Jake Dalzell

by Jake Dalzell, Hazel Garrett, Catriona Forrest, Wayne Liang, Rosalind Mackey, Denis Pavlov, and Josh Simpson  This was a small project we threw together over three days on our Plant Sciences fieldtrip to Portugal. Everyone on the fieldtrip used six different techniques to explore plant physiology and ecology, and each group came up with a research question that could be answered using some of these techniques.

Comparing parasite and host strategies in a Portuguese grassland

7/10/2023 UPDATE: I have discovered other authors came up with this idea before me, as noted by Press and Phoenix¹¹. At least two authors have mentioned it: in the context of mistletoes¹²; and a North American root hemiparasite,

 Triphysaria

¹³. However, as far as I can tell the hypothesis has never been tested.

Total ecosystem evapotranspiration – an overlooked factor in grassland hemiparasite ecology?

In July I spent a few days in Sligo as a field assistant with an international, interdisciplinary team who are studying peritidal stromatolites. Stromatolites are layered rock formations created by microbes, and peritidal stromatolites form in the area of the shore that is affected by seawater from the tide or from sea spray. This makes them unique as they can experience large changes in salinity.

Sligo Stromatolites

This summer I am cataloguing the flora and plant communities of twelve islands [EDIT 20/11/22: it ended up being eleven islands!] in Strangford Lough, with the help of a BSBI Plant Study Grant. The goal is to have a complete species list of each island and a map showing all of the plant communities and where they occur. I have been kayaking to these islands and doing a thorough survey, and also quadrats in each distinct plant community.

The Islands Project

Since 2022 I have been recording the flora of some islands in Strangford Lough. I received a Plant Study Grant from the BSBI to work on this in the Summer of 2022 but as it turned out I had committed to a much bigger project than I first thought! I recently submitted my records and can now finally say I have finished my first large recording project.

Strangford Lough flora: an update

This is an adapted version of my poster, which is also available as a pdf or png. As with all my work on this site it is CC BY 4.0 (free for use/adaptation as long as I am credited).

The Meadow Maker Yellow Rattle (

 Rhinanthus minor

) has become well known in restoration ecology and wildlife gardening as a “meadow maker”, which can transform a species-poor patch of grasses into a biodiverse and flower-rich one.

Yellow Rattle

The model in this post is based on the one in Cameron et al. (2009), improving how seed dispersal is modelled and the dynamics are visualised.

I have become a bit obsessed with hemiparasitic plants recently. When I found Cameron et al.’s paper on how Yellow Rattle can create interesting spatial dynamics in grasslands¹, I got quite excited. I have been learning python this year so I thought I may as well try to replicate their model.

Modelling Yellow Rattle in a simulated grassland

My favourite butterfly, which I first saw during lockdown. They are the most delicate of the white species (Pieridae) found in Ireland and flutter through grasslands in May and June. Taxonomy For a long time, scientists wondered why Wood Whites (

 Leptidea sinapis

) were fairly widespread and common in Ireland, when they were so rare in England.

The Cryptic Wood White butterfly

As someone who has been deeply interested by plants as long as I can remember, I have grown to appreciate all the different habitats I come across in Ireland. Each has its own unique set of species, but one that has always particularly caught my attention is our grasslands.

Why care about grasslands?

Today I had the opportunity to work as a field assistant with a PhD student, Thomas Marquand. He is researching CO₂ and methane dynamics in soil in restored and reclaimed fens in East Anglia (with the Centre for Landscape Regeneration). We took samples from the fen exhibit in the Cambridge Botanic Gardens and used a spectrometer to measure CO₂ and methane concentration, and also δ¹³C-CO₂.

Fen soil carbon flux

Today I went to the BSBI’s November conference in the Natural History Museum in London! I presented my poster on my summer project surveying islands on Strangford Lough, and got to do a 60 second flash talk. It was really great to meet other people who are passionate about plants and learn from them. Read the html version of my poster here. Watch my (very brief!) flash talk here.

British and Irish Botanical Conference, November 22

03/11/2023 EDIT:

actually a rediscovery, this species was recorded in Belfast pre-1930 Earlier this year, I found a species of grass I didn’t recognise, growing as a pavement plant in East Belfast. With a guide I identified it as

 Polypogon monspeliensis

(Annual Beard-grass), a new species for County Down. This species is known from Dublin and Cork, and there is a previous Ulster record from Keady in Armagh.

Annual Beard-grass: a new species for County Down (H38)

Ants visit a Sea Campion flower infected by a

 Microbotryum

smut fungus. Disease is something few people love. There are exceptions though, particularly among those who appreciate the diversity and ecological role of plant disease.

