Where should science funding go?

 by Alastair Potts

I've just finished listening to a podcast by Malcom Gladwell entitled "My Little Hundred Million". In it he highlights that most major donations from super-rich philanthropists go to Universities that already have eye-watering endowments than run into the billions of dollars (Stanford, Harvard etc.). Why? Why give well-endowed prestigious universities more money. Well, because they're going to make intellectual superstars that are going to change the world. Right? Give the richest universities more money so they can produce even better intellectuals. There are big problems with this type of thinking ​— and we'll discuss it in a moment ​— and I believe its being applied to how we fund science in South Africa. 

Gladwell highlights our focus on superstars ​— if we can create superstars in whichever field, sport, art, science etc., then these people are going to have a disproportionate effect on society. Gladwell calls this the "strong-link" approach, where success is determined by the best "player on the field". There is another way to think about this, called the "weak-link" approach ​— the overall success of a team is determined by the weakest players, not the strongest. Some sports, such as basketball, can be dominated by a single player ​— it doesn't matter how many weak players you have on the court as long as you have the superstar who can dribble and shoot up and down the court at will. This demonstrates strong-link thinking. In other sports, the weakest players can affect the entire team irrespective of the superstars. Gladwell cites a book, The Numbers Game, by David Sally and Chris Anderson where they show that soccer leagues should spend less money on getting superstars and rather spend it in improving the quality of their weaker players: "In soccer, what matters is how good your worst player is". It's a weak-link approach, where your chain is only strong as your weakest link. Spend money on strengthening the weakest link.

Gladwell gives other examples of weak-link thinking outside of sport ​— such as, how do you deal with air-traffic congestion across airports? You don't go and spend all your money on the already fanciest, biggest, most state-of-the-art airport (the Stanford of airports). Rather, you spend it on those smaller, less prestigious airports that are not coping. By spending on less-prestigious airports has a far greater impact to the overall air transport system. If philanthropists wanted to have greater impacts on universities, then they should give their money to poorer universities ​— they teach more students, and more money will lead substantially greater improvements, relatively, than giving to a university that already has state-of-the-art everything. 

Two different ways to think about how to spend money and get the best return-on-investment ​— can you guess which type of thinking is followed by science funding in South Africa? We can look at the South African Research Chair Initiative as an example ​— give a few researchers in the community a lot of money to conduct their very important research. Turn them into research superstars. That's strong-link thinking. We create a few super-researchers and they will have an disproportionate effect on society. It's not only the SARCHI chairs. I've sat on review panels where only the "top" projects will be funded. Which projects made it to the "top" (i.e. fulfilled the most criteria)? By-and-large those projects that already had funding from other sources. And, the top projects also, invariably, asked for the most funding. More funding, lots more funding, was given to projects that already had funding! What about those smaller, focused, projects that didn't tick as many boxes? Well, they don't make it high enough up the list to get funded. I can remember one project in particular ​— it would have been run by one researcher on a tiny budget and involved laborious routine data collection that would have provided an absolutely fundamental database for all other researchers in the field. It barely ticked any boxes (no international collaboration, not a big sexy problem etc. etc.) ​— it didn't get funded. Strong-link thinking!   

And by doing this, we leave many researchers and projects out of the funding landscape. I would argue that science is a weak-link game ​— we may think that science is driven by the few superstars, but actually they are supported by a host, swarm, flock, menagerie(?) of other researchers and intellectuals. We also have to admit that scientific discoveries often involve luck, sometimes a lot of luck, and more often than we'd care to really admit. Providing mega-funding to a few scientists reduces the number of scientists that can get lucky. 

