This page is archived

Links to external sources may no longer work as intended. The content may not represent the latest thinking in this area or the Society’s current position on the topic.

Tropical phenology: functional change in ecosystems across space and time

15 - 16 January 2018 09:00 - 17:00

Theo Murphy scientific meeting organised by Dr Katharine Abernethy, Ms Emma Bush, Dr Irene Mendoza Sagrera, Professor Patricia Morellato and Dr Jean-François Bastin.

Long-term phenology data from the tropics show directional change in plant reproduction and productivity. However mechanistic links to climate, as well as intercontinental comparisons, remain elusive. This meeting will provide a cross-disciplinary platform to advance new methods and analytical approaches to detecting functional change in tropical ecosystems, conspicuously missing from contemporary models and meta-analyses of global change.

Recorded audio of the presentations are available on this page.

Attending the event

This meeting has taken place.

Enquiries: Contact the Scientific Programmes team.


Schedule

09:00 - 09:05 Welcome by the Royal Society
09:10 - 09:20 Chair’s session overview and introduction to speakers: What long term tropical phenology could contribute to understanding ecosystems: possibilities and obstacles
OLYMPUS DIGITAL CAMERA

Dr Katharine Abernethy, University of Stirling, UK

09:20 - 09:40 The A-Z of litter-trapping methods

Monitoring plant phenology pattern has a long history nonetheless most studies focused on a few species. In recent years, several forest sites, such as Barro Colorado Island (Panama), Pasoh (Malaysia), Fushan (Taiwan), applied seed trap method to monitor flower and seed production. This design contains a substantial number of seed traps and three seedling plots adjacent to each trap in large, stem-mapped forest plots. Weekly collection and sorting of samples captured by traps provides phenological information for many plant species in the forests. Flower/seed data with high temporal resolution and large spatial scale offer opportunities to investigate community-wide phenology patterns, its association with climate, and to quantify seed availability and seed dispersal distance. In addition, hypotheses regarding community assembly can be tested combining data from seed traps and adjacent seedling plots. However, trap efficiency and representation are yet examined. Professor Yu-Yun uses long-term seed trap data from three forest sites to address these questions. Preliminary results indicate that the rate of species accumulation in traps slows down quickly when number of traps increases. However, adding more traps largely increases average number of individual per species. Census interval greater than 2 weeks is not recommended due to disproportional loss of phenology information.

Dr Yu-Yun Chen, National Donghwa University, Taiwan

09:40 - 10:00 Quantifying phenology data through observations: what does the data mean for different questions?

Phenological patterns represent the rhythms of plant production and as such play a central role in regulating access to resources for animals, many of which are now endangered. In biodiverse tropical forests, factors that influence phenology are complex, involving climatic, edaphic, and biotic variables. As such long-term data it needed to unravel the interactions among variables and many of the long-term data sets involve information derived from direct or indirect observation of trees made over decades. Thus, a critical question that must be asked to advance theory and conservation is: What does observational phenology data mean when addressing particular questions? Professor Chapman considers questions that have often been ignored in the past including; How can patterns differ if trees are considered as fruiting or not or if a magnitude of fruiting is considered? What is the magnitude of interobserver variation? How do the observation of trees vary with data collected with fruit traps? What do short-term data (i.e., a few years) that is often collected by researchers studying particular animal species, contribute to our knowledge of changing phenological patterns? When considering how change in phenological patterns drive population size of endangered species, what do we consider food? Addressing these questions over many geographical locations and different time scales, will help advance our quantification of phenological patterns and help comprehend how climate change will impact the future of tropical forests and the animals they support.

