Hi Andreas,

your are asking a difficult question.

I doubt that we will ever come up with a single parameter defining 'lethal drought'. This is because different plant functional types have different strategies to avoid/resist/tolerate drought and even within a given plant functional type (e.g., trees) strategies to cope with drought are different. As Mitchell et al. showed, there are different strategies in water use and these, in turn, have implications for the carbon balance. The water spenders risk hydraulic failure in order to maintain carbon assimilation while water savers go for carbon. Depending on how severe and how long the drought is, some may die of dessication while others may die of carbon starvation.

For trees, cavitation repair has been observed in different species but this is a viable survival mechanism only when soil water content is sufficient to allow refilling. On the other hand, trees are perennial organisms and the complete failure of xylem transport may force them to 'skip a season worth of growth' only to continue with new functional xylem next spring (if conditions improve).

More and more evidence is accumulating that drought-induced tree mortality results from interactions between these mechanisms. Declining plant water potential reduces stomatal conductance and carbon assimilation, reduced carbon assimilation increases carbon storage dependency, decreased carbon av availability (reduced assimilation and reserves) impedes maintenance of water potential and xylem functioning via osmotic adjustment and cavitation repair and the cycle repeats. Environmental conditions (e.g., soil depth, windiness) and tree characteristics (e.g., biomass, height, rooting depth) come into play and make everyting even more complicated. Hence, drought-induced mortality is a dynamic process and not a state so by definition no single parameter can be used to predict a 'lethal drought dose'.

I hope this answers your question. Please contact me if you want to discuss this in more detail.

Cheers,

Henrik Hartmann

I think that there is no general concept of a LT50 concept for plants and there is not (and should not) be a single way to access it.

Drought induced mortality in higher plants is mostly caused by cavitation of the xylem tissues. The water potential (or water tension) causing this cavitation depends on the soil water potential, the hydraulic conductivity and the rate of transpiration. The rate of transpiration is modified by the weather. As a result the soil drought, light and the water vapour deficit interact on the damage of plants. If you include resprouting etc issues will become even more complicated.

There are publications on the pressure cavitation curves that can help to characterize plants. These are of some value for assessing drought vulnerability of plants.

Dear Andreas

Drought occurrence in a given field or pot experiment is less related to plant than it is related to the soil type or potting medium. It is actually the texture of soil or physical properties of the potting medium which determines when to stop letting the plant to absorb water. For instance, given a certain level of gravimetric water content, a wheat plant might better survive in a more sandy soil than in more of a clay soil. Therefore, any parametrization of "lethal dose" should take into account all these information. In Theory, for a given soil type or potting medium, a lower bound of water content below which plant can not take up water is called "permanent wilting point - PWP". Theoretically, plants that hit to this level wilt permanently and they do not recover once PWP hit. However, I have experienced potting experiments where re-watering recovered the plants. In the field it is hard to investigate and I have not been directly involved. Maybe in the field experiments rate of recovery is different. Perhaps lethal dose as you referred to in the field is nearly PWP.

Sincerely

Mohsen

Andreas, science still doesn't have that answer. Physiological drought is one of the most complex phenomena of botany. Like any complex trait, it must be studied with the most comprehensive technique. It is usually a traditional and simple technique, such as "survivors counting " in some time periods. From the comprehensive technical results, you can try to find one or more factors that better correlate with it. But, can you imagine the effect of DNA and membranes damage? These effects are cumulative and only be measured over time. No use the plant initially survive, but is already doomed to die in a few months.

Hi Andreas,

your are asking a difficult question.

I doubt that we will ever come up with a single parameter defining 'lethal drought'. This is because different plant functional types have different strategies to avoid/resist/tolerate drought and even within a given plant functional type (e.g., trees) strategies to cope with drought are different. As Mitchell et al. showed, there are different strategies in water use and these, in turn, have implications for the carbon balance. The water spenders risk hydraulic failure in order to maintain carbon assimilation while water savers go for carbon. Depending on how severe and how long the drought is, some may die of dessication while others may die of carbon starvation.

For trees, cavitation repair has been observed in different species but this is a viable survival mechanism only when soil water content is sufficient to allow refilling. On the other hand, trees are perennial organisms and the complete failure of xylem transport may force them to 'skip a season worth of growth' only to continue with new functional xylem next spring (if conditions improve).