So where should we be putting our research money? I think we should be spreading it thin, giving it to more scientists (who may have lower impact factors or whichever other bizarre metric you wish to measure scientists by), and try to get as many researchers in the game with their passion, their scientific views, and their luck, to make new notable discoveries. Get these researchers off the bench and into the game, even if they don't have best, fanciest, most-state-of-the-art equipment, their ingenuity will find a way. (A recent example from my own experiences ​— I want to do some work using camera traps to study pollinators. If  I had a huge project grant, I would have bought camera traps that are four to five times the price of a standard trap. I don't have much money for this project. I figured out that buying a pair of reading glasses at my local chemist, popping out the lenses and sticking them onto a normal camera trap works just fine. I can even see ants! I found a way to do something for less, but I needed some research money for the project). 

We need to stop thinking of science funding in terms of big projects or superstars ​— i.e. that science is a strong-link game. We need to find a way to ensure that as many scientists are active as possible. Let us not get misled by the glamour of strong-link thinking.   

"And that’s the problem. Superstars are glamorous, Nobel Prize winners are glamorous. Regional universities in rural South Jersey and solid, capable midfielders [or scientists] are not." Malcom Gladwell

Update: A colleague from another institution read this and said he shared this sentiment: he's been arguing that SARCHI positions probably cost the National Research Foundation R5 million per chair ​— and that there is no way that a single person could out-publish 25 researchers with R200,000 per year. If single researchers get lots of money, there may also be the tendency to play "Godfather" science, as my colleague puts it, as they get their names on papers by sharing their funding with lesser mortals. When these arguments get raised with the upper levels of NRF management, then the retort is that we need superstars for exposure of the rest of National System of Innovation. That's strong-link thinking right there. It's the same reasoning as to why football club owners won't give up their soccer superstars and spend the money on improving the quality of the weaker links in the team ​— they'll lose out on the bragging rights (this point is made by Malcom Gladwell in his podcast). 

Thirst, starve or susceptibility? Elucidating to physiological driver of canopy tree mortality in transformed arid Thicket.

by Daniel Buttner 

As we know Thicket has experienced considerable transformation shifting from a highly conserved resource sink ecosystem to an ephemeral based one (see Lechmere-Oertel et al., 2005). This change in resource dynamics has consistently been demonstrated via fence-line contrasts where extensive browsing, chiefly by goats, has degraded the landscape, removing key species contributing to both ecosystem productivity and functionality, such as the dominant succulent Portulacaria afra (spekboom) and the many "palatable" species.

While extensive work has been conducted on the overall pattern or rather impact of browsing on thicket by quantifying the loss of species as well as looking into the soil geochemical and physical properties, there is little to no physiological research, particularly drought work. This is quite paradoxical considering the general notation and subsequent recognition of thicket species as being drought tolerant, yet citing no empirical data to back up said suppositions. Irrespective of this quantitative data absence several anecdotal observations might through a "spanner in the works" of drought-tolerance assumption in thicket. Yes, the distributional range of thicket, specifically arid subtypes, appear to support this drought tolerance notion. However, much as the subtropical trees in these systems persisting beyond their respective physiological optimum, so does the question as to why in transformed arid thicket do the trees, which have escaped considerable herbivory damage  by goats are slowly beginning to vanish from these transformed landscapes. Such a disappearance would seem to contradict the notion of drought tolerance capabilities of these ancient, resource conservative trees (most commonly Pappea capensis and Euclea undulata). The expatriation has considerable repercussions for the stability and subsequent conservation efforts of subtropical thicket, particularly arid types.

I believe, or more accurately speculate, that many of the tree species are relying on a quite elusive symbiosis. What we in the Potts Research Group have come to fondly call the "Spekboom Sponge Hypothesis" (see The Spekboom Sponge Hypothesis or SSH) which posits that trees and other shrubs exploit P. afra as a physiological refugia, abstracting necessary water requirements during protracted drought (both frost drought and arid drought conditions) periods in arid Thicket. Under the SSH we might expect that thicket trees might experience greater physiological strain than those in transformed sites (spekboom the first causality to goat herbivory).

