Functional Ecology ( IF 4.6 ) Pub Date : 2023-05-14 , DOI: 10.1111/1365-2435.14353 Susanne Kortsch 1 , Leonardo Saravia 2 , Alyssa R. Cirtwill 1 , Thomas Timberlake 3 , Jane Memmott 3 , Liam Kendall 4 , Tomas Roslin 1, 5, 6 , Giovanni Strona 5, 7
|
1 INTRODUCTION
The arrangement of plants in a landscape influences pollinator movements and plant–pollinator encounter rates (Cranmer et al., 2012). This influences visitation frequency and pollination success (Fortuna et al., 2008), which, in turn, affect plant fitness and yield. The level of a pollinator species' generalism affects which flower resources the pollinator can utilise and can therefore play an important role in modulating pollinators' responses to changes in the spatial arrangement of plants at the landscape level. Thus, the interactions between the spatial arrangement of plants and pollinator specialisation are expected to determine visitation rate and, ultimately, pollination success.
It is important to note that from a plant perspective, pollination success is not logically equivalent to the rate at which a plant is visited by a pollinator (King et al., 2013). A direct link between the two assumes that every individual insect which visits an individual plant will deposit conspecific pollen at each visit (Bascompte et al., 2003; Memmott, 1999). Such a scenario, however, is unlikely, because pollinators often visit several plants of different species between visits to conspecific plants. Moreover, pollen transfer is a highly stochastic and wasteful process (Johnson, 2010; Richards et al., 2009). Despite this, given the complexities of measuring effective pollination empirically, many researchers rely on visitation rate as a surrogate measure for pollination. Although visitation rate captures how often pollinator species visit or forage on a given plant species, and thus describe the importance of plants from the pollinator perspective, it disregards the sequence of flower visits by individual pollinators. This is a major shortcoming, given the crucial importance of this sequence in determining the chances of successful conspecific pollen transfer and pollination (King et al., 2013; Morales & Traveset, 2008). As a result, most plant–pollinator network studies using visitation rates do not accurately describe pollination success from the plant perspective (Armbruster, 2017; de Santiago-Hernández et al., 2019; Willmer, 2011).
The distinction between visitation rates and pollen transfer has been repeatedly stressed by several authors (King et al., 2013; Willmer, 2011), as has the importance of evaluating the link between the two. Nonetheless, studies attempting to fill this major knowledge gap are rare (Ballantyne et al., 2015, 2017). Some of the challenges in moving from visitation rate to better estimates of pollination success derive from the crucial distinction between pollinators as species and individuals (Cirtwill et al., under review). Effective pollination requires that a pollinator individual visits the same plant species consecutively, or at least repeatedly, during a plant visitation sequence. Yet, most field-based studies are unable to provide such detailed information at the level of the individual (but see, Arroyo-Correa et al., 2021; Dupont et al., 2014). Instead, information is provided at the level of species, genus or even family. Because of the difficulty in obtaining visitation sequences of individual pollinators (and hence their contribution to conspecific pollen deposition on plant individuals) from field data, the knowledge gap persists—despite recent efforts to disentangle mere visits from effective pollination (Arroyo-Correa et al., 2021; Ballantyne et al., 2015; King et al., 2013). Given the many challenges of assessing pollination success, even in a laboratory or greenhouse setup, the relationship between a given spatial configuration of plants, pollinator specialisation and pollination remains elusive and largely untested (Armbruster, 2017).
Individual-based models can circumvent some of the limitations faced by field studies because they allow the tracking of individual pollinators and their floral visitation through space and time (Newton et al., 2018). Individual-based models permit recording the time and location of all events taking place during a simulation, together with the identity of the individuals (e.g. plants and pollinators) involved in the event. Thus, a properly designed individual-based model can be used to simulate plant–pollinator interactions under a broad range of eco-environmental scenarios, while recording the complete sequence of floral visitation by pollinators and their exact spatio-temporal history. This, in turn, provides a unique opportunity to explore the dynamics of plant–pollinator interactions and disentangle the process of visitation from that of conspecific pollen transfer, therefore improving our understanding of the determinants of pollination success in a spatial context. Additionally, individual-based models are extremely flexible, providing users with freedom to design and manipulate simulation settings. Such flexibility allows the exploration of how variation in the emergent features of pollinator–plant interactions resulting from individual-level behaviour affects the pollination process by, for example, imposing different levels of specialisation for pollinators. Hence, these models can elucidate structure–function relationships and have the potential to catalyse the transformation of network approaches from descriptive to more predictive science (Arroyo-Correa et al., 2021).