Professor Colin Chapman, McGill University, Canada

10:00 - 10:20 State of the art of digital camera methods for phenology

The application of digital cameras to monitor the environment is becoming global and changing the way of phenological data collection. Digital cameras monitoring vegetation phenology (“phenocams”) have an important role by filling the “gap of observations” between satellite monitoring and the traditional on-the-ground phenology. The technique of digital repeated photographs has increased due to its low-cost investment, reduced size, easy set up installation, and the possibility of handling high-resolution near-remote data. The use of imagery data over the traditional phenological observations allows simultaneous multi-sites and long-term monitoring, collecting high-frequency data (daily, hourly), and reduced human labor fieldwork for data acquisition. Phenocams have potential applications for conservation as to document disturbances and changes on vegetation structure, such as deforestation, fire events, flooding and the vegetation recovery. The association of a long-term imagery data with local sensors (e.g., meteorological stations and surface-atmosphere flux towers) allows a wide range of studies, especially linking phenological patterns to climatic drivers; and the impact of climate changes on plant responses. Phenocam networks are growing globally and represent a potential tool for monitoring tropical phenology, for conservation biology and foster networking, as it provides hourly to daily information of monitored systems spread over several sites, ecosystems, and climatic zones, aggregating invaluable information of wide use from phenology to ecosystem dynamics and changes over space and time.

Ms Bruna Alberton, Phenology Lab, São Paulo State University (UNESP), Brazil

10:20 - 10:40 Tropical forest phenology observations from satellites: challenges & opportunities

Space offers a unique vantage point from which to observe the ‘timing’ of recurring phenological phases across Earth’s ecosystems, along with the biotic and abiotic drivers of their timing, and their responses and shifts to climate and environmental change. Satellite sensors are highly calibrated, sophisticated optical instruments that are launched into space to monitor the earth. Satellite data are available at hourly to monthly temporal frequencies; meter to kilometer spatial grids; and include measures of vegetation chlorophyll, fluorescence, air & canopy temperatures, laser scanning, total water storage, active radar and microwave emissions. Through synoptic views, access, and repetitive sampling they generate key measurements that are of immense value in understanding ecosystem functioning. In this talk, Professor Huete will discuss the key challenges and opportunities of satellite based phenological observations across tropical forests. Descriptions of satellite data products, algorithms, data compositing, and quality control/analysis (QA/QC) tools are presented to gain a better understanding of data complexities and improve their utilisation over tropical areas.  We further highlight past and current controversies in satellite phenology applications in the tropics and show how they've advanced our knowledge and improved utilisation of satellite data. Case studies of Amazon greening in the dry season and the confounding influences of seasonality in aerosol loadings from biomass burning, dry and wet season cloud contamination, and seasonal sun angle trajectories are emphasised, all of which can augment forest canopy phenology. Satellite phenology studies of tropical forests are optimised with the use of multiple sensors and in conjunction with ground measurements.  

Professor Alfredo Huete, University of Technology Sydney, Australia

10:40 - 11:10 Coffee
11:10 - 11:30 Wrap-up synthesis: comparative assessment of methods

There are many different ways to observe and record phenology and the current suite of established, long-term, tropical phenology datasets reflects this methodological diversity. Information on the functioning of tropical ecosystems is sparse, and therefore the data recorded in each of these long-term studies is highly valuable. However, the question remains: How can we combine phenology data from different studies to enable meta-analysis of tropical ecosystem function? In this wrap-up talk, Emma Bush will give a comparative assessment of four major methods of phenology data collection - litter-trapping, canopy observations, digital-photography and remote sensing – summarizing for each, the geographic spread across the tropics, the development of protocols and “best practices” for compatibility and the specific ecological information sampled. The major aims of this talk will be to assess which ecological hypotheses can be tested using which data and to identify both opportunities and challenges in combining data for cross-disciplinary, large-scale analyses of tropical ecosystems.

Ms Emma Bush, University of Stirling, UK

11:30 - 12:20 Discussion
12:20 - 12:30 Chair’s final remarks
OLYMPUS DIGITAL CAMERA

Dr Katharine Abernethy, University of Stirling, UK

12:30 - 13:30 Lunch
13:30 - 13:40 Chair’s session overview and introduction to speakers: What would a shared programme of work between specialists in climate science and phenology look like?