The removal of this dominant succulent has a substantial impact on the fundamental ecosystem functions, ranging from soil stabilization to gross carbon sequestration. While the role of P. afra as an ecosystem engineer is undisbuted but its relevance as a nurse plant, as suggested in Adie & Yeaton (2013), and subsequent effects as a physiological refuge from water constrained conditions remains elusive and is need of  discussion. Many studies have attempted to capture nurse plant effects or rather ascertain whether a foundational species (synonymous with ecosystem engineer) can be characterized as a nurse plant. Conventional approaches employ spatial association analyses (proximal distance between individuals to calculate a "Zone of Influence"), the second more commonly applicable approach is removal of the suspected nurse plant and document changes resident species functional traits, often phenological and health indices, with the rare appearance of a physiological metrics (see Brooker et al., 2008) to infer a reduction in stress condition of benefiters near a suspected nurse plant.

While such a removal experiment would not only be impractical but also extremely expensive to undertaken. However, in thicket we are fortunate (I use the term loose here) that goat browsing has provided us with just such an opportunity by preferentially consuming the more palatable species first, i.e. succulents, lianas and eventually spinescent shrubs, leaving only those capable of physical escape, such as the trees who's foliage exists beyond the vertical reach of goats are but the few species to have survived in these landscapes. Although, totally by coincidence have these trees escaped the first round of biophysical disturbance, yet little do they realize the second round is on its way. By removing the undergrowth, goats have unintentionally removed the majority of the Thicket canopy heightening the evaporative demand on near-surface water-soil horizons, actualizing a localized aridity gradation of  the soil matrix placing considerable physiological strain on remnant trees. The impacts of such an aridity gradient is a completely loss of the foundational microclimate needed for seedling establishment and essential for future tree recruitment success in arid thicket. The image below illustrates this quite well, with the sheer abundance of Pappea capensis skeletons outnumbers the living.

Aerial imagery taken in transformed landscape illustrating the numerous Pappea capensis skeletons (red circles) throughout indicative of future expatriation of these trees within this landscape. Note the blue circle, this individual has lost half its canopy cover with the remaining stems bleached white stems indicative of dead tissue. I suspect this individual to have surpassed a critical threshold of mortality likelihood and the foliage persisting is simply the remnant metabolic legacy of the plant, and has cascading hydraulic failures due to the inability to maintain a positive net carbon balance.

While it is quite apparent as to the outcome of goat browsing, i.e. loss of thicket wholeheartedly, the reasons for why the trees and a few spinescent shrubs (pers. obs. Carissa haematocarpa) are eventually expatriated from the landscape — the underlying physiological mechanism rationalizing their lose is yet to ascertain. From the actualization of aridity gradation transitioning between intact to transformed, many would posit drought or rather more specifically drought susceptibility, contrary to the long-held belief of drought tolerance of thicket. While I can only speculate at this moment to the hydraulic vulnerability of these trees there are shared similarities in anecdotal observations made with other classic drought based model systems, such as the Pinyon-Juniper woodlands in the USA (Mueller, et al., 2005). This classic example contrasts to dominant species as polar opposites in their respective drought response strategies and have ever since been cited as the benchmark for analogous reviews and comparative analyses for interpreting drought responses by other arid forest ecosystems or those experiencing seasonal water deficits.