In this study, we use an individual-based modelling approach to address how the spatial arrangements of individuals of different plant species within a landscape combine with pollinator specialisation to influence pollinator visitation rates. We further model the probability of conspecific pollen transfer in different visitation sequences to derive the expected number of plants pollinated. To do this, we formalise two functional metrics of pollination as well as visitation rate: consecutive visits, which describes the number of times an individual pollinator visits the same plant species twice in a row during a visitation sequence, and the expected number of plants pollinated based on pollen contributions from all previous visits to conspecific plants along a visitation sequence. The latter assumes a geometric decay of pollen (i.e. decrease in pollen transfer rates) between visits along the visitation sequence (Bateman, 1947; Harder, 1990).
We expect that the mean number of plants pollinated (per day) in our model differs between metrics considered. Trivially, we expect higher values for visitation rate than for consecutive visits and expected number of plants pollinated, because visitation rate includes all pollinator visits regardless of the position of plant species along each pollinator individual's visitation sequence. In contrast, we expect similar trends in consecutive visits and expected number of plants pollinated, as both these measures take the position of plant species visited along the visitation sequence (and hence the probability of conspecific pollen transfer) into account. A priori, we expected the level of pollinator specialisation to affect pollinator behaviour and plant–pollinator encounter rates, and hence our proxies for pollination. For specialist pollinators, nearly all visits will be consecutive visits to the same plant species. For more generalist pollinators, however, we expect our three proxies for pollination to differ substantially as more heterospecific plant visits occur. All three proxies for pollination are expected to vary in response to the spatial arrangement of plants in the landscape (i.e. with the level of plant intermixing). Furthermore, we expect the level of specialisation to interact with landscape structure in determining pollination success. For example, we expect more consecutive visits to the same plant species in landscapes characterised by no or low plant intermixing, especially for generalist pollinators.
中文翻译：
在基于个体的模型中，景观组成和传粉媒介特征相互作用，影响授粉成功
1 简介