More and more evidence is accumulating that drought-induced tree mortality results from interactions between these mechanisms. Declining plant water potential reduces stomatal conductance and carbon assimilation, reduced carbon assimilation increases carbon storage dependency, decreased carbon av availability (reduced assimilation and reserves) impedes maintenance of water potential and xylem functioning via osmotic adjustment and cavitation repair and the cycle repeats. Environmental conditions (e.g., soil depth, windiness) and tree characteristics (e.g., biomass, height, rooting depth) come into play and make everyting even more complicated. Hence, drought-induced mortality is a dynamic process and not a state so by definition no single parameter can be used to predict a 'lethal drought dose'.

I hope this answers your question. Please contact me if you want to discuss this in more detail.

Cheers,

Henrik Hartmann

Before addressing the equivalent of tolerance limits for physiological drought, we need to go more deeply into the concept of drought. The idea of physiological drought is certainly an advance over the raw concept of drought derived from rainfed cropping. It incorporates the twin ideas of Ernst Detlef-Schulze (1986, Annu Rev Plant Physiol) and of Grieu et al. (1988, Physiol Plant) who distinguished atmospheric drought (high evaporative demand, though "demand" is a problematic concept itself) from soil drought (lack of soil water per se). However, water shortage takes diverse forms as time series, so we should distinguish episodic drought from terminal drought (as occurs in Mediterranean climates). Different species have different physiological acclimation suites and different developmental programs to deal with terminal vs. episodic drought - you don't see species native to Iowa surviving in Western Australia. We should resolve stress (physiological drought) by at least three dimensions: duration, depth, and frequency.

To return to tolerance limits for drought, the concept of drought frequency, especially, brings up the fact that tolerance itself acclimates. An earlier nonlethal drought commonly conditions a plant physiologically to the current drought, as by ABA accumulation, altered rooting patterns, etc. So, there is no LT50, unless one makes it a dynamic thing. Saying that there is an LT50 is like saying there is a fixed magnitude of stomatal conductance, gs. Since gs varies greatly but follows a control program, it is overwhelmingly more useful to view the control parameters (slope and intercept in the simple Ball-Berry model, variously amended in form and to incorporate water stress) as the fundamental descriptors and not gs values at some arbitrary environmental condition. Some regularities appear, such as that, in mesic plants, a Ball-Berry slope near 10 is almost universal. We need a few-parameter description of drought tolerance (not "few" as in oversimplified but few as in concise and fundamental).

Extending this line of thought, we need to use an evolutionary context. Too much drought research harks back to irrigated agriculture of annual crops. Why are there species (nay, genotypes) that are so divergent in their tolerance of environmental conditions? Why isn't there a superstrategy to handle diverse environments? The idea of a superstrategy is unspoken in agricultural research, and it is very misleading. The selection pressures on wild plants and on crop plants are so different and this must be recognized (I wrote about this at length in my 1987 book, A Functional Ecology of Crop Plants).

The evolutionary view of drought tolerance (DT) strategies has not been developed in a comprehensive framework. We have to recognize that drought tolerance is thrown in, for natural selection, with strategies for physiological competitiveness (Iphotosynthetic traits, rooting traits) for autecological performance and for traits for synecological performance (height competition, pollinator attraction, ...). In any one genotype, DT is not fully optimized (nor can it ever be perfect) when it must be in a joint optimization, traded off against degrees of optimization in other performance measures such as those I just mentioned. Mathematically, we have a case of constrained optimization for multiple objectives. It takes a great deal of quantitative formulation of plant performance to calculate this and comprehend its significance. Moreover, we have a system with stochastic drivers in the environment (rainfall stochasticity, e.g.). The proper formulation is one of risk management. A good start on this was made by Paltridge and Denholm, formulating optimal timing for the switch from vegetative growth to reproductive growth in an annual plant under the risk of an uncertain end to the growing season (frost, drought). We have to go up to many dimensions. Another interesting lead was made (Jones and Zur, 1984, J. Irrig. Sci.) in formulating plant performance in drought in terms of something like 20 parameters for stomatal control, rooting, etc., as I recall, and then running multiple simulations to see if an optimal set of parameters could be found. Plant performance optimization studies are common (Cowan and Farquhar, 1976, on to multiple studies by Farquhar and colleagues; Gutschick and Wiegel, 1988 and diverse related studies, including a particularly nice one by Schieving and Poorter). The mathematics of optimizing many parameters is not problematic (e.g., there are methods of simulated annealing or genetic algorithms); the problem is in discovering the fundamental parameters of plant physiology and development. We had a very fruitful week of discussion in 2010 that went into this topic repeatedly (a workshop at the Mathematical Biology Institute at Ohio State; we hope to publish our findings), but we have a long way to go.

