DSpace at VNU: Ecophysiological responses of young mangrove species Rhizophora apiculata (Blume) to different chromium contaminated environments

Science of the Total Environment 574 (2017) 369–380 Contents lists available at ScienceDirect Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv Ecophysiological responses of young mangrove species Rhizophora apiculata (Blume) to different chromium contaminated environments Kim Linh Nguyen a,b, Hoang Anh Nguyen b,c, Otto Richter c,⁎, Minh Thinh Pham a, Van Phuoc Nguyen b a b c Nong Lam University, Ho Chi Minh City, Vietnam National University Ho Chi Minh City, Institute for Environment and Resources, Vietnam Braunschweig University of Technology, Institute of Geoecology, Germany H I G H L I G H T S G R A P H I C A L A B S T R A C T • Young mangrove species Rhizophora apiculata was cultivated in an artificial wetland simulating tidal inundation • Response of R apiculata to chromium contamination strongly depends on soil/water environments and fertilizer • The addition of nutrients amplifies the toxic effect of chromium • The response is modeled by a nonlinear multivariate growth model a r t i c l e i n f o Article history: Received 24 June 2016 Received in revised form September 2016 Accepted September 2016 Available online 14 October 2016 Edited by: F.M Tack Keywords: Mangrove Rhizophora apiculata Chromium Nutrient Microorganism Plant growth modelling ⁎ Corresponding author E-mail address: o.richter@tu-bs.de (O Richter) http://dx.doi.org/10.1016/j.scitotenv.2016.09.063 0048-9697/© 2016 Elsevier B.V All rights reserved a b s t r a c t Many mangrove forests have suffered from the contaminated environments near industrial areas This study addresses the question how these environments influence the renewal of mangrove forests To this end ecophysiological responses of the young mangrove species Rhizophora apiculata (Blume) grown under combinations of the factors heavy metals (here chromium), nutrition and soil/water environment were analyzed We tested the hypothesis that soil/water conditions and nutrient status of the soil strongly influence the toxic effect of chromium Seedlings of R apiculata were grown in three different soil/water environments (natural saline soil with brackish water, salt-leached soil with fresh water and salt-leached-sterilized soil with fresh water) treated with different levels of chromium and NPK fertilizer The system was inundated twice a day as similar to natural tidal condition in the mangrove wetland in the south of Vietnam The experiments were carried out for months Growth data of root, leaf and stem, root cell number and stomata number were recorded and analyzed Results showed that growth of R apiculata is slower in natural saline soil/water condition The effect of chromium and of nutrients respectively depends on the soil/water condition Under high concentrations of chromium, NPK fertilizer amplifies the toxic effect of chromium Stomata density increases under chromium stress and is largest under the combination of chromium and salty soil/water condition From the data a nonlinear multivariate regression model was derived capturing the toxicity threshold of R apiculata under different treatment combinations © 2016 Elsevier B.V All rights reserved 370 K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 Introduction Mangrove wetlands are highly efficient in adsorbing and absorbing wastewater-borne pollutants Mangrove trees have inherent physical, chemical and biological properties for adsorption and/or utilization of nutrients and heavy metals (Zhang et al., 2007; Ke and Tam, 2012; Akhand et al., 2011; Usman et al., 2013; Naidoo et al., 2014) The ability of heavy metal accumulation of mangroves is well established for example in Rhizophora sp and Avicennia marina with the order Cu N Zn N Cr N Ni N Cd N Pb (Akhand et al., 2011; Nazli and Hashim, 2010) Thus, it was proposed to use mangrove vegetation in wastewater treatment systems (Sodré et al., 2013; Tansel