TITLE: Interview presentation for SECO post-doc position
DATE: 2021-06-23
AUTHOR: John L. Godlee
====================================================================
I applied for a post-doc position with the SECO project recently. I
got the job, so I thought I would post the presentation I gave as
part of the interview. The idea for the presentation was to give a
12 minute run-down of "my previous work, how it links to the
post-doc position, and my plans for the position". I've pasted the
slides below with an approximate script of what I said.
[SECO]:
I'll tell you about my previous work and how it links to SECO, give
you an idea of my general research interests, and tell you some of
the ideas I have for this post-doc and what interests me about the
science of the project.
Before I dive into my PhD research, I wanted to highlight some
other projects I've been involved in which I think provided me with
valuable experience that I can apply to this post-doc.
For my undergraduate dissertation I studied tree seedlings in the
Andes and climate induced range shifts. I used leaf functional
traits and physiological stress responses to predict the
elevational limits of different species.
In Canada I worked as a field assistant, on a project studying the
effects of climate warming on the functional traits of shrubs. We
used pheno-cams to monitor phenology, to make predictions about
changes in vegetation structure.
I worked with Lucy Rowland from University of Exeter as a research
assistant on a drought experiment in the Brazilian Amazon, where we
looked at hydraulic traits and leaf gas exchange to investigate the
effect of drought on tree mortality.
In 2017 I managed the fieldwork for a fire experiment in the
Republic of Congo, where I supervised a local field team for 5
weeks, doing mortality surveys and plot burning.
Now I'm getting towards the end of my PhD, which I hope to hand in
before the end of August 2021.
Next I'll cover the skills and insights from my PhD that I think
will be useful for the post-doc.
My PhD is on how tree species composition and diversity affects
woody biomass and ecosystem structure in southern African savannas.
My thesis uses a combination of original field data, plot census
data from SEOSAW, and various remote sensing data.
One of the key parts of the thesis was a large data synthesis model
using 1200 woodland plots in a structural equation modelling
framework to determine how environmental context affects the
relationship between species diversity and biomass.
Then I conducted two studies looking at specific ecological
mechanisms that might drive diversity-productivity relationships in
savannas. The first used plot data and MODIS data to investigate
the effect of species composition on pre-rain green-up and
senescence in Zambia. The second used terrestrial LiDAR data to
investigate how spatial diversity metrics affect canopy structure
and foliage density.
Also, as part of the PhD I set up 15 permanent plots in Bicuar
National Park in Angola with in-country colleagues. We wrote a
paper together on the floristic diversity of the park in relation
to other miombo formations.
I think my PhD project reflects my interest in how community
assembly drives biomass dynamics in different environmental
contexts, and also reflects my interests in both macro-ecology
style modelling and plot-based field studies.
Last year, I took a break from PhD to manage the SEOSAW database
for 5 months.
The original plan was to manage the establishment of new permanent
plots in under-studied areas, but COVID meant we had to change
plans and I took on more of a database manager role.
I spent a lot of time formalising the cleaning process for the
database, to make it more reproducible, including an overhaul of
the species name checking system.
I also incorporated new datasets from various SEOSAW partners and
supported them in data processing and data collection.
I wrote documentation and created digital data collection forms for
colleagues, both as a capacity building tool and to ensure data
quality, and I drafted a mortality recording system that is more
suited to the dry tropics where trees resprout.
I think this experience especially puts me in a really good
position to critically analyse plot data we might get in SECO.
Now I'll switch to talking about SECO itself.
We know that variability in the dry tropical carbon sink is a major
source of uncertainty in the global carbon cycle. There are various
drivers of change pushing dry tropical vegetation in opposing
directions. Importantly, it's been suggested that atmospheric CO2
enrichment will increase woody biomass, particularly in savannas,
by reversing the competitive balance between C4 grasses and C3
trees. But there are also expected to be concurrent changes in
rainfall seasonality, temperature and fire regime.
It's important to consider that there are many functionally
distinct vegetation formations within the dry tropics, even in
similar environmental contexts. Across each continent, dry tropical
vegetation formations are more related to their moist forest
neighbours than to dry tropical formations on other continents.
Meaning these vegetation types are likely respond differently to
drivers of vegetation change based on their adaptive potential.
An interesting avenue for us to explore I think will be using
growth and mortality data from the SECO plots, combined with
functional trait data of species within the plots and their
phylogenetic relationships, to identify the bioclimatic envelopes
of these different vegetation types, and to use that demographic
and functional information to predict which drivers of change are
likely to have the greatest effect on biomass in different regions.