In Defence of Disease

Hemiparasitic plants are those which steal water and nutrients from other plants, but still photosynthesise (so are green). They are an interesting part of the Irish flora, and some can play an important role in grasslands by weakening other plants – particularly Yellow Rattle.

Hemiparasitic plants in Irish grasslands

Slievenacloy Nature Reserve is an Ulster Wildlife reserve in Antrim. It is very well managed for upland grassland, with a wealth of Devil’s-Bit dominated grassland.

Plasmopara euphrasiae: a downy mildew species new to Ireland

On the thirteenth of August, I was at Killard National Nature Reserve, Co. Down, with my mum. We were looking for Frog Orchids and other species rare in the county like Quaking Grass and Field Scabious. She found a plant she didn’t recognise and asked me what it was, and it was Autumn Lady’s Tresses. I knew I hadn’t seen it before and wasn’t sure it had been recorded anywhere near before.

Autumn Lady’s Tresses: an orchid species new to Ulster

In the nineties, the Belfast Naturalists’ Field Club extensively surveyed the urban flora of Belfast, resulting in a book,

 Urban Flora of Belfast

. Many plants new to Down and Antrim were found, and most of these were new to Ireland as well. This was a three-year effort involving some thirty-one recorders.

Belfast’s urban flora

Irish Plants

*The model in this post is based on the one in Cameron et al. (2009),
improving how seed dispersal is modelled and the dynamics are
visualised.*

I have become a bit obsessed with hemiparasitic plants recently. When I
found Cameron et al.'s paper on how Yellow Rattle can create interesting
spatial dynamics in grasslands¹, I got quite excited. I have been
learning python this year so I thought I may as well try to replicate
their model. I did this and then expanded on it a bit.

A spatial model is effectively a field full of virtual plants. This
field is divided up into a grid of cells, which have plants of the
different species growing in them. The plants produce seeds which fall
in the same cell and in cells nearby, which go on to compete with each
other and determine the number of plants in the cell in the next
generation (or "timestep").

## The study system {#the-study-system .wp-block-heading}

Yellow Rattle is a root hemiparasite, stealing nutrients and water from
its neighbours (for more on this see "[hemiparasitic plants in Irish
Grasslands](https://irishplants.org/blog/2022/12/27/hemiparasitic-plants-in-irish-grasslands/){data-type="link"
data-id="https://irishplants.org/blog/2022/12/27/hemiparasitic-plants-in-irish-grasslands/"}"
and my upcoming guide to Yellow Rattle). It is now widely used in
grassland restoration to suppress the growth of grasses and encourage
plant diversity, so understanding how it interacts with other species is
important. 
The other two species in the model are averaged, archetypal "grass" and
"forb" (herbaceous perennial). The species studied to get these averages
were Smaller Cat's Tail, Crested Dog's Tail, and Red Fescue for the
grasses; and Meadow Buttercup, Oxeye Daisy, and Ribwort Plantain for the
forbs. 
The interactions of the species were studied under both high and low
nitrate conditions, nitrate being a very important soil nutrient.

## The model {#the-model .wp-block-heading}

We start with a field, and we sow seeds of rattle, grass, and forb in
equal quantities. From this grow plants which compete with each other,
disperse seeds, and then die. This process continues and we watch what
happens...

<figure class="wp-block-image size-full is-resized">
<img
src="https://irishplants.org/blog/wp-content/uploads/2023/08/field.gif"
class="wp-image-526" style="width:400px" decoding="async" width="400" />
<figcaption>In this case for low soil N.</figcaption>
</figure>

In order to interpret this I'll have to explain how the model is
visualised. Pixels in a GIF have three colour channels (Red, Green, and
Blue), and these combine (like primary colours of paint) to create all
of the colours on the screen. Each of these colour channels represent
the population of a species as a fraction of the maximum possible
population (this is sort of a quirk of how I coded it but you can
imagine it as the maximum number of plants you could physically fit in
the space). If there are no plants of that species in the cell the value
is 0, and if there is 100% of the maximum it is 256. 
The colour channels then combine and we see changes in population as
fluctuations in the colour of each pixel.