Here I illustrate the strong comparative analogs using a single dominant tree species found throughout the Subtropical Thicket biome, Pappea capensis, placing it in context of the proposed hydraulic framework by McDowell et al., (2008) attributing tree mortality to two main hypotheses. The first being the hydraulic failure hypothesis, which attributes the success of a species under drought to have a well structured and supportive conductance system, enabling stomatal activity to go unimpeded during drought, thus anisohydric species where the whole-plant water potential follows closely that of soil (i.e. environment). Anisohydric species fluctuate minimally in their respective stomatal conductance ensuring carbon assimilation is not impact by drought. However, for the plant to ensure water continuity its imperative that a well supported hydraulic architecture capable of enduring extremely water potentials and potential for cavitation and subsequent emboli during protracted dry spells — anisohydric species as the risk takers and most vulnerable to hydraulic failure. Whereas isohydric species coordinate hydraulic traits to avoid the physiological strain expressed upon them during water deficiencies, maintaining whole-plant water potential at more positive pressures than those experienced those of the soil, this ensures that the plant is unlikely to experience cavitation or possible hydraulic dysfunction during drought conditions. However, by limiting or rather stringently controlling stomatal conductance these isohydric species have restricted photosynthetic activity, thus carbon assimilation. Isohydric species play it safe, as most often is the case that they are hydraulically vulnerable and thus take every effort to safeguard xylem conduits from cavitation and possible collapse. However, while attempting to avoid the hydraulic implications of drought they ran the risk of depriving themselves of necessary sugars, thus become carbon starved.

Conceptual hydraulic framework proposed by McDowell et al., (2008), illustrating the mortality trajectory of trees experiencing drought. Biotic agent susceptibility is either amplified or amplifying physiological stress.

An update of the hydraulic framework revised in McDowell (2011), which considers the dual strain placed on non-structural carbohydrates to mitigate hydraulic impairment via osmotic adjustment (production of stable solutes, e.g. proline) and having to ensure net carbon balance maintaining positive carbon turnover. This demand on carbon resources is evident by increased pest and pathogenic susceptibility. 

The carbon starvation and hydraulic failure hypotheses have been proposed the distinction isn't clear cut and is rather blurred. Both are presented as the physiological cause behind drought-induced tree mortality, differentiate on the basis of time to death. Where hydraulic dysfunction is anticipated over a series of months to years while carbon starvation is expected to occur over decades as the tree is carbon turnover is diminished progressively during drought cycles or decadal episodic events. A recent revision of the carbon starvation has additionally been proposed that illuminates hydraulic failure as more significant of the two as effecting mortality risk. McDowell (2011) revised the expectation under carbon starvation noting that during drought trees compensate for hydraulic impairment accumulating non-reactive solutes in their tissues increasing the osmotic potential and enable stomatal conductance to remain high even during water stress states prioritizing photosynthetic metabolism even during stress states. Now whether this is ineffectual photochemical quenching or interlink between nutrient assimilation, specifically nitrogen necessary for hormonal production and other signaling proteins for undertaking simultaneous responses during stressful conditions, remains speculative. However, what many researchers have reached a consensus on is the substantial effects of hydraulic impairment to the survival of trees in any environment experiencing drought or water deficit conditions.

Image of a physiologically taxed Pappea capensis situated in the transformed thicket landscape. A notable reduction in canopy foliage, not imposed by herbivory predominantly but rather as an evergreen (possible semi-deciduous) its leaves are hydraulically conservative (I speculate here no literature or data) thus are susceptible to emboli and subsequent photosynthetic impairment. (Image courtesy of Robbert Duker, May 2020)

Illustration of pest susceptibility of Pappea capensis individual in transformed site, note the numerous insects bore holes. A consequence of non-structural carbohydrate depletion reallocated for production of solutes for osmotic adjustments prioritising hydraulic conductivity over structural vulnerability and secondary metabolites preventing onslaught by pests. (Image courtesy of Robbert Duker, May 2020)

An intriguing outcome of either sequential losses in hydraulic conduits or depression of non-structural carbon turnover via stringent stomatal conductance, is the appearance of the comorbidity of increased  pest and/or pathogen susceptibility. Hence, biotic agents concurrently impact tree mortality either amplifying the physiological stress or enacted as a compounding factor. This impact strongly agrees the notion of allocation optimalty theory were carbon investment is prioritised to tissues or utilised in strategies that pose the greatest risk to survival of the plant. 