Of particular interest to me is whether the functional diversity of
the species pool will constrain the ability of different vegetation
formations to succeed following environmental change. We might
expect for instance that the comparatively species rich dry forests
and savannas of the neotropics might contain enough functional
redundancy so that when the climate changes they can maintain
productivity, albeit with differences in species composition, while
in the species poor savannas of southern Africa, climate change may
severely reduce productivity as fewer species can fill the niche
space.
As an aside, the map on this slide was my attempt to re-create the
map of the SECO working region which I found on the SECO website. I
think just defining the working region of the dry tropics will
prove to be an interesting challenge. Doing this exercise really
highlighted to me that India and southeast Asia will be
particularly controversial, due to their very different structure
to the rest of the tropics. Maybe we can help to find unifying
conditions for the dry tropics during SECO.
We know that one of the key goals of SECO is to constrain important
sources of variation in the carbon cycle of the dry tropics. In
this next section I'll talk about specific biological mechanisms
that I think will be important in driving variation in carbon
cycling, and how the SECO plot data can improve our understanding
of those processes.
This flow chart should be familiar to the panel, it's the DALEC
carbon cycle model, which I imagine will play some part in the
modelling aspect of SECO. I thought the model would be a good way
to structure my thinking on what biological mechanisms affect
variation in the carbon cycle across the dry tropics. I've marked
some key processes in the model which are likely to drive
variability in the carbon cycle across the dry tropics, and I'll
discuss how our understanding of each could be improved through
analysis of the SECO plot data. 
First: Mortality.
Trees in the dry tropics have incredibly complex mortality
dynamics. They resurrect, resprout, and lose parts of the plant to
different extents depending on species, environmental pressures,
and disturbances. We need mortality protocols that are tailored to
the dry tropics and which consider these complex dynamics,
otherwise we could misrepresent the mortality signal.
Particularly I think we should investigate further how fire and
drought affect resprouting behaviour and mortality, because both
these environmental pressures are expected to vary a lot in the
future.
We should look at how this varies among vegetation types, among
species across the conservative/acquisitive trait spectrum, and
among different demographic groups so we can make predictions about
how future vegetation composition will look.
To understand how different vegetation types are expected to fare
under climate change, we need to understand how close their
constituent species are to their physiological limits. We can
assess this using mortality and growth, or even using direct
ecophysiological methods for a subset of plots.
Next: Root:shoot ratios
Allocation to different carbon components varies according to
environment and species life history strategy. Variation in carbon
allocation has knock on effects for decomposition and carbon
residence times as different carbon components turn over at
different rates.
Allocation to wood and roots is normally modelled in large carbon
cycling models as a constant root:shoot ratio. But that ratio is
often poorly parameterised, especially in the dry tropics, and
doesn't account for continental variation in adaptive traits and
evolutionary history that means species allocate differently under
similar environmental forcings.
For example, Australian and African savanna tree species differ in
fire avoidance strategy. Australian Eucalypts grow tall and thin,
while African species more readily invest in root structures,
meaning these two savanna types are likely to vary in how they
allocate to above- and below-ground structures under a changing
environment.
We need to do more to identify environmental and biogeographic
conditions that lead to variation in the root:shoot ratio across
the dry tropics, because without understanding roots we might
majorly misrepresent carbon storage.
Thirdly: leaf traits
Leaf traits and leaf carbon allocation can tell us about
productivity in a system, being the organ chiefly responsible for
photosynthesis. In savannas it can also tell us about competitive
relationships with grasses, and also links to decomposition and
phenology in deciduous systems.
In carbon cycling models leaf carbon allocation most often
estimated using remote-sensing products like Copernicus Leaf Area
Index, but the coexistence of grass and trees in savannas means the
signal is often noisy, and this is complicated even more by
seasonal variation in the relative grass:tree contribution.
We could use terrestrial LiDAR coupled with leaf trait measurements
that represent the leaf economic spectrum to precisely estimate
leaf carbon at selected sites. Then set that in the context of
environmental factors and other aspects of plant physiognomy to
make predictions about the effects of climate change on leaf
production in different functional groups of trees. Then we can
pair these on-the-ground measurements with satellite data, and
phylogenetics to scale up across whole biomes.
Basically there are trade-offs in the leaf and root economic
spectrum which determine allocation to different carbon components
have knock on effects for mortality and decomposition with
consequences for the entire carbon cycle.
To summarise this presentation, I think my combination of skills,
which I've developed in the dry tropics, and which all contribute
to my in depth understanding of carbon dynamics and biogeography as
it applies to this biome, set me up really well to make a decent
contribution to the SECO project.
I hope you can see that I have lots of ideas about this project,
even though I've only had the time to cover a few of them. I think
there are loads of avenues for us to produce research which not
only answers the 4 key questions posed by SECO, but also answers
some intermediate and more mechanistic questions which could shed
light on how the dry tropics operate, both floristically and
functionally.
