Modelling of underwater light in freshwater lakes using survival and failure time analysis

1. This study presents a new approach to modelling subsurface irradiance using concepts from survival and failure time analysis. The model applies a modified Weibull distribution function to predict downwelling irradiance. Data sets from forty-seven Norwegian sites show extremely high coefficients of determination, up to 99.99%, when analysed by the Weibull model.
2. The uncritical use of a single k _d value to model underwater light conditions is likely to result in poor estimates of received irradiance. This error may amount to several hundred per cent. The practice of force-fitting linear least-squares regressions to log-transformed irradiance data inevitably leads to highly biased estimates of the true fraction of incident irradiance entering the water.
3. Wave effects causing fluctuations of subsurface irradiance are modelled with synthetic data and compared with field observations. Fluctuations of surface elevation by waves produce skewed frequency distributions of the underwater light field. The result of these effects, which are to reduce the accuracy of estimated model parameters, can be largely eliminated by carefully designing field procedures used for the acquisition of subsurface light data.     

Modelling the photosensitization-based inactivation of Bacillus cereus

Aims:  To study and to develop a model for the photo-destruction of the foodborne pathogen Bacillus cereus , initially treated with a precursor of endogenous photosensitizers (5-aminolevulinic acid, ALA).
Materials and methods:  The cells were incubated in the presence of ALA (3 or 7·5 mmol l⁻¹) for incubation times ranging from 2 to 60 min, inoculated onto the surface of LB Agar plates and submitted to light irradiation. The Weibull model was used to describe the survival curves of B. cereus . Quadratic equations were used to describe the effects of ALA concentration and incubation time on the Weibull model parameters.
Results:  ALA-based photosensitization proved to be an effective tool for inactivation of B. cereus . The decrease in viable counts observed after 20 min of irradiation, ranged from 4 to 6 log CFU g⁻¹.
Conclusions:  The developed model proved to be a parsimonious and robust solution to describe the observed data.
Significance and Impact of the Study:  The study demonstrates the effectiveness of photosensitization on B. cereus on agar plates. The model developed may be useful to optimize inactivation treatments by photosensitization.     

Statistics of the Underwater Light Field: An Empirical Model

The underwater light field is described as a stochastic process, with the vertical attenuation as a random variable. It is shown that the vertical attenuation integral is well described as a Normal process with uncorrelated increments. The attenuation processes within a water body are found to be quite independent of the incoming irradiance. Thus, the relative light intensity at any depth can be approximated by a Lognormal random variable. Based upon the expectation of this Lognormal variable, and the mean value of the incident irradiance, the irradiance delivered at any depth can be estimated, as well as the statistical distribution of the irradiance. Light data from a number of Norwegian soft-water lakes showed good fit to the model. Two productive lakes included in this study also had a light regime well described by the statistical model. However, the model should be extended to cater for seasonal variations in these applications.     

A seasonal sequence of diel distribution patterns for the planktonic dinoflagellate Ceratium hirundinella in a eutrophic lake

SUMMARY. 1. Diel depth distribution patterns of Ceratium hirundinella were studied during eleven sampling periods, covering the seasonal growth cycle. They were shown to result from short-term periodic or non-periodic external factors, endogenous responses of the alga, and interactions between these.
2. Diel variations in wind stress resulted in the net transport of algae into or out of the sampling position due to lateral water movements. A progressive surface accumulation of Ceratium , leading to a 3-fold increase in cell numbers at the sampling site over a 24 h period, was due to wind-induced upwelling of deeper cells. Near-surface accumulation of Ceratium on a completely overcast afternoon, similar to that associated with migrations, was attributable to advection.
3. Under sufficiently calm conditions, depth-differentiation of Ceratium was regulated by its vertical swimming movements with a diel periodicity. Migration patterns observed in earlier work were confirmed and extended. The alga migrated towards the surface during the daytime and downward during the night; this rhythm had an endogenous component.
4. At high surface illumination, the alga retreated from the surface and formed discrete sub-surface maxima; at low irradiance Ceratium showed positive phototaxis and concentrated near the surface. Under either condition of irradiance Ceratium actively aggregated at depths associated with irradiance levels of about 125–155 μEinsteins m ⁻² s ⁻¹. Downward movement was restricted by anoxic conditions and possibly by thermal/ density gradients. However, given sufficient light penetration, dissolved oxygen and nutrient availability, the thermal density gradient does not apparently eliminate downward movement by Ceratium.     