景观中植物的排列会影响传粉媒介的运动和植物与传粉媒介的相遇率（Cranmer et al., 2012）。这会影响访问频率和授粉成功（Fortuna 等， 2008），进而影响植物的适应性和产量。传粉媒介物种的普遍性水平影响传粉媒介可以利用哪些花卉资源，因此可以在调节传粉媒介对景观水平植物空间排列变化的反应方面发挥重要作用。因此，植物的空间排列和传粉媒介专业化之间的相互作用预计将决定访问率并最终决定授粉成功。
值得注意的是，从植物的角度来看，授粉成功在逻辑上并不等于传粉者访问植物的速度（King 等，2013 ）。两者之间的直接联系假设访问单个植物的每个昆虫都会在每次访问时沉积同种花粉（Bascompte 等， 2003；Memmott， 1999）。然而，这种情况不太可能发生，因为传粉者经常在访问同种植物期间访问不同物种的几种植物。此外，花粉转移是一个高度随机且浪费的过程（Johnson， 2010；Richards 等， 2009））。尽管如此，鉴于凭经验测量有效授粉的复杂性，许多研究人员依靠访问率作为授粉的替代指标。尽管访问率反映了传粉媒介物种访问或在给定植物物种上觅食的频率，从而从传粉媒介的角度描述了植物的重要性，但它忽略了单个传粉媒介访问花朵的顺序。这是一个主要缺点，因为该序列在确定同种花粉转移和授粉成功机会方面至关重要（King 等人， 2013 年；Morales 和 Traveset， 2008 年））。因此，大多数使用访问率的植物-传粉媒介网络研究并不能从植物的角度准确描述授粉成功（Armbruster， 2017；de Santiago-Hernández 等， 2019；Willmer， 2011）。
一些作者反复强调访问率和花粉传播之间的区别（King 等， 2013；Willmer， 2011），以及评估两者之间联系的重要性。尽管如此，试图填补这一重大知识空白的研究却很少（Ballantyne et al., 2015 , 2017）。从访问率转向更好地估计授粉成功的一些挑战源于授粉媒介作为物种和个体之间的关键区别（Cirtwill 等人，正在 审查中）。有效的授粉需要授粉者个体在植物访问序列期间连续地或至少重复地访问相同的植物物种。然而，大多数实地研究无法在个人层面提供如此详细的信息（但请参阅 Arroyo-Correa 等人， 2021；Dupont 等人， 2014）。相反，信息是在种、属甚至科的层面上提供的。由于从田间数据获取个体授粉媒介的访问序列（以及它们对植物个体同种花粉沉积的贡献）的困难，知识差距仍然存在——尽管最近努力将单纯的访问与有效授粉分开（Arroyo-Correa等人，2015）。 ， 2021 年；Ballantyne 等人， 2015 年；King 等人， 2013 年）。鉴于评估授粉成功与否面临许多挑战，即使在实验室或温室设置中，给定的植物空间配置、授粉媒介专业化和授粉之间的关系仍然难以捉摸且基本上未经测试（Armbruster，2017 ）。
基于个体的模型可以规避实地研究面临的一些限制，因为它们允许通过空间和时间跟踪个体授粉者及其花卉访问（Newton et al., 2018 ））。基于个体的模型允许记录模拟期间发生的所有事件的时间和地点，以及参与该事件的个体（例如植物和传粉者）的身份。因此，正确设计的基于个体的模型可用于模拟广泛的生态环境情景下的植物与传粉媒介的相互作用，同时记录传粉媒介访问花卉的完整序列及其确切的时空历史。反过来，这提供了一个独特的机会来探索植物与传粉者相互作用的动态，并将访问过程与同种花粉转移的过程分开，从而提高我们对空间背景下授粉成功的决定因素的理解。此外，基于个人的模型非常灵活，为用户提供设计和操作模拟设置的自由。这种灵活性允许探索个体水平行为导致的传粉者与植物相互作用的新兴特征的变化如何影响授粉过程，例如，通过对传粉者施加不同程度的专业化。因此，这些模型可以阐明结构-功能关系，并有可能催化网络方法从描述性科学向更具预测性的科学转变（Arroyo-Correa 等人，对传粉者实行不同程度的专业化。因此，这些模型可以阐明结构-功能关系，并有可能催化网络方法从描述性科学向更具预测性的科学转变（Arroyo-Correa 等人，对传粉者实行不同程度的专业化。因此，这些模型可以阐明结构-功能关系，并有可能催化网络方法从描述性科学向更具预测性的科学转变（Arroyo-Correa 等人， 2021）。
在这项研究中，我们使用基于个体的建模方法来解决景观中不同植物物种的个体的空间排列如何与传粉媒介的专业化相结合来影响传粉媒介的访问率。我们进一步对不同访问序列中同种花粉转移的概率进行建模，以得出预期授粉植物的数量。为此，我们形式化了授粉和访问率的两个功能指标：连续访问，它描述了单个授粉者在访问序列期间连续两次访问同一植物物种的次数，以及预期授粉的植物数量基于之前所有访问同种植物的花粉贡献（按照访问顺序）。后者假设沿着访问序列的访问之间花粉呈几何衰减（即花粉转移率降低）（Bateman， 1947；Harder， 1990）。
我们预计模型中授粉植物的平均数量（每天）在所考虑的指标之间有所不同。简单地说，我们预计访问率的值高于连续访问的值和预期授粉植物的数量，因为访问率包括所有传粉媒介的访问，无论植物物种在每个传粉媒介个体的访问序列中的位置如何。相比之下，我们预计连续访问和预期授粉植物数量的趋势相似，因为这两种措施都考虑了沿访问序列访问的植物物种的位置（以及同种花粉转移的概率）。首先，我们预计传粉媒介的专业化水平会影响传粉媒介的行为和植物与传粉媒介的相遇率，从而影响我们的授粉代理。对于专业传粉者来说，几乎所有访问都是对同一植物物种的连续访问。然而，对于更通用的授粉媒介，我们预计随着更多异种植物访问的发生，我们的三个授粉代理将有很大不同。预计授粉的所有三个代理都会根据景观中植物的空间排列（即植物混合的水平）而变化。此外，我们预计专业化水平与景观结构相互作用，决定授粉成功。例如，我们预计在没有或很少有植物混合的景观中，人们会更多地连续访问相同的植物物种，特别是对于通才传粉者来说。
京公网安备 11010802027423号