I have made several observations of pest susceptibility of the Thicket tree dominant Pappea capensis  that strongly agree with what we might expect under the hydraulic framework proposed by McDowell et al., (2008). There is a great demand for drought research both from a physiological and hydrological standpoint in Thicket as the existing climate change management outline for this biome operates under the historical notion of drought tolerant trees, yet for all we know these trees are persisting beyond their respective physiological optimum surviving within their "pseudo-realized niche" via interspecific facilitation with P. afra. Hence, the loss of this dominant ecosystem engineer has major ramifications for the sustainability of Thicket tree diversity in arid extensions of the Albany Subtropical Thicket biome. A great time to be a budding drought physiologists and Thicket ecologist, so many questions to ask and better still more experiments to run.

References:

Brooker, R.W., Maestre, F.T., Callaway, R.M., Lortie, C.L., Cavieres, L.A., Kunstler, G., Liancourt, P., Tielbörger, K., Travis, J.M., Anthelme, F. and Armas, C., 2008. Facilitation in plant communities: the past, the present, and the future. Journal of ecology, 96(1), pp.18-34. 

Lechmere‐Oertel, R.G., Cowling, R.M. and Kerley, G.I., 2005. Landscape dysfunction and reduced spatial heterogeneity in soil resources and fertility in semi‐arid succulent thicket, South Africa. Austral Ecology, 30(6), pp.615-624. 

McDowell, N., Pockman, W.T., Allen, C.D., Breshears, D.D., Cobb, N., Kolb, T., Plaut, J., Sperry, J., West, A., Williams, D.G. and Yepez, E.A., 2008. Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought?. New phytologist, 178(4), pp.719-739.

McDowell, N.G., 2011. Mechanisms linking drought, hydraulics, carbon metabolism, and vegetation mortality. Plant physiology, 155(3), pp.1051-1059. 

Mueller, R.C., Scudder, C.M., Porter, M.E., Talbot Trotter III, R., Gehring, C.A. and Whitham, T.G., 2005. Differential tree mortality in response to severe drought: evidence for long‐term vegetation shifts. Journal of Ecology, 93(6), pp.1085-1093.

Frost synonymous with fire? The hydraulic death hypothesis in Thicket

by Daniel Buttner

Frost has been defined as a fundamental factor in dictating the biome boundary between Thicket and Nama-Karoo shrubland (see Image 1). The impacts of frost are numerous and have been discussed in a previous post by Dr. Robbert Duker (see Thicket restoration and frost: the forgotten enemies) , overall having detrimental impacts on photosynthetic efficiency of exposed individuals ( Duker et al. 2015).

Abrupt biome  boundary observed between Spekboom-dominated Thicket and Nama Karoo shrubland. (Photo: Robbert Duker, 2014).

This got me thinking that the effects of frost might also extend beyond the photosynthetic machinery of Thicket species (spekboom being primary focus here) and impact the structural components as well, specifically hydraulic conduits in the stem.

At this moment I would like to take a little detour, and chat about fire or more specifically the way in which fire is proposed to kill plants and relating this to the hydraulic death hypothesis (Midgley et al. 2011). To be brief, the hydraulic death hypothesis suggests that plant mortality following intense fire events can be attributed to hydraulic failure by cavitation due to sudden water deficiencies induced by heat plumes or by what Midgley et al. (2011) referred to as run-away hydraulic damage due to leaf water deficits where stomatal response to limit water loss was lagging behind the sudden burst of heat from the fire. Whether by run-away hydraulic failure or cavitation, there where three key predictions/expectations suggesting failure in hydraulic conductivity as the prime cause for mortality. Midgley et al. (2011) described these three predictors/expectations as: (1) post-fire mortality occurred a few days after the fire event, (2) death or damaged tissue was observed forming from the base upwards, (3) the mortality correlated negatively with stem diameter as opposed to height.