The role of phytoplankton in determining the underwater light climate in Lake Constance

At all seasons, the underwater light field of meso-eutrophic large (480 km²) deep (mean: 100 m) Lake Constance was studied in conjunction with the assessments of vertical distributions of phytoplankton chlorophyll concentrations. Vertical profiles of scalar, downwelling and upwelling fluxes of photosynthetically available radiation, as well as fluxes of spectral irradiance between 400 and 700 nm wavelength were measured.The overall transparency of the water for PAR is highly dependent on chlorophyll concentration. However, the spectral composition of underwater light is narrowing with water depth regardless of phytoplankton biomass.Green light is transmitted best, even at extremely low chlorophyll concentrations. This is explained by the selective absorption of blue light by dissolved organic substances and red light by the water molecules. Nevertheless, significant correlations were found between vertical attenuation coefficients of downwelling spectral irradiance and chlorophyll concentrations at all wavelengths. The slopes of the regression lines were used as estimates of chlorophyll-specific spectral vertical light attenuation coefficients (K _c()).The proportions of total upwelling relative to total downwelling irradiance (reflectance) increased with water depth, even when phytoplankton were homogeneously distributed over the water column. Under such conditions, reflectance of monochromatic light remained constant. Lower reflectance of PAR in shallow water is explained by smaller bandwidths of upwelling relative to downwelling light near the water surface. In deeper water, by contrast, the spectra of both upwelling and downwelling irradiance are narrowed to the most penetrating components in the green spectral range. Reflectance of PAR was significantly correlated with chlorophyll concentration and varied from 1% and 1-% at low and high phytoplankton biomass, respectively. Over the spectrum, reflectance exhibited a maximum in the green range. Moreover, in deeper layers, a red maximum was observed which is attributed to natural fluorescence by phytoplankton chlorophyll.     

Effects of light history on primary productivity in a phytoplankton community dominated by the toxic cyanobacterium Cylindrospermopsis raciborskii

1. The effects of instantaneous irradiance and short‐term light history on primary production were determined for samples from a subtropical water reservoir dominated by the toxic cyanobacterium Cylindrospermopsis raciborskii. ¹⁴C‐bicarbonate uptake incubations were conducted on water samples from the reservoir, for irradiance (photosynthetically active radiation) ranging from 0 to 1654 μmol quanta m⁻² s⁻¹. Prior to the ¹⁴C incubations, cells were pre‐treated at irradiance levels ranging from 0 to 1006 μmol quanta m⁻² s⁻¹. 2. The average irradiance experienced by cells during the 2–2.5 h pre‐treatment incubations affected the productivity–irradiance (P–I) parameters: exposure to high light in pre‐treatment conditions caused a substantial decrease in maximum rate of primary production P_max and the photoinhibition parameter β when compared to cells pre‐treated in the dark. 3. While the data collected in this study were not sufficient to develop a full dynamic model of C. raciborskii productivity, P_max and β were modelled as a function of pre‐treatment irradiance, and these models were applied to predict the rate of primary production as a function of both instantaneous and historical irradiance. The results indicated that while cells with a history of exposure to high irradiance will be the most productive in high irradiance, production rates will be highest overall for dark‐acclimated cells in moderate irradiance. 4. Our results may explain why optically‐deep mixing favours C. raciborskii. If the mixing depth z_m exceeds the euphotic depth z_eu, cells will be dark‐acclimated, which will increase their rate of production when they are circulated through the euphotic zone. These results also predict that production rates will be higher during morning hours than for the same irradiance in the afternoon, which is consistent with other phytoplankton studies. 5. Since the rate of production of C. raciborskii‐dominated systems cannot be described by a single P–I curve, accurate estimates of production rates will require measurements over the daily light cycle.     

