More about measuring biomass …

The previous post showed how the proportions of green algae and diatoms changed as the total quantity of algae in the River Ehen waxed and waned over the course of a year.   The BenthoTorch, however, also measures “blue-green algae” and so let’s look at how this group changes in order to complete the picture.

Before starting, though, we need to consider one of the major flaws of the BenthoTorch: its algorithms purport to evaluate the quantities of three major groups of algae yet, in my posts about the River Ehen I have also talked about a fourth group, the red algae, or Rhodophyta (most recently in “The only way is up …”).  Having pointed a BenthoTorch at numerous stones with thick growths of Audouinella,we can report that Rhodophyta seem to be bundled in with the blue-green alga signal, which is no great surprise given the similarity in their pigments.  It is, however, one of a number of examples of the need to interpret any BenthoTorch results with your brain fully engaged, and not just to treat outputs at face value. Similar questions need to be asked of the Xanthophyta and Chrysophyta, though the latter tend not to be common in UK streams.

Relationship between the proportion of “blue-green algae” (Cyanobacteria and Rhodophyta) and the total quantity of benthic algae (expressed as chlorophyll concentration) in the River Ehen (c.) and Croasdale Beck (d.).  The blue lines show quantile regression fits at p = 0.8, 0.5 and 0.2.  

In contrast to the green algae and diatoms, the Cyanobacteria/Rhodophyta signal shows a strong negative relationship as biomass increases though, again, there is enough scatter in this relationship to make it necessary to approach this graph with caution.  I suspect, for example, that the data points on the upper right side of the data cloud represents samples rich in Audouinella, which tends to occur in winter when biomass, generally, is much greater.   On the other hand, Croasdale Beck, in particular, has a lot of encrusting Chamaesiphon fuscus colonies which are pretty much perennial (see “a bigger splash …”) but whose relative importance in the BenthoTorch output will be greatest when the other two groups of algae are sparse.   I suspect that encrusting members of this genus are favoured by conditions that do not allow a high biomass of other algae to develop, as these will reduce the amount of light that the Chamaesiphonreceives.

Thicker biofilms in the River Ehen often have some narrow Phormidium-type filaments as well as small bundles of nitrogen-fixing Calothrix, but the overall proportion is generally low relative to the mass of diatoms and green algae that predominate.    But that is not really telling us the whole story.  I finished my previous post with a graph showing how the variation in biomass increases as the biomass increases.  The heterogeneity of stream algal communities, however, cannot be captured fully at the spatial scale at which the BenthoTorch works: there is a patchiness that is apparent to the naked eye: one of our sites has distinct mats of Phormidium autumnale towards one margin, and dense Lemaneagrowths in the fastest-flowing sections, largely attached to unmovable boulders, which makes biomass measurement very difficult. I’ve also written about distinct growths of Tolypothrix and its epiphytes (see “River Ehen … again”), another alga which forms discrete colonies at a few locations. I try to collect a random sample of stones from a site but there are constraints, including accessibility, especially when the river rises above base flow.   In the River Ehen we also have to take care not to disturb any mussels whilst removing stones.

Whilst our sampling cannot really be described as “random” I do think that there is sufficient consistency in the patterns we see for the results to be meaningful. We could spend a lot more time finessing the sampling design yet for little extra scientific gain.   I prefer to think of these measurements as one part of a complex jigsaw that is slowly revealing the interactions between the constituents of the dynamic ecosystem of the River Ehen.   The important thing is to not place too much faith in any single strand of evidence, and to have enough awareness of the broader biology of the stream to read beyond the face value indications.

The complexities of measuring mass…

Once upon a time, measuring the quantity of algae growing on the beds of streams and rivers was a painstaking, slow process that invariably revealed large amounts of spatial and temporal variation that, very often, obscured the ecological signals you were looking for. That has changed in the last decade thanks to the availability of field fluorimeters such as the BenthoTorch.  This makes it much quicker and easier to measure chlorophyll concentrations, the usual proxy for algal quantity.  Thanks to devices such as this it is now much easier to discover that your ecological signal is masked by spatial and temporal variation.

We’ve generated a lot of data about the fluxes of algae in the River Ehen using a BenthoTorch over the past five years and are in a position where we can start to make some generalisations about how the quantity of algae vary over the course of a year.  In broad terms, the results I showed in “The River Ehen in January” back in 2014 have not varied greatly over subsequent years, with peak biomass in mid-winter and low biomass in the summer (due, we presume, to intense grazing by invertebrates).  Curiously, we see a much less distinctive seasonal pattern in the nearby Croasdale Beck, but that is a story for another day….