et al., 2013; Ouyang and Guo, 2016) On the other hand, mangrove forests provide many ecosystem services such as flood control, shoreline stabilization and storm protection, sediment and nutrient retention and are hot spots of biodiversity in tropical and subtropical regions Their ecological and socioeconomic significance has been recognized in many studies (Acharya, 2002; Lee et al., 2014) Therefore, protection and conservation of mangrove ecosystems are important issues Thus, to balance between the needs to use mangrove vegetation for filtering pollutants and the necessity to protect mangrove ecosystems, the tolerance thresholds of mangrove vegetation to the concentrations of pollutants need to be investigated This motivated our investigation on the physiological responses of mangroves under elevated contamination levels of pollutants Growth responses of some mangrove species under the effects of salinity (Takemura et al., 2000), nutrition and light intensity have been discussed (Ye et al., 2005; Chen and Ye, 2014; Dangremond et al., 2015; Alongi, 2011) The significance of trace metals on plant physiological processes, such as photosynthetic performance and salt secretion, was stated (Naidoo et al., 2014; MacFarlane and Burchett, 2002) as well as the studies on metal accumulation in roots, stems and leaves of mangroves (Lewis et al., 2011; Keshavarz et al., 2012; Mahdavi et al., 2012) It has been addressed that, the uptake by the roots of mangrove trees and the uptake by the benthos and microorganisms are the main processes causing the retention and degradation of substances (Hawkins et al., 1998; Yim and Tam, 1999), while the bioavailability of substances depends on multiple factors, such as salinity (Yim and Tam, 1999; Ye et al., 2005; Jayatissa et al., 2008), temperature, redox potential (Lewis et al., 2011) and nutrient supply (Alongi, 2011) However, studies on the responses of mangroves to the combinations of stress factors and the toxicity thresholds of mangrove species to heavy metals contaminated environments are still limited In our study, we investigated the hypothesis that the combined effects of soil type, plant nutrition and contamination by heavy metals (here chromium) generate complex response patterns both of plant growth and plant physiology To this end, an experimental system in an artificial mangrove wetland was constructed in the LongThanh mangrove forest (106.9358° N, 10.6475° E), located in the south of Vietnam LongThanh mangrove forest is situated in the vicinity of a large industrial area principally involved in plating, alloying, tanning, textile dyes and other activities The area is moderately polluted by organic substances and heavy metals, especially chromium, which accumulated in the sediments and in the vegetation (Nguyen et al., 2014) The mangrove species Rhizophora apiculata Blume which is a mangrove tree species of tropical coastal environments, propagates through viviparous propagules, is the dominant species in this forest We used chromium in our experiments Although chromium is not an essential element, it can be taken up by the plants along with essential elements like iron It has been documented that an excess level of chromium will exert strong adverse effects on plant growth leading to a decrease in productivity and ultimately to death (Oliveira, 2012) The paper focuses on the assessment of the chromium toxic tolerance of young R apiculata through plant physiological responses A three factorial experiment with the factors chromium load, soil nutrition and soil type was conducted For each factor combination, the response of the trees was observed during months monitoring the variables root length, root number, root cell density, tree diameter, tree height, leaf number, leaf area and stomata density Since the characteristics of soil and the associated diversity of microbial communities are important components in the retention and transformation of substances in plants, the soil factor comprises