Image 2: A spekboom individual planted in a frost-prone area. (Photo: Robbert Duker, 2014)

This fundamental idea of cavitation got me thinking about spekboom and the Thicket Nama-Karoo biome boundary. First off when water freezes it expands and crystallizes, keep this in the back of your mind while we undertake a thought experiment. Imagine a straw with the ends of the straw capped and the tube filled with water, then placing the straw in your freezer. What would happen to that straw as it freezes, it would begin to expand until it burst at any point along its length.  This expanding straw metaphor is analogous to what we might expect in a plant stem comprising of hundreds and thousands of xylem vessels, little straws if you will, running the length of the stem. When frost occurs or rather the temperatures inducing frost formation, the water within these vessels freezes, crystalizes, and expands resulting in hydraulic dysfunction. There are two potential mechanisms of hydraulic dysfunction under frost, firstly, cellular puncture by water crystals resulting in cytoplasmic leakage or in the case of xylem vessels protrusions into surrounding tissue diminishing water driven potential created by intact xylem tissue. The second mechanism is expansion damage of vessel conduits preventing water mobility up the stem. Regardless of the mechanism the outcome is the same, failure in hydraulic conductivity subsequently resulting in mortality.

Bringing us back to the hydraulic death hypothesis in the case of fire there are three expectations, delayed mortality, bottom-up necrosis, and negative correlation of mortality with stem diameter as opposed to height. And while I can speak to the first two of these expectations under the hydraulic death hypothesis induced by frost conditions (see Image 2 as a summary image), the third requires some more data.

Image 3: Frost damage of a single spekboom individual transplanted into a frost-prone area at study site. Notice the damage is from the bottom progressing upwards conforming to the second expectation of the hydraulic death hypothesis (photo: Robbert Duker, 2014).

This "frost hydraulic death hypothesis" has yet to be tested explicitly but has great potential as an explanatory process beyond existing photosynthetic interference concepts explaining this biome boundary. Applying this hypothesis to other frost tolerant and susceptible Thicket species diverging in plant functional traits (e.g. wood density, SLA, turgor loss point, and xylem water potential) may highlight potential species-specific strategies to mitigating frost-water related influences, particularly hydraulic death.


References:

Duker, R., Cowling, R.M., du Preez, D.R., van der Vyver, M.L., Weatherall‐Thomas, C.R. and Potts, A.J., 2015. Community‐level assessment of freezing tolerance: frost dictates the biome boundary between Albany subtropical thicket and Nama‐Karoo in South Africa. Journal of biogeography, 42(1), pp.167-178.

Midgley, J.J., Kruger, L.M. and Skelton, R., 2011. How do fires kill plants? The hydraulic death hypothesis and Cape Proteaceae “fire-resisters”. South African Journal of Botany, 77(2), pp.381-386.

The loss of thirst-refugia and implications for thicket restoration

by Alastair Potts

In a previous post (The flattening of the waterscape and unrecorded loss of thirst-refugia: how does this affect plant biodiversity? ), I discussed how the area that is readily accessible to animals has increased dramatically because a fundamental resource ​— water ​— has been homogenised across the landscape (by humans).

A major driver of this homogenisation process is the advent of fencing the landscape to control both domestic and wild animals. This has a long, and relatively recent, history with many fences being put up from the 1950s onwards using government farming subsidies. However, if animals are going to be kept on a parcel of land, there has to be water. And thus, fencing went hand-in-hand with building dams, waterholes and watertroughs (initially windmill-driven). Every parcel of land had to have at least one water point.

Fast forward a few decades to the present where many farms have been consolidated or repurposed into larger farms ​— usually resulting in the removal of inner boundary fences ​— we have larger parcels of land, but watering points are very rarely removed. Thus, herbivores have greater freedom to roam the landscape, but their densities can remain high because of the short-distances to available water.

Here we need to jump briefly to a different topic ​— high failure rates of restoration initiatives in the subtropical thicket using Portulacaria afra (commonly known as "Spekboom"). We have demonstrated that part of this is due to planting in the wrong parts of the landscape (e.g. in frost zones; Duker et al. 2015a,b). But there is a strong herbivory component to this failure ​— although this has not been directly measured, but field observations suggest that this is a very important part of the puzzle.