Mixotrophic ciliates in an Andean lake: dependence on light and prey of an Ophrydium naumanni population

1. Planktonic ciliates were examined during a spring–summer period (November 1998–April 1999) in the ultraoligotrophic Lake Moreno Oeste (41°5' S and 71°33' W, 758 m a.s.l), which belongs to the Nahuel Huapi System (Patagonia, Argentina). The lake is deep ( Z _max=90 m) and warm monomictic.
2. Sampling was performed at a mid-lake station, where vertical profiles of temperature and light were measured in situ , and samples for bacteria and ciliates enumeration were taken throughout the water column.
3. The peritrich Ophrydium naumanni , a freshwater pelagic ciliate with endosymbiotic Chlorella , was the dominant ciliate in the lake.
4.  Ophrydium naumanni and autotrophic picoplankton exhibit a clear coincidence in their vertical distribution ( P  < 0.05), preferring levels at or near the 1% of surface photosynthetically active radiation (PAR) irradiance. Both may have the same light requirements, or the coincidence may reflect a trophic relationship.
5. Dependence on light and prey by O. naumanni were studied using field experiments, in which we analysed ciliate grazing on bacteria, and in laboratory experiments, in which we compared particle uptake under dark and light conditions.
6.  Ophrydium naumanni was able to ingest particles [latex microspheres and fluorescently labelled bacteria (FLB)] in field and laboratory experiment, indicating that it has the potential to affect bacteria population of Lake Moreno Oeste.
7. Ciliate particle ingestion was observed to be dependent on light availability because under dark conditions, the ingestion was lowered ( P  < 0.05).     

Inherent and apparent optical measurements in the Hudson/Raritan Estuary

Duringan August, 1999 field campaign, measurements were made to establish hydrologicoptical properties of the Hudson/Raritan Estuary (New York-New Jersey): 1)concurrent above-and below-surface spectral irradiance; 2) sampling forlaboratory determination of inherent optical properties; and 3) concentrationsof optically-important water quality parameters. We used a bio optical modelbased on to predict thesubsurface irradiance reflectance from optically important water constituents.This model was then validated with the measured reflectance spectra from thefield spectroradiometers. Modeling of reflectance is a prerequisite forprocessing remote sensing data to desired thematic maps. These are key input tothe geographic information system (GIS) used to manage the water qualitycondition of the estuary.     

Comparison of methods for analysis of selective genotyping survival data

Survival traits and selective genotyping datasets are typically not normally distributed, thus common models used to identify QTL may not be statistically appropriate for their analysis. The objective of the present study was to compare models for identification of QTL associated with survival traits, in particular when combined with selective genotyping. Data were simulated to model the survival distribution of a population of chickens challenged with Marek disease virus. Cox proportional hazards (CPH), linear regression (LR), and Weibull models were compared for their appropriateness to analyze the data, ability to identify associations of marker alleles with survival, and estimation of effects when all individuals were genotyped (full genotyping) and when selective genotyping was used. Little difference in power was found between the CPH and the LR model for low censoring cases for both full and selective genotyping. The simulated data were not transformed to follow a Weibull distribution and, as a result, the Weibull model generally resulted in less power than the other two models and overestimated effects. Effect estimates from LR and CPH were unbiased when all individuals were genotyped, but overestimated when selective genotyping was used. Thus, LR is preferred for analyzing survival data when the amount of censoring is low because of ease of implementation and interpretation. Including phenotypic data of non-genotyped individuals in selective genotyping analysis increased power, but resulted in LR having an inflated false positive rate, and therefore the CPH model is preferred for this scenario, although transformation of the data may also make the Weibull model appropriate for this case. The results from the research presented herein are directly applicable to interval mapping analyses.     

The spectral absorption and scattering properties of dissolved and particulate components in relation to the underwater light field of some tropical Australian fresh waters

SUMMARY. 1. The inherent optical properties (scattering coefficients and absorption coefficients across the photosynthetic waveband) are presented from diverse tropical water bodies (billabongs) in the Alligator Rivers Region of northern Australia.
2. The data are used to interpret observed characteristics of the underwater light field as exemplified by the spectral distribution, and overall rate of attenuation, of photosynthetically available radiation (PAR).
3. Attenuation of PAR, especially in the blue waveband, is caused primarily by intense light absorption by the yellow-brown humic pigments, both soluble and particulate, in the water.
4. It was estimated that in six moderately turbid billabongs, light scattering increased attenuation by an average of 58% above that attributable to absorption alone, whereas in a highly turbid billabong the increase was 111%.
5. A distinguishing feature of the optical character of these billabongs, compared with previously studied water bodies in southern Australia, is the great contribution to light absorption made by the particulate humic material.     