The BenthoTorch uses an algorithm to partition the fluorescence signal between three major algal groups and, though this is not without issues (see below), I thought it might be interesting to see how these groups varied with biomass trends, and consider how this links to ecological theory.  The first group I’m considering are the green algae which, in this river, are mainly filamentous forms.   The general pattern, seen in the graph below, is for a gradual increase in the proportion of green algae, which fits with the current understanding of thicker biofilms having greater structural complexity with filamentous algae out-competing attached single celled algae to create a “canopy” of algae that are more effective at capturing light and other resources.  The relationship is, however, strongly wedge-shaped so, whilst many of the thickest biofilms have a lot of green algae, there are also thick biofilms where green algae are scarce or even non-existent.  Croasdale Beck shows a similar, but less pronounced, trend.

Relationship between the proportion of green algae and the total quantity of benthic algae (expressed as chlorophyll concentration) in the River Ehen (a.) and Croasdale Beck (b.).   The blue lines show quantile regression fits at p = 0.8, 0.5 and 0.2.   The image at the top of the post shows Ben Surridge using a BenthoTorch to measure algal biomass beside Croasdale Beck in Cumbria.

The second graph shows that this pattern of a gradual increase in proportion is also the case for diatoms and, once again, there is a broad wedge of points with an upward trend.  But, once again, there are also samples where biomass is high but diatoms are present in very low numbers or are even absent.   What is going on?

The problem is clear I think, if one looks at the final image in “The only way is up …” where the very patchy nature of algal communities in the River Ehen (and, indeed, many other rivers).   There are plenty of algae on this boulder, but not organised in a homogeneous manner: some zones on the boulder are almost pure diatom whilst others are almost pure green algae (and there are also zones that are almost pure Lemanea– I’ll come to that in a future post).   We try to sample the stones as randomly as possible so you can see the potential for getting very different numbers depending on where, on a stone, we point the BenthoTorch’s sensor.

Relationship between the proportion of diatoms and the total quantity of benthic algae (expressed as chlorophyll concentration) in the River Ehen (c.) and Croasdale Beck (d.).   The blue lines show quantile regression fits at p = 0.8, 0.5 and 0.2.  

With experience, you can make an educated guess about the types of algae present in a biofilm.  I’ve tried to capture this with my watercolours, using washes of raw sienna for the diatoms and a grass-green for the green algae, which roughly matches the colour of their respective growths in the photo in my earlier post.   The two groups of algae a are relatively distinct on that particular boulder.   The top row roughly matches the upper “edge” of the graph showing variation in diatoms, whilst the bottom row emulates the upper “edge” of the graph showing variation in green algae.  These are the two extreme situations; however, we also often see darker brown growths in the field, which can be recreated by mixing the raw sienna and grass-green together.  When I peer through a microscope I often see green algae smothered in diatoms: genera such as Oedogoniumare particularly prone as they have less mucilage than some of the others we find in the Ehen. Their filaments often host clusters of Fragilariacells as well as Achnanthidium minutissimum, whilst stalked Gomphonemaand chains of Tabellaria flocculosaoften grow through the tangle of green filaments.   The dark brown colour is deepened yet further by the colour of the underlying rock, so my effort on white watercolour paper is a little misleading.

A colour chart showing how different proportions of green algae and diatoms influence the colour of biofilms.

The final graph shows how, as the average biomass increases in the River Ehen, so the variability in biomass also increases.   The River Ehen is one of the cleanest rivers I know but I suspect that this pattern in benthic algal quantity could be reproduced in just about any river in the country. What I would not expect to see in any but the purest and most natural ecosystems is quite so much variation in the types of algae present.   Once there is a little enrichment, so I would expect the algae to become more of a monoculture of a dominant filamentous alga plus associated epiphytes.  Like much that happens in the microscopic world of rivers, it is easier to describe than it is to measure.

That, however, is only part of the story but I’ll come back to explain the patterns in the other main groups of algae in the Ehen and Croasdale Beck in my next post.

The relationship between mean chlorophyll density and the standard deviation (based on measurements from five separate stones) for samples from the River Ehen and Croasdale Beck. 

 

Secular icons?

I’ve got two pictures on display as part of an exhibition at Durham University Botanic Gardens, both showing abstract or semi-abstract views of algae.  One is my sextych of the alga Apatococcus(see “Little round green things …”) and the other is a triptych based on Haematococcus, an alga which I wrote about in “An encounter with a green alga that is red” back in 2013.   Both pictures were painted for my final degree show back in 2008 and both addressed questions about the boundaries between abstract and representational art.