natural saline soil, saltleached soil and salt-leached-sterilized soil To quantify the toxic effects, a mathematical growth model was then developed for simulating the growth pattern of the young mangrove R apiculata in dependence of chromium and nutrients in different soil/water environments Materials and methods 2.1 Design of the artificial mangrove wetland 2.1.1 Experimental set-up The experiments were carried out in a greenhouse containing ten aquaria and ten reservoirs each with a volume of 750 L Each pair of reservoir-aquarium was used for one treatment combination Water was pumped from the reservoir into each aquarium by a submersible pump at a rate of 750 L/h lasting for h A valve at the bottom of each aquarium allowed for a slow discharge of the water from the aquarium back to the reservoir (1 h) Pots with young mangrove plants were placed into aquaria which were kept flooded twice a day simulating the conditions in a tidal zone Water infiltrated the pots via the surface and by holes The experimental-setup simulated the semi-diurnal tidal condition in the area The spatial dimensions of the set-up are given in Fig 2.1.2 Plant culture Young mangrove species R apiculata were collected from the LongThanh mangrove forest with heights between 35 and 40 cm and leaf numbers between and They were cultivated individually in 40 cm × 40 cm (diameter × depth) plastic pots with drainage holes, each pot contained kg of soil Thirty pots were placed in each aquarium All plants were grown under natural light in the air-conditioned glasshouse at ~35 °C (cf Fig 1) 2.1.3 Experimental design A three factorial trial was performed with the factors “Soil/Water”, “Fertilizer” and “Chromium Load” The soil/water factor comprised i) Brackish water + natural saline sediment (NN) ii) Fresh water + salt-leached sediment (WN) iii) Fresh water + salt-leached and sterilized sediment (AN) Natural saline soil (NN) was taken from the LongThanh mangrove forest For the production of salt-leached soil (WN), the natural saline soil (NN) was immersed in fresh water, was stirred and replaced by new fresh water every day until soil salinity was zero Basic soil characteristics were analyzed, at the beginning of the study, to obtain some basic descriptive data of the soils in the forest Soil samples were homogenized in an agate mortar and air dried at room temperature A sample of 0.5 g of the oven dried homogenized soil samples were directly digested with aqua regia according to DIN EN 13346 (2000) The concentrations of elements in the aqua regia extracts were determined with an ICP OES Total nitrogen was analyzed using an elemental analyzer LECO TruSpec CHN Macro The standard deviation, determined by multiple measurements for all elements b2% Basic soil properties and concentrations of some heavy metals in the natural saline soil and salt-leached soil are shown in Table K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 371 Fig The experimental system: twice a day the contaminated water was pumped from the reservoirs to the aquaria The dimension unit of the system is in cm Salt-leached-sterilized soil (AN) was obtained from salt-leached soil by sterilization in the oven at 121 °C, atm for 10 h Brackish water used for the experiment with natural saline soil (NN) was pumped from the DongKho river nearby, 50 m away from the greenhouse Fresh water was pumped from a groundwater well in the area and was used for the experiment with salt-leached soil and saltleached-sterilized soil The fertilizer factor comprised two levels zero and 15 g The fertilizer consisted of ammonium, phosphorus and potassium with mass fractions 16, 16 and respectively (NPK 16-16-8) Fertilizer level 15 g was applied three times in three months, g per pot each time, the first time at day 10 of the experiment Chromium levels were mg/L, 500 mg/L and 1000 mg/L of Cr (III)+ Chromium was added to the water in the form of chromium sulphate Cr2(SO4)3 Each treatment was arranged as randomized complete