Phytoremediation removal rates of benzene,toluene, and chlorobenzene

Phytoremediation is a sustainable remedial approach, although performance efficacy is rarely reported. In this study, we assessed a phytoremediation plot treating benzene, toluene, and chlorobenzene. A comparison of the calculated phytoremediation removal rate with estimates of onsite contaminant mass was used to forecast cleanup periods. The investigation demonstrated that substantial microbial degradation was occurring in the subsurface. Estimates of transpiration indicated that the trees planted were removing approximately 240,000 L of water per year. This large quantity of water removal implies substantial removal of contaminant due to large amounts of contaminants in the groundwater; however, these contaminants extensively sorb to the soil, resulting in large quantities of contaminant mass in the subsurface. The total estimate of subsurface contaminant mass was also complicated by the presence of non-aqueous phase liquids (NAPL), additional contaminant masses that were difficult to quantify. These uncertainties of initial contaminant mass at the site result in large uncertainty in the cleanup period, although mean estimates are on the order of decades. Collectively, the model indicates contaminant removal rates on the order of 10^?2–10⁰ kg/tree/year. The benefit of the phytoremediation system is relatively sustainable cleanup over the long periods necessary due to the presence of NAPL.     

Energy requirement for foliage formation is not constant along canopy light gradients in temperate deciduous trees

Foliage construction cost (glucose requirement for formation of a unit foliar biomass, G , kg glu kg⁻¹), chemical composition and morphology were examined along a light gradient across the canopies in five deciduous species, which ranked according to increasing shade-tolerance as Populus tremula < Fraxinus excelsior < Tilia cordata = Corylus avellana < Fagus sylvatica . Light conditions in the canopy were estimated by a hemispheric photographic technique, allowing ranking of sample locations according to long-term light input incident to the sampled leaves (relative irradiance). G and foliage carbon concentration increased with increasing relative irradiance in F. excelsior , T. cordata and C. avellana , but wereindependent of irradiance in F. sylvatica and P. tremula . However, if G of non-structural-carbohydrate-free dry mass was considered, it also increased with increasing relative irradiance in P. tremula . A positive correlation between the concentration of carbon-rich lignin and irradiance, probably a result of the acclimation to greater water stress at higher light, was the major reason for the light-dependence of G . Lignin concentrations were highest in more shade-tolerant species, resulting in greatest carbon concentrations in these species. Since carbon concentration and G are directly linked, the leaves of shade-tolerant species were also more expensive to construct. As the result of these effects, G increased faster with increasing leaf dry mass per area which was mainly determined by relative irradiance, in shade-tolerators. Given that shade-tolerant species had lower leaf dry mass per area at common irradiance and that this saturated at lower relative irradiance than leaf dry mass per area in the intolerant species, it was concluded that enhanced energy requirements for foliage construction might constrain species morphological plasticity and the upper limit of leaf dry mass per area attainable at high light.     

Ecological and physiological investigations in continental Antarctic cryptogams

Summary Microclimate and CO₂ exchange of the lichen Usnea sphacelata were measured during summen on a hill near Casey Station, Bailey Peninsula, Wilkes Land, Antarctica. Within a period of 52 days (November 10 until December 31, 1985), 8 diurnal courses of net photosynthesis were measured in naturally snow-covered lichen thalli, and 9 diurnal courses in thalli experimentally sprayed with melt water. Photosynthetic performance of a light-form of Usnea sphacelata was compared with that of a shade-form. Net photosynthesis was reversibly depressed in snow-covered lichen thalli of both forms when irradiance was higher than 600 mol m^–2 s^–1 photosynthetic active radiation (PAR), the depression persisting several hours after a period of strong light. These responses suggest photoinhibition. Models of photosynthesis were established for the light-form by non-linear regressions with field data from water-sprayed thalli (Model W) and field data measured in snow-covered lichens (SNO I, SNO II). Model SNO I is based on median values of photosynthetic rates and SNO II on maximum values for each light/temperature combination. Photosynthetic rates were calculated using model W; the results showed values approximately three times higher than measured in the field with naturally moistened thalli. Photosynthetic rates according to model SNO II fitted the data of naturally moistened lichens measured during the day, before strong light (> 600 mol m^–2s^–1 PAR) caused reversible decrease of net photosynthesis. Model SNO I fitted the data measured during and after a phase of strong irradiance. Model SNO I demonstrated that light stress was highest at temperatures below 2 °C. This study has shown that long-term calculation of the photosynthetic productivity must take into account decreases in net photosynthesis rate caused by strong light, as well as effects of water content and temperature. For the investigated period of the austral summer, a carbon production of 3.44 gm^–2 was estimated for U. sphacelata.     