The point that I was trying to make with these images is that the boundary between abstract and representational art depends partly on what the viewer knows about the subject matter and, in the case of algae, this is usually not very much.  In cases such as these, the legend becomes very important as a means of bridging the gap between abstraction and reality by providing just enough information to help viewers make sense of the content (see “How to win the Hilda Canter-Lund Prize (2)”).   In the case of microscopic images, this should always include some indication of scale, written in terms that non-biologists can easily understand (I would always write “1/100^thof a millimetre”, rather than “10 micrometres”, for example).

Haematococcus. Triptych.   2008 50 x 130 cm Acrylic, resin and PVA on canvas.

This issue of viewers being able to “unlock” the meaning of images extends beyond the abstract/representational boundary that I encounter when displaying images of the microscopic world.  Exactly the same challenges occur when, for example, secular western Europeans look at eastern Orthodox icons, a subject that occasionally creeps into this blog (see, for example, “Unorthodox icons”.   My own curiosity about this art form led me to spend a week studying icon painting at the Quaker College in Woodbrooke, in the suburbs of Birmingham.  About ten minutes away from Woodbrooke there is the Serbian Orthodox church of St Lazar (built after the second world war by Yugoslav refugees with financial support from the Cadbury family, the Quaker philanthropists who also established Woodbrooke).

I talked a little about the practice of icon painting in “The art of icons …”.  Today, I am more interested in the symbolism.   A secular westerner can look at many of the icons I’ve depicted and broadly catagorise the contents: most would recognise that Fr Nenad, the priest of the Selly Oak Orthodox church, is holding an icon that depicts the crucifixion, for example, or that the icon just to the left of the centre of the doors in the iconostasis in the lower image depicts the Virgin and Child.  However, the symbolism goes much deeper.   I have a spotter’s guide to icons (sad, I know …) and it lists twenty eight different variants on the basic depiction virgin and child, differing in the physical relationship of Mary and Jesus, their facial expressions and the setting.  Each of these have a slightly different meaning for the Orthodox faithful.   The westerner sees “virgin” and “child”, the eastern Orthodox devotee sees so much more.

The Serbian Orthodox Church of the Holy Prince Lazar in Selly Oak, Birmingham with, right, Father Nenad displaying an icon of the crucifixion. 

What’s all this got to do with painting algae, you may ask.   Scientific illustratation and icon painting are two branches of applied art, where the subject matter serves a higher purpose.  Both, in their own way, try to help viewers understand their place in the world.  If you are not religious, you may not be comfortable with this comparison but, for most of Europe, east and west, until the Enlightenment, this would have been the case.   In both cases, however, the image cannot be viewed in isolation, the viewer needs the appropriate “keys” to unlock meaning.   Even then, the viewer is not a passive observer.   The icon requires a response from the viewer, it is the focus for contemplation and meditation and, I suggest, scientific images, when displayed as “art” should play a similar role, inviting viewers to reflect upon the mysteries of the natural world and demanding a response.

The iconostasis at the Serbian Orthodox Church in Selly Oak.

The only way is up …

How does an alga move upstream?   I’m curious because, I am now seeing populations of Lemanea fluviatilisabout four kilometres further upstream in the River Ehen than when I first started my regular visits in 2013.   I can explain the presence of the organism partly through changes in the hydrology of the river: a small tributary, Ben Gill, that had been diverted into the lake in Victorian times was reconnected to the river in 2014 and this introduced periodic pulses of intense energy to the river that had immediate effects on the substrate composition.  Lemanea fluviatilisis a species that thrives in the fastest-flowing sections of streams so I am quite prepared to believe that even a small shift in the hydrology of this very regulated river might make the habitat more conducive.

But that does not explain how it got there in the first place.   If the alga was occurring a few kilometres further downstream we would not have any such problems: the upstream populations would provide innocula and, if the habitat conditions changed at the downstream location, then some of those propagules might be able to establish at the downstream locations.   But what about movement in the other direction?

There has been relatively little published on this topic in recent years.  I have a review by Jørgen Kristiansen from 1996 that considers the dispersal of algae but most of the references that he cites are quite a lot older than this and I have not seen much published subsequently.   He lists our options: dispersal by water, by organisms, by air currents and by human activity.   Let’s consider each in the context of Lemaneain the River Ehen.   Lemanea, like most red algae, has a complicated life cycle with the potential for dispersal in both the haploid and diploid phases, but that is probably more detail than we need right now.  We’ll just outline the options in broad terms:

Water:the linear flow of the river means that it is almost impossible for the downstream population to provide inocula for the new upstream locations.  It may be possible for populations from further upstream in the catchment to seed the new locations.  I have not seen Lemaneain any of the streams that flow into Ennerdale Water (from which the Ehen emerges) but my knowledge of the catchment is not exhaustive.   Likelihood: very low to low.