Herbivory can come in the form of domestic or wild herbivores. Farmers may keep some stock animals on degraded land ​— but their effect on growth and survival of planted spekboom can be devastating. What is the primary reason why limited stock can be kept on degraded land? ​— the presence of a water point.

Dr Robbert Duker and Dr Marius van der Vyver (both excellent thicket ecologists) report high densities of Greater Kudu retarding the growth rate of spekboom planted in the great Thicket-Wide Plot experiments and neighbouring intact thicket. I suspect the increased densities of Greater Kudu across the subtropical thicket is cause for alarm ​— this is not in line with how this ecosystem worked previously. For example, Jack Skead records that the first sighting of Greater Kudu in 300 years in the Steytlerville area was in 1956 ​— Kudu were unknown in this region before and yet their densities are so high now that driving at night is considered dangerous as the chance of hitting a Kudu is quite high.

And these Kudu are also found in very degraded areas (note that most fences are not barriers to Kudu who readily leap >2 m). Why? The availability of water.

Thus, if we are to give restoration efforts the best chance at succeeding, we need to try and increase the thirstscape ​— to do this, we need to close watering points! There will be areas that are have natural water points (i.e. near rivers), so don't target these for restoration. Thus, any analysis of spekboom restoration potential for the Eastern Cape landscapes should include the waterscape. Target areas that are away from natural or anthropogenic water sources, or include the shutdown of anthropogenic water sources as part of the restoration toolset.


References

Duker, R., Cowling, R.M., du Preez, D.R., Potts, A.J., 2015a. Frost, Portulacaria afra Jacq., and the boundary between the Albany Subtropical Thicket and Nama-Karoo biomes. South African Journal of Botany 101, 112-119.

Duker, R., Cowling, R.M., du Preez, D.R., van der Vyver, M.L., Weatherall-Thomas, C.R., Potts, A.J., 2015b. Community-level assessment of freezing tolerance: frost dictates the biome boundary between Albany Subtropical Thicket and Nama-Karoo in South Africa. Journal of Biogeography 42, 167–178.

The Spekboom Sponge hypothesis

By Alastair Potts

We know that Portulacaria afra, commonly known as "Spekboom" (and the focus of much misguided hype around carbon sequestration), is an ecosystem engineer in its native range in the Eastern Cape province (and the Little Karoo of the Western Cape). 

An ecosystem engineer is any organism that significantly contributes to the creation, modification or destruction of a habitat. In this case, P. afra, increases the water availability in the landscape (van Luijk et al. 2013) for other subtropical thicket tree species that would not normally be able to survive in that landscape (Wilman et al. 2014). It does this by increasing the rainwater infiltration and creating a thick mulch leaf litter (Lechmere-Oertel et al., 2008) which traps water and decreases evaporation.

Our research group was recently in the field exploring how this dynamic helps subtropical trees ​— i.e. what effect does having P. afra around a tree help its water dynamics. Our focus species was Pappea capensis (Jacket-Plumb) and we were measuring the tension of the water column in trees that were either emerging from Spekboom clumps or were in herbivore-degraded parts of the landscape, and therefore standing alone.

Examples of Pappea capensis trees that we assessed for water pressure potential (a measure of plant water stress). Trees were on the same slope and within ~200 meters of one another. Exposed trees occurred lower on the slope where overbrowsing by goats and sheep has removed P. afra. The exposed trees have likely been in this state for >50 years. 

What we found was unsurprising...

The mean water pressure potential (uMPA= mean MPa of 3 replicates per plant) of trees of Pappea capensis occurring alone or within a clump of P. afra (Exposed: n=9, In-clump=11). {Don't ask about the difference in sample size! All I can say, this measurement mission was done between 12h00 and 15h00 with our highest temperature of 47°C ​— our brains were a tad fried}. More negative values means greater water stress!