design (RCD) with 24 replications for non-destructive samples and three replications for destructive samples The experimental design is shown in Table Table Soil characteristics at the beginning of the experiments Parameter Unit Natural saline Salt-leached soil (NN) soil (WN) Pore water salinity Sand Silt Clay Total nitrogen Total Phosphorous Total Kali Total Na As Pb Cd Cu Zn Cr PSU 15.1 0.0761 0.0147 24.61 32.27 43.12 0.0536 0.00755 % % % % % % 0.53 % 0.39 mg/kg mg/kg mg/kg mg/kg mg/kg mg/kg 22.8 0.49 0.69 6.07 12.7 b0.0167 29.4 57.9 15.6 Threshold values Notes Practical salinity unit 12 120 70 200 Sediment quality guidelines of Vietnam (QCVN 03: 2008) 2.2 Morphometric measurements Stem height, stem diameter and leaf number were measured monthly for each plant Plant mass, leaf area, root cell number, stomata number, root length and root number were recorded after months and months respectively Plants were removed from their pots to ensure roots were intact and salt-leached to remove soil and other debris Stems, roots and leaves were blotted dry and weighed fresh Plant parts were then placed in labelled paper bags and dried at 60 °C for days, after which they were dry weighed Biomass data of this study (not shown here) were used to validate a comprehensive growth model in terms of a system of differential equations (Richter et al., 2016) Root cell density were determined according to Glime and Wagner (2013), the main roots were sliced at the tips and immersed in Javel water (sodium hypochlorite NaOCl) for 15 min, then rinsed with distilled water After that the samples were dipped into acetic acid for min, rinsed again with distilled water and then dried by filter paper These samples were then colored and cut into 1-mm2-slices to count the number of cells using a microscope For the determination of stomata number, the second pair of leaves from each plant shoot was chosen The abaxial epidermis of the leaf was cleaned first using a degreased cotton ball, and then carefully smeared with collodion for approximately 15 The thin film was peeled off from the leaf surface Numbers of stomata for each film unit (1 mm2) were counted under a microscope 2.3 Statistical analysis and regression model The data were analyzed by a factorial analysis of variance appropriate to the experimental design The statistical model implies the factors soil (S), nutrient (N) and chromium (Cr) In addition to the three experimental factors, the factor time was taken into consideration as a fourth factor The interactions of the factor time with the other factors allow to draw conclusions on the time profiles of the dependent variable, e.g., if there is a significant interaction between the factors chromium and time, then the shapes of the curves are different yijklm ẳ ỵ Si ỵ Nu j ỵ Crk ỵ Tl ỵ ẵS Nuij ỵ ẵS Crik ỵ ẵS Til ỵ ẵNu Crjk ỵ ẵNu Tjl ỵ ẵCr Tkl ỵ ẵCr Nu S ỵ ẵNu S T ỵ ẵCr S T ỵ ẵCr Nu T ỵ ẵCrNu S T ỵ ijklm 372 K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 Table Experimental design Soil level ID Water and soil AN0-0 AN15-0 AN0-500 AN15-500 AN0-1000 AN15-1000 WN0-0 WN15-0 WN0-500 WN15-500 WN0-1000 WN15-1000 NN0-0 NN15-0 NN0-500 NN15-500 NN0-1000 NN15-1000 Fresh water + wash-sterilized sediment Nutrient level (NPK 16-16-8) Chromium (Cr III+) level (0 mg/L) (0 g/pot) (500 mg/L) (15 g/pot) (1000 mg/L) Fresh water + washed sediment (0 mg/L) (0 g/pot) (500 mg/L) (15 g/pot) (1000 mg/L) Brackish water + natural sediment i ¼ 1; 2; j ¼ 1; k ¼ 1; 2; l ¼ 1; 2; 3; 4; 5; m ¼ 1; 2; :::; n Post hoc tests were performed according to the method of Bonferroni, i.e adjusting the significance level with respect to the number of multiple comparisons Furthermore, multivariate nonlinear regression models for the growth of the variables height, diameter and leaf number were derived For the growth of the young trees an exponential growth law is appropriate, where the growth rate depends on the fertilizer level Nu and chromium load Cr The models were applied to each soil type