A final point is that we need a forensics of plant death, as Marilyn Ball of the ANU and I talked about long ago. Why does any one plant die? There has been some very good discussion in the past decade, such as work put out by Nate McDowell, Craig Allen, Dave Breshears, and others. They worked to distinguish carbon starvation from water-stress-induced cell death and other phenomena. The phenomena are linked, of course. Carbon starvation disallows a plant from expressing tolerance and recovery mechanisms such as growth of new vascular tissue or new roots. Cell death from stress is both a result of stress and (in key places, such as some vascular tissue) a cause of stress. Back to the first sentence - we need a coherent framework to discuss mortality of plants (and, as physicians also say, morbidity, a significant decline in function). When a human patient dies from complications of AIDS, such as pneumonia from an opportunistic infection, what is the cause? There are proximate causes (pneumonia...or the microbial species itself, or its availability...) and more nearly ultimate causes (AIDS...). Death is more of an ecosystem function (the patient, the HIV virus, the pneumonia microbe, the social milieu in which the patient acquired exposure to both) than an individual function or phenomenon. A view above the level of the organism is also made more appropriate when we consider that it's genes, not finite-lived individuals, that are the currency of the biosphere.

I've laid out a very long roadmap into some terra incognita. However, like other explorers we have to go there. Otherwise we are in little parochial enterprises, seen in the large. Valuable as they may be in some area, our research efforts are then self-limiting. At least when we do explore in this mode, I think we won't be extinguishing native populations (oops - the Green Revolution did some of that, to small farmers; careful!).

Dear all, thanks a lot for your replies. I am sorry to be not precise enough with my question. So I will try to clarify my thoughts. It is clear for me that there will be no single concept that can be applied, and I am also quite familiar about the ongoing debate on hydraulic failure vs. carbon starvation as well as the complexity of drought-induced mortality as a process.

I would like discuss about the definition of a 'lethal dose' of drought (e.g. measured by xylem water potential as a measure for 'physiological drought') experienced by a plant population that lead to the proven death of 50% of the individuals of a tree population. This would be a 'fuzzy' empirical indicator, off course. But with this approach any problems with pot size, soil structure, climatic drought intensity and length etc. should be avoided. And this idea is different from producing pressure-cavitation curves for individuals. I am not trying to explain the mechanisms behind drought-induced mortality, but try to get an idea about the impact dose (as a complex factor).

Vincent, I completely agree with you that a LT50 value according to your definitions is static. But could it be a solution for the dynamic responses of plants to show the full spectrum of LT50. So isn't it perhaps more a question how to apply and evaluate such a static indicator?

This is an interesting question with possibly a complex answer. However Kursar et al 2009 explore the use of Lethal Dessication 50 which in a way is similar to Lethal Dose as in this study was related to plant/species performance in the field. I would think that relating loss of conductivity might be tricky as you mention vessel reffilling capacity could play a huge role. Also by looking at the shape of vulnerability curves makes me think that vessel and pit resistance to cavitation distribution might makes things different for two species with similar, lets say P50 or P75.

May be, species specific PWP 50 could be standardized in a given set of conditions.

Thanks, Omar. That was a good hint.

I think it depends on the type of plant, it differs between plants having an avoiding strategy and resisting strategy.

I know I tried on Arabidopsis thaliana, although it isn't a drought resisting plant, but I studied the variation of Sugar transporter genes experssion during drought and I performed a phenotypic study in parallel on three trials of 300 plants each. Plants were studied individually. In general, I measured the stomatal conductivity, leaf surface, mass variation, soil water and water ration in all plants, I assume that the critical point in short days conditions was reached when water ration inside plants dropped till an interval if 30 to 20%.

