
<30
30 to <35 35 to <40 40 to <45 45 to <50 50 to <55 55 to <60 Days in Wet Season
60
45
30 Days
15
0 Dry Season Wet Season
Figure C2.2: Histograms and Box & Whisker Plot showing between year variability in the number of Rainy Days in the Dry Season and in the Wet Season from 1991 to 2010.
Page 64 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6
4
2 Frequency
0
<5.4
5.4 to <5.85.8 to <6.26.2 to <6.66.6 to <7.07.0 to <7.47.4 to <7.87.8 7o <8.2 Wet Season (mm)
16.0
13.0
10.0 millimetres
7.0
4.0 Dry Season Wet Season
Figure C2.3: Histogram and Box & Whisker Plot showing between year variability in Rainfall Intensity in the Dry Season and in the Wet Season (measured as Average Rainfall per Rainy Day) for the 20 year period 1991-2010.
27.5
24.5
21.5
MJ/m2 18.5
15.5
12.5 Dry Season Wet Season
Figure C2.4: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Solar Radiation for the Dry Season and the Wet Season from 1991-2010.
Page 65 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
8 6
6 4
4
2
Frequency Frequency 2
0 0
<22.8 <13.5
22.8 to <23.123.1 to <23.413.4 to <23.723.7 to <24.024.0 to <24.324.3 to <24.624.6 to <24.9 15.5 to <13.813.8 to <14.114.1 to <14.414.4 to <14.714.7 to <15.015.0 to <15.3 Dry Season (oC) Wet Season (oC)
26.0
22.0
18.0 Celsius
14.0
10.0 Dry Season Wet Season
Figure C2.5: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Temperature for the Dry Season and the Wet Season from 1991 to 2010.
8
6
4 Frequency 2
0
<19.0
19.0 to <19.419.4 to <19.819.8 to <20.220.2 to <20.620.6 to <21.021.0 to <21.421.4 to <21.8 Wet Season (oC)
34.0
30.0
26.0 Celsius
22.0
18.0 Dry Season Wet Season
Figure C2.6: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Maximum Temperature during the Dry Season and the Wet Season for the 20 year period from 1991 to 2010.
Page 66 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6 8
6 4
4
Frequency 2 2 Frequency
0 0
<14.8 <7.2
14.8 to <15.115.1 to <15.415.4 to <15.715.7 to <16.016.0 to <16.316.3 to <16.6 7.2 to <7.57.5 to <7.87.8 to <8.18.1 to <8.48.4 to <8.78.7 to <9.09.0 to <9.3 Dry Season (oC) Wet Season (oC)
16.8
14.3
11.8 Celsius
9.3
6.8 Dry Season Wet Season
Figure C2.7: Between year variability in the Average Monthly Minimum Temperature for the Dry Season and the Wet Season in the period from 1991 to 2010.
6 8
6 4
4
Frequency 2 2 Frequency
0 0
<14.8 <7.2
14.8 to <15.115.1 to <15.415.4 to <15.715.7 to <16.016.0 to <16.316.3 to <16.6 7.2 to <7.57.5 to <7.87.8 to <8.18.1 to <8.48.4 to <8.78.7 to <9.09.0 to <9.3 Dry Season (oC) Wet Season (oC)
18.0
16.0
14.0 Celsius
12.0
10.0 Dry Season Wet Season
Figure C2.8: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Diurnal Temperature Range for the Dry Season and the Wet Season in the 20 year period between 1991 and 2010.
Page 67 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
C3 Four Seasons Values
6
4
2 Frequency
0
<20.0
20.0 to <35.035.0 to <50.050.0 to <65.065.0 to <80.080.0 to <95.0 95.0 to <110.0 Spring (mm)
300.0
225.0
150.0 millimetres
75.0
0.0 Summer Autumn Winter Spring
Figure C3.1: Histograms and Box & Whisker plots showing between year variability in Rainfall for each of the 4 seasons for the 20 year period between 1991 and 2010. Summer season is from December to February, Autumn covers March to May, Winter is June to August, and Spring falls in the months of September to November.