Young shoots of Lemanea fluviatillis(bottom right) growing on a submerged boulder in the River Ehen at a location where I have not previously seen it.   These are growing alongside thick growths of diatoms (yellow-brown in colour) and patches of green filamentous algae.

Organisms:much of the older literature is concerned with the possibility of living algae or their propagules being transported in mud attached to bird’s feet or feathers and this cannot be ruled out.   There is also a recent study showing how mink may act as a vector for Didymosphenia geminata in Chile.  The Ehen also has aquatic mammals (such as otters) that could be acting as vectors for Lemanea, as well as migratory fish such as salmon and trout that could move propagules upstream.   There is also some evidence that some algae can survive passage through mammalian and invertebrate guts, and this, too, may provide a means for Lemaneato spread upstream.    Likelihood: low to medium.

Air currents / wind:quite a lot has been written about airborne dispersal of algae, with even Darwin making a contribution (see reference in Kristiansen).  The key hazard in airborne dispersal is desiccation so, in the case of Lemanea, the most likely lifecycle stages that could be dispersed in this way would be the diploid carpospores or haploid monospores. This, however, would assume that there were times during the year when the relevant life-cycle stages were exposed and, as Lemaneais a species that I usually find in the Ehen only fully-submerged, this is not very feasible.  Likelihood: low.

Human activity:there is evidence that Didymosphenia geminatacan be transported between sites attached to waders and new records often correspond with patterns of recreational use (references in Bergey & Spaulding – see below).   When we work in the Ehen we prefer to move downstream in order to minimise the risk of moving organisms on our kit, and we also clean our kit before we start.   However, a lot of people work in this part of the Ehen and it only takes one dirty wader to introduce a propagule.   Likelihood: low to medium.

We’ll almost certainly never know for sure why Lemanea fluviatilisis now thriving four kilometres further upstream than it was five years ago.  It is, however, worth bearing in mind that, given enough time, even a low probability may yield a positive result.   So none of the four hypotheses can be ruled out for sure.   Three of the possibilities are entirely natural, with one – movement by the stream itself – being constrained by the direction of flow.  Biological vectors look like a very plausible means of moving algal propagules around catchments but, for this to work, we need wildlife-friendly corridors around the river to support the animals and birds.  The upper Ehen has these, but many other rivers do not.

Actually, having a number of options all with a relatively low likelihood adds to the sense of mystery that every ecologist should have when they approach the natural world.  When cause and effect are too predictable, we tend to focus on engineering the right “solution”.  The truth, in our muddled and unpredictable world, is often that nudging several factors in the right direction will give us a more resilient outcome, even though we may have to wait longer for it to happen.

Reference

Bergey, E.A. & Spaulding, S.A. (2015). Didymosphenia: it’s more complicated.  BioScience65: 225.

Kristiansen, J. (1996).  Dispersal of freshwater algae – a review.  Hydrobiologia336: 151-157.

Leone, P.B., Cerda, J., Sala, S. & Reid, B. (2014).  Mink (Neovision vision) as a natural vector in the dispersal of the diatom Didymosphenia geminata.  Diatom Research29: 259-266.

Raven, J.A. (2009).  The roles of the Chantransia phase of Lemanea (Lemaneaceae, Batrachospermales, Rhodophyta) and of the ‘Mushroom’ phase of Himanthalia (Himanthaliaceae, Fucales, Phaeophyta).  Botanical Journal of Scotland46: 477-485.

Croasdale Beck in February

My latest trip to the west Cumbria coincided with the period of freakily warm weather that marked the end of February (in marked contrast to a year previously when we were in the midst of the “Beast from the East”).   It felt like spring had come early although the skeletal outlines of leafless trees were incongruous against the backdrop of blue skies and, despite feeling the warmth of the sun on our faces as we worked, the water still had a wintery chill when the time came to plunge in my arm.

There were thick growths of algae on the bed of Croasdale Beck: a quick check with my microscope later showed this to be mostly Odontidium mesodonand Gomphonema parvulumand this piqued my curiosity to see how different species responded to the fluctuations in biomass that we observe in the streams in this region. I’ve talked about this before (see “A tale of two diatoms …”), suggesting that Platessa oblongellatended to dominate when biofilms were thin whilst Odontidium mesodon preferred thicker biofilms.  That was almost two years ago and I now have more data with which to test that hypothesis, and also to see if any other common taxa had an equally strong preference for particular states.