Light limitations to algal growth in tropical ecosystems

1. Spatial and temporal variations in algal concentrations are controlled in many aquatic ecosystems by the availability of solar irradiance, rather that nutrients or grazing. In such light limiting conditions, changes in the optical or hydrological characteristics of the water column will directly impact biomass concentrations. Here we develop and test an approach based on the relationship between available solar irradiance within the mixed layer and algal biomass concentrations. 2. As with most nutrient/biomass relationships, an increase in available solar energy favours an increase in biomass when light limitation prevails. The ratio between light/biomass is then used to determine a critical light requirement that can be used to estimate critical depth and compensation irradiance and open the way to exploring how changes in mixing depth and vertical attenuation may influence algal biomass concentrations. 3. This approach is used to describe real conditions in two disparate algal communities; the phytoplankton community in Lake Victoria, East Africa and the microphytobenthos community in the lacustrine system of Esteros del Iberá (South America). 4. Differences in the critical light requirement were used to examine the relative efficiency of the algal communities in their use of available solar energy. The tropical phytoplankton community showed similar energetic requirements to theoretical estimates and were found to be less efficient when compared with the phytobenthos community.     

A portable flow chamber for in situ determination of benthic metabolism

1. Many stream ecologists are interested in determining the metabolic rates of benthic organisms, particularly those of production and respiration. It is often necessary to make these measurements on fresh material in the field at remote sites. Recirculating chambers are commonly used for this purpose.
2. A broad variety of recirculating chambers are described in the literature, but each design has inherent limitations. The most common are inability to control flow in the chamber and match it with external flow rates, and a lack of the power required to do this for extended periods. Alteration of spectral irradiance, temperature rise and elevated internal chamber pressures are additional limitations that have received little attention.
3. We have designed and constructed a flow chamber that eliminates some of these problems. The chamber utilizes a DC motor-driven propeller as an efficient recirculator (axial impeller), minimizing power requirements and it is constructed of UVB transparent acrylic to allow a full spectral complement of solar irradiance in the interior. Modular components allow the chamber to be taken apart quickly for cleaning and replacement of parts, making it more functional than some previous designs.
4. The axial impeller chamber was compared to a similar sized conventional chamber that had a small diameter return line and a high capacity centrifugal pump. The axial impeller chamber had less of a temperature rise during field incubations, lower power consumption and less internal pressure in the return line when producing equivalent water velocities.
5. The reported axial impeller design had relatively homogeneous velocity across the working section relative to other chambers and was capable of water velocities in excess of 1 m s^–1.     

Response of Photosynthesis Models to Light Limitation

An investigation into the effect of time step on a common photosynthesis algorithm reveals that the predicted phytoplankton production and biomass depend strongly on the length of the time step. This time step dependence is due to the assumption that a light limitation factor derived from integrating the irradiance over the time step is equivalent to the integrated light limitation factor over the time step. This subtle inaccuracy in defining the factor for light limited phytoplankton production produces a substantial difference in the biomass estimates derived from the two models. To illustrate the difference, the light limitation factor integrated over the time step is implemented in the one dimensional water quality model DYRESM-WQ. The new version of DYRESM-WQ is used to simulate chlorophyll N concentrations in Prospect Reservoir, New South Wales. These results are compared to concentrations predicted using the original algorithm. The comparison shows that the new algorithm for phytoplankton production is relatively insensitive to time step, which decreases the difficulty of calibrating the model for chlorophyll a.     

A model for predicting the emergence of dragonflies in a changing climate

1. Precise models for the phenology of different species are essential for predicting the potential effects of any temporal mismatch of life cycles with environmental parameters under different climate change scenarios. Here we investigated the effects of ambient water temperature on the onset and synchrony of emergence for a widespread European riverine dragonfly, Gomphus vulgatissimus .
2. Long-term field data on the annual emergence from two rivers in northern Germany, and additional data from a laboratory experiment with different temperature regimes, were used to develop a model that predicted the onset of emergence by using mainly the temperature sum (degree days) as a parameter.
3. Model predictions of the onset of emergence fitted the observations well and could be transferred between localities. This was particularly so when weighting early winter temperature data by using a day length and a temperature-response function, implying potential additional control mechanisms for the onset of emergence.
4. We simulated effects of different winter temperature regimes on the emergence curves in order to predict the effects of climate change. These indicated an acceleration of emergence by 6–7 days per 1 °C temperature increase, which is corroborated by the laboratory data and is in the upper range of data published for other dragonflies.     