separately since soil type is a qualitative factor The nonlinear multivariate regression problem was solved by the NonLinearFit procedure in Mathematica® The detailed form of two models and the model fits are presented in section Model comparisons were performed based on the Akaike information criterion AIC (Akaike, 1974) Results I: statistical data analysis Detailed ANOVA tables are given in the appendix A An overview of all ANOVA results is presented in Table 3.1 Root growth Table shows mean values and standard deviation of the root length and root number of young R apiculata for each treatment combination The data shows the negative effect of the chromium treatment on root elongation, this value is lower than that of treatments with no chromium input for all soil types Among different soils, at chromium level (no chromium) root length has the order AN N WN N NN However, when compare between the treatments with chromium, root length has the order WN N NN N AN With the addition of nutrient, root length of soil WN at chromium level (500 mg/L) is the highest among the three soils, the lowest value happened with soil AN at the third month but with soil NN at the sixth month Root number is highest at treatment WN in all nutrient levels and Cr levels The lowest values occurred with the treatment AN at month and with treatment NN at month for all nutrient levels These features are corroborated by the results of the ANOVA: The table for the root length (cf Table A.1, Appendix A) shows significant main effects of the factors soil, chromium and time (p value b 10−3) There is no significant nutrient effect The interactions between the (0 mg/L) (0 g/pot) (500 mg/L) (15 g/pot) (1000 mg/L) factors chromium and soil (Cr S), soil and time (S T), nutrient and time (Nu T) and chromium and time (Cr T) are significant (p value b 0.05) The three-way interaction between the factors chromium, nutrient and soil (Cr Nu S) is highly significant (p value b 10−3), showing that the combined effect of chromium and nutrients depends on the soil type For the root numbers (cf Table A.2, appendix A), we find significant main effects of the factors soil, nutrient and time The interactions between chromium and soil (Cr S), soil and time (S T), nutrient and time (Nu T) are significant 3.2 Stem growth Fig 2a–c shows the box plots of stem height for each soil type under all combinations of the chromium and nutrient factors Stem growth is highest at treatment WN and lowest at treatment AN The growth curves show different forms under the influence of the chromium loads With no chromium, the height increases over the whole time period The addition of nutrient enhances tree growth in the WN and NN soils Surprisingly, in soil AN (sterilized soil) the addition of nutrients does not have positive effect and even leads to a diminished growth in the case of high chromium levels The data of tree diameter shows similar result The ANOVA tables for plant height and plant diameter (Tables A.3 and A.4, Appendix A) show highly significant main effects of all factors Second order interactions are also highly significant with the exception of the soil-time interaction (S T) at the height table For both variables, diameter and height, the three-way interaction (Cr Nu S) is highly significant This means that the combined effect of chromium and nutrient is differing with respect to soil type as seen in Fig 3.3 Leaf growth Fig 3a–c shows the development of leaf number in each soil under all combinations of chromium and nutrient The highest value of leaf number occurred at treatment WN and the lowest occurred at the treatment NN At the beginning of the experiment, each plant had leaves During the first months, plants in all treatments formed from to new leaves per month After to months the effect of chromium is noticeable Leaves were severely damaged at the high chromium concentrations during the last two months of the experiment The results of the ANOVA given in Table A.5 in appendix A confirm