Douglas Sheil, Norwegian University of Life Sciences (NMBU), Norway

13:40 - 14:00 Tree phenological monitoring and environmental change in dry tropics

Tree phenological monitoring is of immense significance because it provides a potential tool to address critical questions related to climate change. The tropical dry forests constitute a mosaic composed of several phenological functional types adapted to seasonal drought in different ways. These functional types differ with respect to phenological timing and triggering factors, water relations, extent of deciduousness (~ leaflessness), etc. Deciduousness reflects interacted effect of seasonal drought, tree characteristics and resource use rates. Dry tropical forest tree species show functional diversity in deciduousness, stem wood density (SWD), leaf mass area (LMA) and leaf strategy index (LSI, reflecting resource use rate) to overcome water limitations. Leaf-exchanging short-deciduous species  (having higher SWD and LMA and low LSI) show  flowering coinciding with leaf transitional state when vegetative growth is at its minimum, and both fruit formation and leaf flushing being supported at the same time. In contrast, long-deciduous species (with lowest SWD and LMA, higher LSI) show predominantly dry season flowering, subsequent fruiting on leafless shoots and distinct separation of vegetative and flowering phenophases. Intermediate species show wider flowering range through summer, rainy, autumn or winter seasons. The ability of tree species to withstand (leaf-exchange) or avoid (deciduousness) drought stress and varying seasonal flowering timings appear to be the principal mechanism for successful survival and reproduction under extremely dry and seasonal climate. There is a need to develop capabilities to detect and predict the impact of climate change on deciduousness through long-term phenological network in tropics. This talk will explore the ecological significance of deciduousness in tropical trees with emphasis on inter- and intra-species plasticity in deciduousness, various capacity adaptations related with the duration of deciduousness, and the relationship between tree stem water status and deciduousness. The impending environmental change (i.e. temperature and precipitation) in dry tropics are likely to affect deciduousness and other phonological traits significantly.

Dr Chandra P Kushwaha, Axis Colleges, India

14:00 - 14:20 Towards a robust framework to quantify LAI and radiative transfer variations at tree and landscape levels in the tropics

Unprecedented opportunities are now at hand to bridge the gap between forest seasonal dynamics, flux tower data and apparent temporal signals in vegetation indices. They pass through progress in observational systems at different scales: UAV and ground based LiDAR and multispectral sensors (phenocams) and growing constellations of (nano)satellites. These systems provide detailed structural and temporal information complementing more traditional destructive data collection campaigns, permanent forest plots and tree climbing. The unprecedented amounts of data obtained allow documenting detailed foliage dynamics at tree to landscape levels, and hence in turn to quantify the budget of matter and energy exchange between plants and the atmosphere. This budget can potentially become so resolute as to allow characterizing plant-plant or even within-plant interactions, and thus feed into next generation individual plant growth models (FSPMs). This talk will present some results of ongoing collaborative work in Central Africa and French Guyana, highlight some difficulties as well as intended research directions.

Dr Nicolas Barbier, Botany and Modelling of Plant Architecture and Vegetation, France

14:20 - 14:40 Modelling large-scale vegetation phenology using the JULES land surface model

The Joint UK Land-Environment Simulator (JULES) is the land surface component of the Met Office’s Unified Model. It is a community model that is co-developed by both the Met Office and NERC science community (particularly the Centre for Ecology and Hydrology), and forms a key part of the UK’s contribution to the global model inter-comparison projects (e.g. CMIP6). JULES can be run as a standalone model or as the fully integrated land surface component of the Unified Model for weather forecasts and/or climate and earth system simulations. This presentation will give a brief overview of how vegetation, in particular phenology, is modelled within JULES. It will demonstrate key strengths and weaknesses in the modelling of vegetation phenology using recent results comparing Leaf Area Index (LAI) derived from satellite observations (Copernicus Global Land Service 1km LAI product) with JULES modelled LAI across global biomes and climate zones, and will finish with ideas on how to improve phenological modelling within JULES. A major aim of this presentation is to encourage greater collaboration between ecological experts and climate modellers to ensure the most up-to-date ecological science and observations are included in model developments.

Dr Deborah Hemming, Vegetation-Climate Interactions group, Met Office Hadley Centre, UK

14:40 - 15:10 Tea
15:10 - 16:00 Discussion
16:00 - 16:20 The potential for incorporating phenology in environmental models for global policies