Page 68 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6 6
4 4 Frequency 2 Frequency 2
0 0
<5 <15 5 to <7 7 to <9 9 to <11 15 to <19 19 to <23 23 to <27 27 to <31 31 to <35 35 to <39 11 to <13 13 to <15 15 to <17 Winter Rainy Days Spring Rainy Days
40
30
20 Days
10
0 Summer Autumn Winter Spring
Figure C3.2: Histograms and Box & Whisker Plot showing between year variability of the Number of Rainy Days in each of the 4 seasons during the period 1991 to 2010
Page 69 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
5
4
3
2 Frequency
1
0 <5.0 5.0 to 5.5 to 6.0 to 6.5 to 7.0 to 7.5 to ≥8.0 <5.5 <6.0 <6.5 <7.0 <7.5 <8.0 Winter Rain Intensity (mm)
22.0
18.0
14.0
10.0 millimetres
6.0
2.0 Summer Autumn Winter Spring
Figure C3.3: Histogram and Box & Whisker Plot showing between year variability in Rainfall Intensity (measured as Average Rainfall per Rainy Day) across the 4 season for the 20 year period from 1991 to 2010.
Page 70 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6
4
Frequency 2
0
<10.6
10.6 to <11.111.1 to <11.611.6 to <12.112.1 to <12.612.6 to <13.113.1 to <13.613.6 to <14.1 Winter (MJ/m2)
31.0
24.0
2 MJ/m
17.0
10.0 Summer Autumn Winter Spring
Figure C3.4: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Solar Radiation for the 4 seasons over the period 1991-2010.
Page 71 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
27.0
23.0
19.0 Celsius
15.0
11.0 Summer Autumn Winter Spring
Figure C2.5: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Temperature for the 4 season in the 20 year period 1991 to 2010.
Page 72 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
36.0
32.0
28.0
Celsius 24.0
20.0
16.0 Summer Autumn Winter Spring
Figure C2.6: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Maximum Temperature during the 4 seasons between 1991 and 2010.
Page 73 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
7 6
6 5 5 4 4 3 3
2 Frequency 2 Frequency
1 1
0 0
<16.1 <12.2
16.1 to <16.416.4 to <16.716.7 to <17.017.0 to <17.317.3 to <17.617.6 to <17.919.9 to <18.2 12.2 ro <12.612.6 to <13.013.0 to <13.413.4 to <13.813.8 to <14.214.2 ro <14.614.6 to <15.0 Summer (oC) Autumn (oC)
17.0
13.0 Celsius
9.0
5.0 Summer Autumn Winter Spring
Figure C2.7: Between year variability in the Average Monthly Minimum Temperature during the 4 seasons between Summer and Spring in the years 1991 to 2010.
Page 74 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
20.0
17.0
14.0 Celsius
11.0
8.0 Summer Autumn Winter Spring
Figure C2.8: Histograms and Box & Whisker Plot showing between year variability in the Average Monthly Diurnal Temperature Range for each of the 4 seasons from 1991 to 2010.
Page 75 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
C4 Six Noongar Seasons
8
6
4 Frequency 2
0
<5.0
5.0 to <15.0 15.0 to <25.025.0 to <35.035.0 to <45.045.0 to <55.055.0 to <65.0 Kambarang (mm)
250.0
200.0
150.0
100.0 millimetres
50.0
0.0
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.1: Histograms and Box & Whisker Plot showing between year variability (1991- 2010) in Rainfall for the six Noongar seasons of the SWBP.
Page 76 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6 6
4 4 Frequency Frequency 2 2
0 0 1 2 3 4 5 6 7 >7 1 2 3 4 5 6 7 >7 Birack Rainy Days Bunurua Rainy Days
6 6
4 4 Frequency Frequency 2 2
0 0 ≤5 6 7 8 9 10 11 >11 ≤10 12 14 16 18 20 22 >22 Djeran Rainy Days Makuru Rainy Days
6 6
4 4 Frequency 2 Frequency 2
0 0 ≤10 13 16 19 22 25 ≥28 ≤2 3 4 5 6 7 8 ≥8 Djilba Rainy Days Kambarang Rainy Days
30
25
20
15 Days 10
5
0
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.2: Histograms and Box & Whisker Plot showing between year variability in the number of Rainy Days in each of the Noongar seasons in the period 1991 to 2010.