these complex patterns All main effects and all second and third order interactions are highly significant (p values b 0.05) The different shapes of the time curves seen in Fig are reflected by the significant interaction terms in the ANOVA table Furthermore, the soil specific K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 373 Table Mean values and standard deviation of physiological data of young R apiculata for each treatment combination AN: Fresh water + salt-leached and sterilized sediment, WN: Fresh water + salt-leached sediment, NN: Brackish water + natural saline sediment Root cell number (cell/mm2) Stomata number (s/mm2) Root length (cm) Root number (per plant) Leaf area (cm2) Soil type Nutrient (mg) Concentration of Cr(III+) (mg/L) Three months Six months Three months Six months Three months Six months Three months Six months Three months Six months AN 0 252.67 ± 363.33 ± 41.67 ± 42.33 ± 22.83 ± 39.93 ± 4±0 5.66 ± 474.33 ± 959.33 ± AN 15 54.012 306.33 ± 17.926 411.33 ± 3.51 40.33 ± 3.21 47.33 ± 2.93 29.83 ± 4.81 39 ± 4.33 ± 1.53 7.33 ± 21.78 518.67 ± 213.03 1131.70 ± AN 500 13.796 227.67 ± 14.978 320.33 ± 6.66 52.67 ± 4.16 59.33 ± 1.07 23 ± 34 ± 0.58 4±1 1.53 5.33 ± 107.36 293.1 436 ± 24.98 772 ± AN 15 500 22.679 263.33 ± 18.903 347.33 ± 3.06 53.67 ± 3.79 59.67 ± 17.3 ± 18.33 ± 4±1 1.53 5.33 ± 484.67 ± 154.07 924.33 ± AN 1000 36.226 182 ± 44.23 201 ± 4.04 62.33 ± 2.52 65.33 ± 2.07 11.53 ± 2.08 19.36 ± 2.66 ± 2.08 5±1 78.01 384.33 ± 148.18 409.67 ± AN 15 1000 23.302 252.67 ± 31.796 5.13 178 ± 67.29 60.33 ± 3.51 66.00 ± 2.72 16.06 ± 2.05 26.36 ± 0.58 3±1 6±1 32.624 425 ± 63 41.02 326.33 ± WN 0 43.247 262.33 ± 377.33 ± 2.52 45.33 ± 46.67 ± 2.18 21.46 ± 0.80 42.5 ± 5.33 ± 5±1 544.33 ± 93.18 1073.3 ± WN 15 37.899 358 ± 55.194 489 ± 7.77 47.67 ± 5.51 50.33 ± 1.11 22.03 ± 4.44 30 ± 1.15 5.33 ± 4.33 ± 57.59 106.93 616 ± 83.21 1353.3 ± WN 500 29.866 256 ± 18.358 336 ± 3.51 55.00 ± 3.51 55.67 ± 5.32 17.16 ± 24.33 ± 0.58 5.33 ± 1.15 5±1 175.1 486 ± 31.05 831.67 ± WN 15 500 28.844 281.67 ± 33.779 373 ± 7.55 57.67 ± 2.52 58.00 ± 3.06 20.16 ± 2.52 30.16 ± 1.15 4.67 ± ± 2.65 592.67 ± 77.84 977.67 ± WN 1000 31.533 226.33 ± 39.661 267.33 ± 4.51 66.33 ± 4.58 68.33 ± 4.75 26 ± 2.65 8.69 34 ± 0.58 4.67 ± 5.67 ± 37.42 473.67 ± 65.59 521.67 ± WN 15 1000 29.4 265.67 ± 55.537 261.67 ± 2.08 65.00 ± 3.51 66.67 ± 18.83 ± 7.69 24 ± 3.5 0.58 4.67 ± 1.15 7.67 ± 94.31 543.67 ± 80.63 517.33 ± NN 0 43.097 221.33 ± 49.943 324.33 ± 4.36 53.67 ± 4.04 58.00 ± 6.25 18.83 ± 22.67 ± 0.58 3.67 ± 0.58 6.33 ± 54.23 421.33 ± 26.41 760 ± NN 15 84.69 282.67 ± 30.567 385.33 ± 4.04 52.00 ± 54.67 ± 3.82 20.83 ± 3.01 28.5 ± 0.58 3.33 ± 0.58 7±1 23.86 477.67 ± 121.06 883.67 ± NN 500 24.786 209.33 ± 44.602 291 ± 3.61 69.33 ± 4.51 75.00 ± 7.37 15.67 ± 3.77 21.5 ± 0.58 3.33 ± 5±1 51.25 45.63 394 ± 28.83 643.33 ± NN 15 500 32.146 242 ± 38.974 330.33 ± 6.03 70.33 ± 3.61 72.67 ± 3.06 17.17 ± 3.5 21.67 ± 0.58 3.33 ± 5.67 ± 450.67 ± 86.22 794 ± 78.26 NN 1000 41.581 122.67 ± 27.538 188 ± 3.21 71.33 ± 5.03 72.00 ± 1.90 16.97 ± 3.51 20 ± 0.58 3±0 1.53 4.33 ± 41.86 352.67 ± 393.33 ± 1000 25.502 198 ± 21.794 146 ± 28 2.08 73.00 ± 2 76.33 ± 4.10 16.97 ± 2.18 14.83 ± 3.67 ± 1.15 4.67 ± 50.94 355.33 ± 40.27 274.33 ± 3.06 4.13 2.36 0.58 0.58 75.79 86.0 NN 15 28.618 effect of the nutrients is reflected by the significant two-way interaction between the soil and nutrient factors (Nu S), the three-way interaction (Cr Nu S) and the three-way interaction (Cr Nu T) The latter significant interaction confirms that the shapes of the leaf number depend on the combinations of Cr and nutrient For leaf area (cf Table 3), a similar response pattern was observed, the highest value occurred at the treatment WN and the lowest one occurred at the treatment NN At chromium level 3, the decrease in leaf area is amplified by the nutrient at the treatments AN and NN The results of the ANOVA given in Table A.6 in Appendix A confirm these complex patterns Almost all main effects and all second order interactions