The Food and Agriculture Agency of UN has as mission to achieve food security for all. FAO’s three main goals are: the eradication of hunger, food insecurity and malnutrition; the elimination of poverty and the driving forward of economic and social progress for all; and, the sustainable management and utilization of natural resources, including land, water, air, climate and genetic resources for the benefit of present and future generations. In the framework of supporting developing countries in their fight against emergencies (diseases, disasters, etc.), FAO Forestry helps to develop adoptable, adaptable and feasible tools based on satellite data to help in forecasting of tropical human, and livestock disease as well as nature emergences. Moreover in order to help countries, FAO Forestry assists developing countries to prepare and implement national strategies for global policies (as e.g. REDD+). For the monitoring, reporting and verification, FAO supports the countries to develop national satellite forest monitoring systems that allow for credible measurement, reporting and verification (MRV) of REDD+ activities. These are among the most critical elements for the successful implementation of any REDD+ mechanism. The UN-REDD Programme is supporting countries to develop cost- effective, robust and compatible national monitoring and MRV systems, providing tools, methodologies, training and knowledge sharing that help countries to strengthen their technical and institutional capacity for effective MRV systems. To develop strong nationally-owned forest monitoring systems, technical and institutional capacity building is key. The UN-REDD Programme, through FAO, has taken on intensive training together with other partners (e.g. INPE, USAID, Silvacarbon), and has provided technical help and assistance for in-country training and implementation for national satellite forest monitoring. In this talk, some concrete examples of incorporation of phenology in environmental applications for the global climate change policy will be presented.

Dr Inge Jonckheere, Food and Agriculture Organization (FAO) of the UN, Italy

16:20 - 16:50 Discussion
16:50 - 17:00 Closing synthesis

Douglas Sheil, Norwegian University of Life Sciences (NMBU), Norway

09:00 - 09:10 Chair’s session overview and introduction to speakers: What trends have been detected in tropical ecosystems? What are the analytical challenges and possible solutions?

Dr Irene Mendoza Sagrera, São Paulo State University (UNESP), Brazil

09:10 - 09:35 Environmental controls of the timing of leaf fall and flowering in tropical forests

Many species flower multiple times each year in humid tropical forests. Unfortunately, the many phenology models developed for temperate and boreal plants predict a single first flowering date for each year. Here, we present a new model framework able to predict the timing of multiple annual flowering events. This presentation evaluates Dr Wright's model against 29-year flowering records for Hybanthus prunifolius (Violaceae) and Handroanthus guayacan (Bignoniaceae) from Barro Colorado Island, Panama. Previous experiments demonstrate that both species flower when an extended dry period precedes a brief, intense rain. Our model captures this sequence of rainfall events and, for the first time, quantifies the length and intensity of the preceding drought and the rainfall event. The close match between best-fit model parameters and previous experimental results for these two species suggests the new model framework will provide novel insights for the thousands of humid tropical forest species that lack experimental study.

Dr S Joseph Wright, Smithsonian Tropical Research Institute, Panama

09:35 - 10:00 Phenological strategies and trends witnessed across Africa

Using various data sources, we addressed the diversity of phenological strategies and trends witnessed across Africa, and the consequences for forest management. Using field observations in several sites (n=11) across Central Africa, Dr Fayolle evaluated size at reproduction, a key aspect of species life history, for most major timber species (n=31 species). Dr Fayolle found a great variability in diameter at reproduction among species that was correlated with adult stature (maximum diameter and height). For some species, reproductive diameter thresholds showed substantial variation between sites, and crown exposure to light had a significant positive effect on reproductive status. The majority of minimum cutting diameter limits (MCDLs, imposed by national forest administration) was found to be suitable, trees having high probability for being seed trees at MCDL. Using field and herbarium data, Dr Fayolle examined the spatial variation in flowering phenology across the whole range of species, and using remotely-sensed data of vegetation activity (EVI from MODIS satellite), Dr Fayolle specifically examined the reversed phenology on both side of the climatic hinge, the area in Atlantic Central Africa where the rainfall seasonality reverses. Dr Fayolle found that herbarium data can be used to derive phenological information, and specifically the onset of flowering, probably because most central African tree species have a regular and annual phenology. She also found strong support for the spatial variation in tree phenology associated with the shift in rainfall seasonality, for flowering (field and herbarium data) and for forest functioning (EVI seasonality), with however, differences among species.

Dr Adeline Fayolle, Gembloux Agro-Bio Tech and University of Liege, Belgium

10:00 - 10:20 Discussion
10:20 - 10:50 Coffee
10:50 - 11:15 What are the analytical challenges and possible solutions?