Thus, the subtropical trees within a P. afra clump, in general, had significantly lower tension (more +ve values on the figure) on their water column than those on exposed slopes without P. afra. {Note: these are preliminary results that require replication ​— that's up to Daniel B.}

But, while we were waiting around for the poor sod who was fetching the latest samples (we tag-teamed as to who had to venture forth from our basecamp — the centre of large P. afra clump — to collect the next sample) ​— we decided to also measure the water tension in P. afra. We were expecting it to be quite a bit lower, but it was consistently around -0.83 to -0.96 MPa.

This is quite a difference in water pressure potential between the succulent P. afra and canopy tree P. capensis. Which got us thinking...! What if P. afra increased water availability in the landscape more than simply by increasing infiltration and decreasing evaporation? If the roots of P. capensis and P. afra grew close together (i.e. touching), then the difference in water pressure potential might be enough for the tree to suck water out of the succulent!

We call this the "Spekboom Sponge" hypothesis: specifically, as trees become more stressed and the tension on their water columns increased, that they might ​— at some point ​— be able to suck water out of P. afra. 

P. afra is able to suck up a lot of water during a rainfall event ​— data from Dr Kathleen Smart's PhD found that P. afra increased its stem diameter by 5-7% post rainfall (×1.05-1.07 of original size). Based on the size of the stems she was analysing (ø: 84.5 & 145.5 mm),  that's ~0.4-0.8 mm³ per cm of stem. The subtropical trees do not have this same ability, and cannot store much water in their stems, and are thus largely reliant on soil moisture for their physiological requirements {ref needed}. Thus, the P. afra, may provide the surrounding plant community with additional water via touching roots systems as the environment becomes drier and drier.

This is quite an exciting hypothesis, as it suggests that P. afra may be more than just an ecosystem engineer but also a "nurse plant" for subtropical trees during the common droughts and generally low rainfall conditions experienced by arid and valley subtropical thicket in the Eastern Cape landscapes.

And of course, Daniel B. (the Duracell Bunny in our research group) found this paper within hours of our return from the field:

"Nurse shrubs can receive water stored in the parenchyma of their facilitated columnar cacti" by Alicia Montesinos-Navarro and colleagues published in the Journal of Arid Environments in 2019.

In this paper, these researchers demonstrate that the columnar cacti, Neobuxbaumia tetetzo, provides water to neighbouring shrubs, Mimosa luisana. 

We'll be testing this hypothesis in the coming year using the same methods as Montesinos-Navarro et al. We'll provide an update as to what we find.

Alastair out...

Update: 20-Feb-2020. After chatting to Ed February, we're also going to be looking at leaf SLA of trees in and out of P. afra clumps as well as leaf δ13C to determine if there have been long-term water contraints (beyond short hot and dry periods, like our pressure chamber measurement mission).

Update: 22-May-2020. Thanks Daniel for picking that I don't know how to convert PSI to MPA! Excuse the crazy readings in the previous figure version...!

References

Lechmere-Oertel, R.G., Kerley, G.I.H., Mills, A.J., Cowling, R.M., 2008. Litter dynamics across browsing-induced fenceline contrasts in succulent thicket, South Africa. South African Journal of Botany 74, 651-659.

Montesinos-Navarro, A., Verdú, M., Querejeta, J.I., Valiente-Banuet, A., 2019. Nurse shrubs can receive water stored in the parenchyma of their facilitated columnar cacti. Journal of Arid Environments 165, 10-15.

van Luijk, G., Cowling, R.M., Riksen, M.J.P.M., Glenday, J., 2013. Hydrological implications of desertification: Degradation of South African semi-arid subtropical thicket. Journal of Arid Environments 91, 14-21.

Wilman, V., Campbell, E.E., Potts, A.J., Cowling, R.M., 2014. A mismatch between germination requirements and environmental conditions: Niche conservatism in xeric subtropical thicket canopy species? South African Journal of Botany 92, 1-6.