A fully predictive model for one-dimensional light attenuation by Chlamydomonas reinhardtii in a torus photobioreactor

The light attenuation in a photobioreactor is determined using a fully predictive model. The optical properties were first calculated, using a data bank of the literature, from only the knowledge of pigments content, shape, and size distributions of cultivated cells which are a function of the physiology of the current species. The radiative properties of the biological turbid medium were then deduced using the exact Lorenz-Mie theory. This method is experimentally validated using a large-size integrating sphere photometer. The radiative properties are then used in a rectangular, one-dimensional two-flux model to predict radiant light attenuation in a photobioreactor, considering a quasi-collimated field of irradiance. Combination of this radiative model with the predictive determination of optical properties is finally validated by in situ measurement of attenuation profiles in a torus photobioreactor cultivating the microalgae Chlamydomonas reinhardtii, after a complete and proper characterization of the incident light flux provided by the experimental set-up.     

Light availability as a possible regulator of cyanobacteria species composition in a shallow subtropical lake

1. Variations in the relative biovolumes of dominant cyanobacterial taxa were evaluated in the context of environmental conditions using canonical correlation analysis (CCorrA) and Redundancy Analysis (RDA). The objective was to test a conceptual model in which underwater irradiance determines dominance by bloom-forming (high light adapted) or non-blooming (low light adapted) taxa. 2. The data set consisted of 404 contiguous observations, collected over a 3-year period at eight pelagic sites, in shallow Lake Okeechobee, Florida, U.S.A. Data included species biovolumes, total phosphorus (TP), total nitrogen (TN), dissolved oxygen (DO) and chlorophyll a concentrations, as well as two indices: underwater irradiance (Secchi depth) and the ratio of Secchi:total depth. 3. The first environmental canonical variable was strongly correlated with the two light-related indices, and negatively correlated with TP. This reflects the predominant role of resuspended P-rich lake sediments in controlling underwater irradiance in the shallow lake. The first species canonical variable displayed a strong negative correlation with Lyngbya limnetica and L. contorta, and positive correlations with Anabaena circinalis, Aphanizomenon flos aquae and Microcystis spp. The results support the conceptual model; the first pair of canonical variables explained 55% of the variation in the species–environmental data set. RDA results provided further support for the hypothesis that irradiance was the major force controlling community structure. 4. One unexpected result was a positive association between Oscillatoria spp. dominance and indicators of high irradiance. This conflicts with past research indicating that Oscillatoria is a low light adapted taxon, and the finding that it is the most abundant taxon in Lake Okeechobee. This may reflect the fact that the two Lyngbya taxa were more strongly associated with low light conditions than Oscillatoria. CCorrA results indicated that Oscillatoria densities are strongly controlled by water temperature. There is a need for more detailed studies of cyanobacteria ecophysiology in order to explain fully the seasonality of phytoplankton in this and other shallow subtropical lakes.     

Non-linear effects of drought under shade: reconciling physiological and ecological models in plant communities

The combined effects of shade and drought on plant performance and the implications for species interactions are highly debated in plant ecology. Empirical evidence for positive and negative effects of shade on the performance of plants under dry conditions supports two contrasting theoretical models about the role of shade under dry conditions: the trade-off and the facilitation hypotheses. We performed a meta-analysis of field and greenhouse studies evaluating the effects of drought at two or more irradiance levels on nine response variables describing plant physiological condition, growth, and survival. We explored differences in plant response across plant functional types, ecosystem types and methodological approaches. The data were best fit using quadratic models indicating a humped-back shape response to drought along an irradiance gradient for survival, whole plant biomass, maximum photosynthetic capacity, stomatal conductance and maximal photochemical efficiency. Drought effects were ameliorated at intermediate irradiance, becoming more severe at higher or lower light levels. This general pattern was maintained when controlling for potential variations in the strength of the drought treatment among light levels. Our quantitative meta-analysis indicates that dense shade ameliorates drought especially among drought-intolerant and shade-tolerant species. Wet tropical species showed larger negative effects of drought with increasing irradiance than semiarid and cold temperate species. Non-linear responses to irradiance were stronger under field conditions than under controlled greenhouse conditions. Non-linear responses to drought along the irradiance gradient reconciliate opposing views in plant ecology, indicating that facilitation is more likely within certain range of environmental conditions, fading under deep shade, especially for drought-tolerant species.