Using a case study of long-term records of the flowering and budding of eight eucalypt species, E. camaldulensis, E. goniocalyx, E. leucoxylon, E. macrorhyncha, E. melliodora, E. microcarpa, E. polyanthemos and E. tricarpa Professor Hudson will examine the advantages of 2 analytic approaches:  singular spectrum analysis (SSA) and wavelets. Professor Irene Hudson and Dr Marie Keatley, recently showed at an individual species level, SSA decomposition/reconstruction/ (cross-)correlation analysis: determined the strength, directionality and lagged nature between climate and flowering; identified the primary climatic drivers at peak flowering intensity, and confirmed the dynamic nature of the relationship between peak flowering and climate over the year. This study showed how SSA can add further understanding about the interplay between climate and flowering. Change point years for flowering based on SSA sub-components align with some years of major shift in global ENSO signal. Using  wavelets Hudson and Keatley (2015a, b) investigated the influence of climate on budding, and the effects of current and past budding (up to 3 months prior), and climate on peak flowering for E. tricarpa and E. leucoxylon. The commencement of the budding and flowering phenophases were generally 4 months apart. Minimum temperature positively influenced bud development (above a min temp), and maximum temperature negatively influenced budding (below an upper threshold). This complex interplay between climate and budding on flowering needs further work with the examination of additional species, but, given that flowering is dependent on budding, our postulate makes sense. Unravelling the additional role of budding in all eight species is a focus for future research, using multivariate time series models, methods not used to date to unravel flowering and budding, with potential value-add to phenology studies involving multiple phenophases.

Professor Irene Hudson, Swinburne University of Technology, Australia

11:15 - 11:40 Forest productivity change over time

A synthetic understanding of the impacts of global environmental changes on tropical tree phenology will require an understanding beyond flower, fruit and seed production. Such changes may co-occur with increases or decreases in net primary productivity (NPP), with rising NPP potentially suggesting more resource availability for plants to allocate to reproduction. In this talk I assess recent evidence for large-scale regional, continental and pan-tropical changes in forest productivity, with a particular focus on utilising data from long-term inventory plots across Amazonia (RAINFOR network), Africa (AfriTRON network), and SE Asia (T-FORCES network) which can assess changes in woody production across many tropical forests. By combining with satellite and model estimates of changes in forest productivity it is possible to make clear statements about these changes for large areas of the tropics.

Professor Simon Lewis FRS, University College London and University of Leeds, UK

11:40 - 12:20 Discussion
12:20 - 12:30 Closing synthesis

Dr Irene Mendoza Sagrera, São Paulo State University (UNESP), Brazil

12:30 - 13:30 Lunch
13:30 - 13:45 Chair’s session overview and introduction to speakers: Present and future of tropical phenology research

Professor Patricia Morellato, São Paulo State University, Brazil

13:45 - 14:00 The Nature’s Calendar experience – from legacy data to citizen science

The fortunes of phenological recording schemes have waxed and waned (prospered and faltered) in different decades in different countries. Government-sponsored schemes that incorporate fixed recording locations have experienced funding restrictions that have inevitably led to a reduction in coverage. The alternative format of Citizen Science schemes has become well established in several countries, including the USA, Netherlands and UK. In the UK, the current Nature’s Calendar scheme was piloted in 1998 and has since been involved in both the recording of current data as well as backloading the legacy of older data, going as far back as the late 17th century. The database now contains several million records. Such Citizen Science schemes have their own unique issues, including recruiting, retaining and motivating their recorders. Publicity, for example by national TV coverage, can boost numbers hugely, but such support can be ephemeral. This presentation will discuss some of the challenges concerning Citizen Science schemes, legacy data and the merging of the two. Examples of how the data have been used, and what they show, will also be presented.