Page 77 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
40.0
30.0
20.0 millimetres 10.0
0.0
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.3: Histograms and Box & Whisker Plot of between year variability (1991-2010) in Rainfall Intensity (i.e. Average Rainfall per Rainy Day) for the six Noongar seasons.
Page 78 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
32.5
26.5 2
20.5 MJ/m
14.5
8.5
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.4: Histograms and Box & Whisker Plot showing between year (1991-2010) variability in the Average Monthly Solar Radiation for the six Noongar seasons.
Page 79 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6 6
4 4 Frequency 2 Frequency 2
0 0
<16.9 <11.7
16.9 to <17.317.3 to <17.717.7 to <18.118.1 to <18.518.5 to <18.9 189.9 to <19.319.3 tto <19.7 11.7 to <12.012.0 to <12.312.3 to <12.612.6 to <12.912.9 to <13.213.2 to <13.513.5 to <13.8 Djeran (oC) Makura (oC)
27.0
23.0
19.0 Celsius
15.0
11.0
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.5: Histograms and Box & Whisker Plot showing between year variability (1991 to 2010) in the Average Monthly Temperature for each of the six Noongar seasons.
Page 80 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
36.5
29.5 Celsius 22.5
15.5
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.6: Histograms and Box & Whisker Plot showing between year (1991-2010) variability in the Average Monthly Maximum Temperature for the six Noongar seasons.
Page 81 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6
4
Frequency 2
0
<16.0
16.0 to <16.316.3 to <16.616.6 to <16.916.9 to <17.217.2 to <17.517.5 to <17.8 Birak (oC)
17.0
13.0 Celsius 9.0
5.0
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.7: Between year variability in the Average Monthly Minimum Temperature for the Dry Season and the Wet Season in the period from 1991 to 2010.
Page 82 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
6
4
Frequency 2
0
<14.3
14.3 to <14.814.8 to <15.315.3 to <15.815.8 to <16.316.3 to <16.816.8 to <17.3 Bunurua (oC)
6
4
Frequency 2
0
<10.9
10.9 to <11.511.5 to <12.112.1 to <12.712.7 to <13.313.3 to <13.913.9 to <14.5 Dejeran (oC)
6
4
Frequency 2
0
<15.2 <17.0
15.2 to <15.515.5 to <15.815.8 to <16.116.1 to <16.416.4 to <16.716.7 to <17.0 Kambarang (oC)
20.0
18.0
16.0
14.0 Celsius 12.0
10.0
8.0
Birak Djilba Djeran Makuru Bunurua Kambarang
Figure C4.8: Histograms and Box & Whisker Plot showing between year variability in the
Average Monthly Diurnal Temperature Range for each of the six Noongar seasons.