Thanks to all for the ongoing discussion. I would like to add another point to the discussion. Perhaps you are able to follow the idea of lethal or critical xylem water potential for different species (as a rough, emprical indicator with a broad value range and with all justified limitations mentioned by Henrik and others above) where 50% or even 75% of a plant population is seriously endangered to get lost (e.g. P50/P75 according to Omar). Then, it should be possible to link this 'lethal dose' of drought to a critical soil matrix potential at the lower end of the effective rooting zone (the lowest soil depth where the plants can deplete soil water resources). This relationship was shown by Breda et al. (1995)* in oak stands in France. Based on this idea we introduced the term 'Critical limit of soil water availability (CL-SWA)' by Czajkowski et al. (2009, see link attached), since soil matrix potential is linked to soil water availability by PF curve characterstics. We defined the CL-SWA as the proportion of plant-available water within the variable effective rooting depth (ERD) that meets both the critical soil water potential at the lower limit of the ERD and the critical plant xylem water potential (cf. the 'lethal' dose of drought).

I am very interesting in your opinion about these ideas.

*Bréda N, Granier A, Barataud F, Moyne C (1995) Soil water dynamics in an oak

stand : I. Soil moisture, water potentials and water uptake by roots. Plant

Soil 172:17-27

Article Critical limits of soil water availability (CL-SWA) for fore...

I think that a lethal dose as a concept for comparison of species makes sense. However, since soil drought will interact with other conditions (light, VPD) the responses across field experiments will not be comparable. I think that for small plants this makes definitely sense. Large plants (=trees of a certain size) will have or not have access to ground water and deep soil layers and concepts of soil water potential will become a bit diffuse. However, the approach might work (but would require that the user of the results knows the limits).

I agree with Henrick All the best

This lethal dose depends mainly upon the following:

1- the genotype ( senestive or tolerant)

2- time of application ( i.e. stage vegatativ, flowering or fruiting)

3- duration of drought ( moderate drought or severe drought

There are many excellent answers already, but I woúld like to add some importnat point. I thjnk that one shoukl consider that even in one plants and in one organs cells may have different sentivity to drought. For example, in dicotyledon plants mesophyll cells in the fully expanded leaf are very sensitive to drought because because they have high water contents and large lytic vacuol (ratio between fresh and dry weight are very high) At the same time, meristemic tissue have a storage vacuol with low water contents and are much more resitant to drought stress.

There is a great body of research ion this. You might look at the distributio of algae on a seashore and note the recovery rates of various species from extreme draught. The deep water eps have less ability to recover, while those higher up the shore can often appear to be dead but will still recover when the tide eventually covers them or soaks a rock pool. Thus it is a variation of the structure an physiology of the cell contents.

It is a great idea but difficult to identify. Drought induced physiological changes of plants are diverse, the reasons for plant died beause of grought may vary. I am wondering that a " lethal potential" for plants should relate to SPAC concept. In another words, the "lethal potential" may be a further development for SPAC...

This is good stuff people! Thanks for the discussion.

To add another dimension to this discussion one should also consider that the capability of plants to avoid or tolerate drought shifts seasonally with changes in their phenological development. Drought that is lethal at certain times of the year can cause minimal stress at other times of the year. All plants have this ability to shift their capability to handle drought. For example:

Grossnickle, S.C. 1989. Seasonal shoot phenology and water relations of Picea glauca. Can. J. For. Res. 19:1287-1290

Hello all,

thanks again for all these interesting comments and hints. Maik, the stress recovery (time) is off course an interesting approach. However, I guess that lethal dose concept (perhaps similar to the PWP concept) addresses another 'trait' than recovery after a non-lethal drought stress. Both 'traits' can correspond or not.

Inspired by this discussion and the very interesting paper of Kursar et al. (2009) about the lethal dessication (LD 50) I formulated following approach:

"Based on the ‘physiological drought’ concept, we present a ‘lethal dose’ approach, related to soil water deficit (LSWD) for the major central European forest tree species.

The pre-dawn water potential (ψpd) is the key parameter for assessments of ‘physiological drought’. According to the equilibrium concept of the soil-plant water potential at pre-dawn time, ψpd relates to the soil water potential at the lowest soil depth in which plant’s root system is able to deplete water resources. This rooting depth meets the definitions of the ‘effective rooting depth’ (ERD). The ‘lethal dose’ of soil water deficit (LSWD) represents the critical proportion of plant-available water within the effective rooting depth (ERD) that meets both the critical soil water potential at the lower limit of the ERD and the critical ψpd.