Professor Tim Sparks, Coventry University, UK

14:00 - 14:15 Discussion
14:15 - 14:40 Linking phenology across the North and South Hemispheres: thoughts on building a framework for networking

Phenology around the world has benefited from collaborations between regions and countries, but to date, collaborations between the northern and southern hemisphere are limited. The benefits of a formal north-south phenological network are many. It would establish a common language, contribute to the sharing of methods and ensure sensible consistency in details such as a standard reference period and data collection protocols. Such a network would contribute to the global picture of the effects of climate change as well as programs aimed at adapting to climate change, international policy and conservation efforts. It would open up new opportunities for funding and also provide a basis for assisting countries with less well-developed phenological programs. Establishing a north-south phenological collaboration framework requires a number of decisions to be made. For example, Is phytophenology the focus? At what level(s) should such a framework operate: between individuals, scholarly societies, academic institutions, governments or inter-governmental agencies? What role should the World Meteorological Organization and the Phenology Commission of the International Society of Biometeorology play? How would the framework function in terms of establishing standards for collection, recording and reporting of phenological data? The aim of this presentation is not to resolve these issues, but to open the discussion and hopefully motivate champions to develop an effective north-south phenological network.

Dr Marie Keatley, University of Melbourne, Australia

14:30 - 14:45 Discussion
14:45 - 15:15 Tea
15:15 - 15:30 Phenology, phylogeny and species distribution modelling applied for conservation

Climate change affects plant phenology culminating in temporal shifts such as early or late reproduction, detected by the analyses of long-term time series. Phenological changes have serious consequences for both animals that rely on plants for food resources, and for plant communities that undergo alterations in their dynamics and dispersion that ultimately alter species’ composition and geographical distribution. The effects of climatic change on tropical plant phenology are still poorly understood mainly due to the lack of long-term studies. Most phenological data from tropical regions are restricted to local scales and cover short periods of time (1-3 years). In the Phenology Lab Dr Staggemeier used two alternative approaches to understand and predict potential effects of climate change in the tropics. In the first, she combines ecological niche modelling with phenological information based on herbarium records to estimate the potential responses predicted for tropical plants. Dr Staggemeier predicts size, displacement and direction of change in geographical distribution in light of future global warming scenarios. In the second approach, Dr Staggemeier and her colleagues are mapping the flowering phenology in the phylogeny to determine the occurrence of phylogenetic signal in phenological patterns. This last approach allows to identify clade sensitivity to climate change. The analyses presented in this talk were applied to Atlantic forest Myrtaceae and the results indicating early reproduction in tropical plants could minimize the negative effects of global warming. Additionally, clades with a recent history of speciation showed conservative flowering patterns. This presentation will show both approaches are good tools, providing a more comprehensive assessment of future functioning and dynamics of tropical forests under climatic change.

Dr Vanessa G Staggemeier, Phenology Lab, São Paulo State University (UNESP), Brazil

15:30 - 15:45 Discussion
15:45 - 16:05 Linking phenology to ecosystem carbon budgets and trade-offs

Controversy remains over the roles of precipitation, solar irradiance, and temperature as the main drivers of seasonal dynamics of the net primary productivity of tropical forests. The relationship between seasonal peaks in precipitation and wood productivity is relatively well established for Amazonian forests, but the predominant drivers of the seasonal production of leaf, flowering, and fruiting remain unclear. To explore the dynamics of leaf and reproductive phenology in tropical forests, we need to improve our understanding of the seasonal trends of carbon allocation in trees between canopy, aboveground coarse woody material, and fine root production. A question that remains unresolved is the extent to which a seasonal increase in canopy leaf production reflects an increase in overall productivity, or alternatively simply a shift in carbon allocation amongst these components. Dr Girardin recently demonstrated that allocation shifts are more important than overall changes in productivity in explaining interannual variability in canopy, wood, and fine root growth rates in a Bolivian forest. Here, she investigates the physical drivers of carbon allocation across a wider range of tropical forest sites: is there evidence of clear trade-offs in carbon allocation driving seasonal variation of woody growth, canopy production, and fine root production? If so, how do these trends vary across sites? Finally, this talk will explore the nature and timing of reproductive phenology in tropical forests, how this relates to the overall seasonality of climate and productivity, and the amount of productivity invested in reproduction. Systematic ground-based monitoring of the major components of biomass productivity is needed to understand the drivers of annual forest carbon allocation seasonality. The Global Ecosystems Monitoring network has been monitoring the carbon cycle in 17 sites across the tropics since 2009. We propose to explore the seasonal trends of carbon allocation in lowland Amazonia, central Africa, and Borneo.

Dr Cécile Girardin, Oxford University, UK

16:05 - 16:55 Panel discussion
16:55 - 17:00 Final remarks and future directions