Page 83 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa D. Outcomes of Least Squares Regression Analyses of Correlation between Annual Climate Metrics The Pearson‟s Coefficient of Correlation (R2) for each pair of climate metrics is reported on the lower left and the significance of each R2 value is reported on the upper right. Significant correlations are shown in red. Annual Avg. Annual Avg Annual Avg. Annual Avg. Annual Avg. Annual Annual Annual Rain Monthly Monthly Monthly Monthly Monthly Rainfall Rainy Days Intensity Diurnal Solar Rad Temp. Max. Temp. Min. Temp. Range
Annual F = 29.06 F = 7.046 F = 1.506 F = 8.262 F = 1.304 F = 14.07 F = 5.691
Rainfall P-value<0.001 P-value=0.016 P-value=0.235 P-value=0.010 P-value=0.268 P-value=0.001 P-value=0.029
Annual F = 0.181 F = 1.402 F = 10.25 F = 2.349 F = 23.00 F = 7.210 61.8% Rainy Days P-value=0.676 P-value=0.252 P-value=0.005 P-value=0.143 P-value<0.001 P-value=0.015
Annual Rain F = 0.269 F = 0.230 F = 0.102 F = 0.049 F = 0.074 28.1% 1.0% Intensity P-value=0.610 P-value=0.638 P-value=0.753 P-value=0.827 P-value=0.789
Annual Avg F = 44.87 F = 14.98 F = 1.828 F = 14.38 Monthly 7.7% 7.2% 14.7% P-value<0.001 P-value=0.001 P-value=0.193 P-value=0.001 Solar Rad. Annual Avg. F = 0.565 F = 25.49 F = 29.94 Monthly 31.5% 36.3% 1.2% 71.4% P-value=0.462 P-value<0.001 P-value<0.001 Temp. Annual Avg. F = 5.943 F = 0.374 Monthly 6.7% 11.5% 0.6% 45.4% 3.0% P-value=0.025 P-value=0.548 Max. Temp. Annual Avg. F = 10.17 Monthly 43.9% 56.1% 0.3% 9.2% 58.6% 24.8% P-value=0.005 Min. Temp. Annual Avg. Monthly 24.0% 28.6% 0.4% 44.4% 62.4% 2.0% 36.1% Diurnal Range
Page 84 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
E. Correlation Analyses for Noongar Seasons Temperature Metrics
Biruk (Dec-Jan) Average Monthly Avg Monthly Avg Monthly Diurnal Range Maximum Temp Solar Radiation Avg Monthly F = 24.26 F = 19.89
Diurnal Range P-value<0.001 P-value<0.001 Avg Monthly F = 15.90 R2 = 0.574 Max Temp P-value<0.001 Avg Monthly R2 = 0.525 R2 = 0.470 Solar Radiation Bunurua (Feb-March) Average Monthly Avg Monthly Avg Monthly Diurnal Range Maximum Temp Solar Radiation Avg Monthly F = 29.38 F = 13.59
Diurnal Range P-value<0.001 P-value=0.002 Avg Monthly F = 4.139 R2 = 0.620 Max Temp P-value=0.057 Avg Monthly R2 = 0.430 R2 = 0.051 Solar Radiation Djeran (April-May) Average Monthly Avg Monthly Avg Monthly Diurnal Range Maximum Temp Solar Radiation Avg Monthly F = 7.023 F = 7.016
Diurnal Range P-value=0.016 P-value=0.015 Avg Monthly F = 0.144 R2 = 0.281 Max Temp P-value=0.708 Avg Monthly R2 = 0.283 R2 = 0.008 Solar Radiation Makuru (June –July) Average Monthly Avg Monthly Avg Monthly Diurnal Range Maximum Temp Solar Radiation Avg Monthly F = 10.39 F = 10.33
Diurnal Range P-value=0.005 P-value=0.005 Avg Monthly F = 0.130 R2 = 0.366 Max Temp P-value=0.723 Avg Monthly R2 = 0.365 R2 = 0.121 Solar Radiation
Page 85 of 86 Impact of climatic factors on germination of the rare shrub V. staminosa
Djilba (Aug-Sept) Average Monthly Avg Monthly Avg Monthly Solar Diurnal Range Maximum Temp Radiation Avg Monthly F = 20.05 F = 12.57
Diurnal Range P-value<0.001 P-value<0.001 Avg Monthly F = 16.91 R2 = 0.52.7% Max Temp P-value<0.001 Avg Monthly R2 = 0.41.1% R2 = 0.48.4% Solar Radiation Kambarang (Oct-Nov) Average Monthly Avg Monthly Avg Monthly Solar Diurnal Range Maximum Temp Radiation Avg Monthly F = 26.82 F = 6.311
Diurnal Range P-value<0.001 P-value=0.022 Avg Monthly F = 4.846 R2 = 0.59.8% Max Temp P-value=0.042 Avg Monthly R2 = 0.26.0% R2 = 0.21.2% Solar Radiation
Page 86 of 86