A lethal impact dose for plant collectives is normally defined as a threshold of 50% mortality. By applying this concept, the soil water deficit (ERD), where 50% mortality in forest tree populations occurs, can be used for determining the L50SWD. This L50SWD indicator can be easily implemented in combined climate and soil water models in order to assess potential sensitivity of different forest tree species or species provenances to (future) increased drought events."

Andreas, i just inquire that for annual plants wheat/brassica etc the rooting in soil could be more uniform, as well as the soil; how do you see these two factors in trees particularly in natural ecosystem of forests?

I think that growth vigor of root system especially root length, root distribution and root biomass as well as root shoot ratio. These previous characters can be used as selection criteria for drought tolerance

Rooting is a very essential trait for drought adaptation of plants (annual plants and perennial plants like trees). However, it is not an easy job to determine root system functioning in plants' water supply. Off course, you can look for fine root traits like root length, biomass, surface area and tip number. But this might be a uncertain proxy for root system ability to deplete soil water resources, since only a little part of the finest roots have active surfaces for water uptake, and this area is not easy to measure. And this parameter need not to be correlated to typical fine root triaits mentioned above. Thus, I am more convinced of measurements on the functioning xylem system and its integrity (like xylem water potential, loss of hydraulic conductivity and sapflow). These data are intergrating the whole water transport system (root-stem-twig-leave xylem). Concerning my LSWD concept an increased drougt adaptation due to increased rooting (e.g. shown by a raising root/shoot ratio) are included by a variable effective rooting depth (ERD) and also by inceasing the lethal soil water deficit if plant water uptake is increased by structural variations of the fine root system (more tips, active surfaces).

Hi Andreas, I'm a plant pathologist not a physiologist so my answer should be taken with that warning in mind. To me, your question seems best addressed from an evolutionary perspective. Your questions seems to me to contain two elements; a generic truth and a variable truth. The generic truth is that all plants, as all organisms, will die if their supply of water is restricted below their absolute requirement for survival. The idea that there is a lethal dose of water unavailability seems to flow (if you'll pardon the pun) from this generic truth. On the other hand, given adaptation to diverse life strategies the actual lethal dose of unavailability of ater (and hence its 50% value) is likely to be context and species specific. In the ecologica literature I think Tillman's R_star concept seems to capture the concept quite well. Since I recenrtly moved from Scotland to California I have personal experience of needing to revise my concept of water availability. Here in northern C one of the worst invasive plants in star thistle (Centaurea solstitialis) which creates local droughts in habitats that it invades by extracting water from the soil to a percentage content at which most other competitors cannot survive.

Undoubtedly then, there is a "lack of water availability" at which we would expect all individuals of a species to die. Consequently there is a value (probably more accurately a distribution) at which half of the population would be expected to die. However, these values will vary with the local range of genotypes and the particular details of the water holding characteristics of the soil (not to mention the activities of the other species in the community).

I think that we might be losing focus here. To use a famous phrase," it depends on what you mean by physiological drought."

Are we including the algae in this discussion? If so, stressed xylem can not be a measure. Take a look at all the mechanisms of morphology to prevent water loss. Xerophytes are an example. Anatomical features such as thick waxy cuticles and responses to acute cold such as intense curling of leaves. Looking at the anatomy of the roots of terrestrial plants might be a thought. However when stripping all these variables away from the exercise we are left with tissues. Tissues are also variable according to cell contents, thickness and composition of the cell walls. Now, what are we left with? Basically living tissue which will vary in physiological differences according to individual sps. Here lies the answer. Hydraulic conductivity may be the indicator but ultimately tissue failure may be the answer as this would eliminate regeneration from the roots or dormant storage organs.

Plant death is not not a stress but the end of life. When using a lethal dose-50 we probably mean than 50% of plants under scrutiny are dead or the reverse metric is mean water deficit at 50 % of survival, sort of translate from mean life duration in water units. The function boundaries are the values at 0% and 100% and the functional link is experimental, but there is a variety of water contents that do not affect survival, but leaf number RGR and DW. I wonder if lethality simplifies the description, but the empirical value of a Gomperz equation is related to cavitation in a restricted number of cases.