A cobble from the bed of Croasdale Beck in February 2019 showing a brown biofilm (approx. 1.7 micrograms per square centimetre) dominated by Gomphonema parvulumand Odontidium mesodon.   The photograph at the top of the post shows Ennerdale Water photographed on the same day.

I should also be clear that, in Croasdale Beck especially, diatoms are the main algal component of the biofilm, so they are not so much responding to a particular state of the biofilm as actively contributing biomass to create that state.  The other photosynthetic organism that is obvious to the naked eye in this part of Croasdale Beck is the cyanobacterium Chamaesiphon fuscus (see “A bigger splash …”) but this forms crusts on stone surfaces rather than contributing to the superstructure of the biofilm itself. We do find other filamentous algae, but intermittently and in smaller quantities.

We’ll look at Platessa oblongellafirst, bearing in mind that this was shown to be a mixture of two species about halfway through our study (see “Small details in the big picture …”).   The graph below, therefore, does not differentiate between these two species although, from my own observations, I have no reason to believe that they behave differently.   What I have done in these graphs is to divide the biomass measurements and the percent representation of these taxa in each sample into three categories: low, middle and high.   In each case, “low” represents the bottom 25 per cent of measurements, “high” represents the top 25 per cent of measurements and “middle” represents all the rest. The left-hand graph shows biomass (as chlorophyll a concentration) as a function of the relative abundance of the diatom whilst the right-hand graph shows the opposite: the relative abundance of the diatom as a function of the biomass.  These graphs bear out what I suggested in my earlier post: that Platessa oblongella(and P. saxonica) are species whose highest relative abundances occur when the biofilm is thin.  So far, so good.

Relationship between relative abundance of Platessa oblongella (including P. saxonica) and biomass in Croasdale Beck, Cumbria.  a. shows biomass (as chlorophyll a) as a function of the relative abundance of the two species (Kruskal-Wallis test, p = 0.047) whilst b. shows the relative abundance as a function of biomass (p = 0.057).

My second prediction in my earlier post was that Odontidium mesodonpreferred moderate or thick biofilms; however, whilst there is a clear trend in the data, differences between low, middle and high values of neither biomass nor relative abundance are significant.   The explanation may lay in the strong seasonality that O. mesodondisplays, thriving in spring but less common at other times of year (see “More about Platessa oblongella and Odontidium mesodon”).  However, there are no strong seasonal patterns in biomass in Croasdale Beck, and this disjunction introduces enough noise into the relationship to render it not significant.

Relationship between relative abundance of Odontidium mesodon and biomass in Croasdale Beck, Cumbria.  a. shows biomass (as chlorophyll a) as a function of the relative abundance of O. mesodon (Kruskal-Wallis test, p = 0.568) whilst b. shows the relative abundance as a function of biomass (p = 0.060).

I then tried looking at the relationship between relative abundance and biomass for a few other common taxa but with mixed results.   None of Achnanthidium minutissimum, Gomphonema parvulum complex or Fragilaria pectinalis showed any clear relationship; however, when I looked at Fragilaria gracilis, a different pattern emerged, with a significant relationship between the quantity of biomass and the proportion of this species in the sample.  That, too, is not a great surprise as I often see clusters of Fragilaria gracilis cells growing epiphytically on filamentous algae within the biofilm.  Whilst Platessa oblongella, which sits flat on the stone surface, seems to be a species that thrives when the biofilm is thin, so Fragilaira gracilisis favoured by a more complex three-dimensional structure, where it can piggy-back on other algae to exploit the light.   I suspect, however, that in a stream such as Croasdale Beck, where the substratum is very mobile, Fragilaira gracilis will also be one of the first casualties of a scouring spate which will, in turn, open up the canopy allowing Platessa oblongella back.   Even though my results for Odontidium mesodonare not significant, I still think it plays a part in this sequence, occupying the intermediate condition when some biomass has accumulated.  It looks to me as if it also likes cooler conditions which then complicates interpretation of my results.

Indeed, I am being rather selective in the results that I have included here.  Three of the six species I investigated showed no response and one of the three that I did include showed a trend rather than a statistically-convincing effect.  I suspect that the situation will rarely be as simple as I have shown for Platessa oblongella and Fragilaira gracilis.  Nonetheless, there is enough here to make me want to scratch a little more and try to understand this topic better.

Relationship between relative abundance of Fragilaria gracilis and biomass in Croasdale Beck, Cumbria.  a. shows biomass (as chlorophyll a) as a function of the relative abundance of F.gracilis (Kruskal-Wallis test, p = 0.010) whilst b. shows the relative abundance as a function of biomass (p = 0.036).

Croasdale Beck, photographed in February 2019.