## How I made the model {#how-i-made-the-model .wp-block-heading}

### Methods from the original paper {#methods-from-the-original-paper .wp-block-heading}

In the original paper, the species studied were grown together pairwise
in pots, and the final size of the plants were compared -- does one
outcompete the other? To what extent? The results of these were
summarised as numbers which can be plugged into difference equations
([wiki
link](https://en.wikipedia.org/wiki/Recurrence_relation){data-type="link"
data-id="https://en.wikipedia.org/wiki/Recurrence_relation"}), which
describe how the abundance of each species changes between one timestep
to the next (e.g. from one year to the next). 
These equations are run simultaneously in each cell of the model
(represented as one pixel). Each timestep, every plant produces seeds,
which are distributed among the cells around it. These seeds "compete"
(i.e. the equations are run to determine how many survive) and this
determines the number of plants of each species in each cell. Then the
process repeats. 
This produces some interesting dynamics. In particular the authors point
out the dynamics with high nitrate are different in the spatial model to
a non-spatial model (effectively one single cell). This is because the
populations of the different species can be different in different
places, whereas in the non-spatial model one wins out.

### Changes from the original paper {#changes-from-the-original-paper .wp-block-heading}

In the original model, seeds only fall in the original cell, the 4
neighbouring cells, and a small fraction somewhere random in the field.
To make a more realistic model of seed dispersal I implemented a
different dispersal kernel, where seeds fall in a bell-jar shaped
distribution around the parent plant (any cross-section is a normal
distribution):

<figure class="wp-block-image size-full">
<img
src="https://irishplants.org/blog/wp-content/uploads/2023/08/myplot.png"
class="wp-image-531" loading="lazy" decoding="async" width="137"
height="138" />
<figcaption>The yellow is where the greatest number of seeds fall, while
the blue is where only very few seeds fall.</figcaption>
</figure>

I also changed the model so that each cell contains only a finite number
of plants -- this makes extinction much more likely, and the periodic
steady state with high N is much less likely. When the model does not
reach this steady state there is no visible difference in dynamics from
the non-spatial model. The original authors discussed this when they
investigated a stochastic spatial version of their model:

> With high-nutrient parameters, stochasticity per se always reduced
> rather than extended the coexistence times. Under low-nutrient
> parameters, stochasticity made coexistence less likely as low
> population sizes tended to lead to extinction of the grass, especially
> if populations were started away from the equilibrium state. However,
> when equilibrium was reached, the stochastic version of the
> non-spatial model resulted in oscillations among the components (Fig.
> 1c). This cycling has the superficial appearance of being
> deterministic, but is absent from the deterministic version of the
> same model; it is actually the result of stochastically induced
> resonance, a phenomenon observed generally in systems described by a
> set of non-linear equations and occurs even at large population sizes
> (Alonso, McKane & Pascual 2007).
>
> Cameron et al., 2009

The last difference, which I made at the end, is that plants of the same
species compete with each other. This brings back the periodic cycling
for high N. This to me is a far more realistic scenario as plants always
"compete" with others of the same species, either through direct
competition or increased disease².

## Conclusion {#conclusion .wp-block-heading}

My main takeaway from this has actually been that these models are not
particularly useful for understanding the interactions of different
species in grasslands. While they are a nice way to illustrate
population dynamics, it is essentially impossible to make any meaningful
predictions as there are so many factors involved (species, soil type,
climate, seed dispersal...) that accurately parameterising the model is
incredibly difficult. 
On the other hand, the dynamics do provide a nice illustration of how
hemiparasites can affect plant communities. I used this to create an
illustration of the ideal case of Yellow Rattle introduction:

<figure class="wp-block-image size-full">
<img
src="https://irishplants.org/blog/wp-content/uploads/2023/08/timeline.png"
class="wp-image-525" loading="lazy" decoding="async"
srcset="https://irishplants.org/blog/wp-content/uploads/2023/08/timeline.png 3507w, https://irishplants.org/blog/wp-content/uploads/2023/08/timeline-300x180.png 300w, https://irishplants.org/blog/wp-content/uploads/2023/08/timeline-768x461.png 768w, https://irishplants.org/blog/wp-content/uploads/2023/08/timeline-1536x921.png 1536w, https://irishplants.org/blog/wp-content/uploads/2023/08/timeline-2048x1228.png 2048w"
sizes="(max-width: 3507px) 100vw, 3507px" width="3507" height="2103" />
</figure>

## References {#references .wp-block-heading}

1. D. D. Cameron, A. White, J. Antonovics, Parasite--grass--forb
 interactions and rock--paper-- scissor dynamics: predicting the
 effects of the parasitic plant Rhinanthus minor on host plant
 communities. *Journal of Ecology*. **97**, 1311--1319 (2009).
2. J. S. Petermann, A. J. F. Fergus, L. A. Turnbull, B. Schmid,
 Janzen-Connell Effects Are Widespread and Strong Enough to Maintain
 Diversity in Grasslands. *Ecology*. **89**, 2399--2406 (2008).