Hi all

very interesting discussion group.

We published some years ago a couple of papers, the first related to recovery capacity under drought stress and different degrees of xylem cavitation. The second related with coordination among symplastic and apoplastic processes during dessication. Unfortunately we could only study two species but both were drought resistant and results very interesting. Here the references if you are interested in them. Regards. Alberto

- VILAGROSA, A., ET AL. 2010. Are symplast tolerance to intense drought conditions and xylem vulnerability to cavitation coordinated? An integrated analysis of photosynthetic, hydraulic and leaf-level processes in two Mediterranean drought-resistant species. Environmental Experimental Botany 69: 233-242. DOI. 10.1016/j.envexpbot.2010.04.013.

- VILAGROSA, A., et al. 2003. Cavitation, stomatal conductance, and leaf dieback in seedlings of two co-occurring Mediterranean shrubs during an intense drought. Journal of Experimental Botany 54: 2015-2024.

Thanks to dr Vilagrosa, lethal "dose" is probably a bad metric for describing life support relative to cavitation. Cavitation mimics a survival curve but in terms of % vessels cavitated versus pressure in a twig. This suction depends on transpiration and flow resistance. Pre-desertic vegetation has a very low leaf-to-sapwood area ratio and limited transpiration. A lot of leaves did not unfold or were shed, but with the plant alive. Osmotic adaptation occurs and by going further, is the time of colonization of ephemerals or succulents. Cold at 5°C or below, and dry long exposure are however unbeatable. Temperature makes a strong variability also on cavitation.

Hi all, thanks again for these new contributions and the interesting papers, Alberto. Yes, I agree with nearly all of your concerns on the limitation of my proposed concept. However, looking on the state-of-the-art ecological modelling on drought limitation for plants I see that often very basic 'expert assessments' for critical drought limits are used like 'European beech cannot grow below a annual precipitation rate of less than 500 mm' or climate envelope models that use annual means for temperature and precipitation for presence/absence assessments . Thus, I see a large gap between the very sophisticated functional research in plant physiology and the very basic information used in large-scale ecological modelling. So I am very open and interested to alternative ideas how to bridge this gap.

Yours, Andreas

Hi Andreas,

the discussion is very interesting, moving continuously among different perspectives.

Probably your final comment is not more related to a single plant perspective as in first original question, but more related to the ecosystem level.

When you consider a well-established forest (not a young tree plantation), the common observation is that after a certain level of stress (your parameter could be useful in that sense), the defence mechanisms are violated by different other organisms which in turn can determine the plant death. The development of the pathogens will follow the dynamic of a growing population, which produce in turn the dynamic of trees mortality.

Additionally, in a forest you should also consider that different structural characteristics (age, dimension, spatial arrangements, species composition - as resulting from the management) interacts with the dynamic of the trees mortality after an important strees event.

Probably a different way to see your question could be about the weight of the different ecosystem characteristics, when compared to a common LSWD (tree based), on the dimension of trees mortality.

Take a look at this paper that provides a thoughtful and detailed discussion on this topic.

McDowell N, Pockman WT, Allen CD, Bershears DD, Cobb N, Kolb T, Plaut J, Sperry J, West A, Williams DG, Yepez EA (2008) Mechanisms of plant survival and mortality during drought: why do some plants survive while others succumb to drought? New Phytol 178:719-739

Dear all,

thanks again for the inspiring discussions that we have used for developing the L50SWC indicator. Attached you will find a link to a recent publication in Frontiers in Plant Science where we are describing the indicator and applied them to study extreme drought response in European beech population througout Europe. I would appreciate your comments on this.

Thanks and best regards,

Andreas

Article Desiccation and Mortality Dynamics in Seedlings of Different...

Hello Andreas,

Interesting paper. Thanks!

What I find heartening is that we also used SWA (in a different format) to harden, but not cause mortality in operational nursery cultural programs. In these programs we used a measure of Container Capacity (i.e. the amount of water retained in the soil profile after drainage of the macro pores from a point of container saturation). Your figure 4 is exactly what we determined. Thus we could grow plants down to around 50% CC, harden plants between 50% & 30% CC, but avoid going below 25% CC so we would not have any seedling death. It seems that many tree species respond in a similar manner to the SWA around their root systems.

Steve