are highly significant, except the interaction effects of nutrient and soil (Nu S), chromium and soil (Cr S) and nutrient and time (Nu T) It shows that these interaction effects are not significant for the development of leaf area Similar to leaf number, the three-way interaction chromium – nutrient – time (Cr Nu T) is significant Table Summary of the results of the ANOVAs: **: p b 10−3, *: p b 0.05, –: p ≥ 0.05 Factor Height Diameter Number of leaves Leaf area Stomata density Root cell density Root length Root number Soil Nutrient Chromium Time Nutrient-soil interaction Chromium-soil interaction Soil-time interaction Chromium-nutrient interaction Nutrient-time interaction Chromium-time interaction Chromium-nutrient-soil interaction Nutrient-soil-time interaction Chromium-soil-time interaction Chromium-nutrient-time interaction ** ** ** ** ** ** – ** ** ** ** – – – ** ** ** ** ** ** ** ** ** ** ** – – ** ** ** ** ** ** ** ** ** ** ** * * ** ** ** ** ** ** – – * * – ** – – – * ** – ** ** – ** – – – – – – – – ** ** ** ** – – – * – ** – – – * ** – ** ** – ** * – * * ** – – – ** * – ** – * * – * – – – ** – 374 K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 Fig a–c The development of stem height (in cm) of R apiculata when exposed to different concentrations of Cr and nutrient during months in three different soils (a: AN, b: WN, c: NN) In all soils tree growth is inhibited by chromium The addition of nutrient enhances the tree growth in the WN and NN soils but give no effect to AN soil 3.4 Cell number and stomata number The response pattern of cell number and stomata number shows interesting features Root cell density (cf Table 3) was increasing from month to month in the presence of nutrient in almost all treatments except in the experiments with highest Cr load (Cr level 3), where root cell numbers decreased (in the treatment AN) or were retarded (treatments WN and NN) from month to month For all three soil types, maximum values occurred in soil type WN (salt-leached soil) and minimum values occurred in soil type NN (natural saline soil) The results of K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 375 Fig a–c The development in time of leaf number of young R apiculata exposed to different concentrations of Cr and nutrient during months in three different soils (a: AN, b: WN, c: NN) In all three soils, leaf numbers decrease after about months under the highest chromium level In the sterilized soil (AN) and in the natural saline soil (NN), there is a strong synergistic effect of the nutrient at the maximum chromium load the ANOVA (cf Table A.7, Appendix A) show significant main effects for all factors In addition, the two-way interactions between chromium and nutrient (Cr Nu), chromium and time (Cr T) and the tree-way interaction chromium-nutrient-time (Cr Nu T) are significant In contrast to the retarded trend of root cell number, stomata density increases with increasing chromium load (cf Table and Fig 4) After three months, the minimum value of stomata density occurred at the treatment AN but after six months the lowest value occurred at the treatment WN The density of stomata is highest at the treatment NN in both and months periods The effect of nutrient on stomata density is not clear The ANOVA table (Table A.8, Appendix A) for the stomata density shows significant main effects of the factors soil, chromium and time 376 K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 Fig Stomata density at month and at month in dependence on chromium load and nutrient Note that stomata density increases with increasing chromium load There are no significant interactions except a significant interaction between chromium and soil (Cr S) There are no significant effects of the factor nutrient Result II: nonlinear regression model Preliminary remark: the model development was motivated by the exploratory data analysis and the results of the ANOVAs Model selection was guided by the use of the Akaike information criterion (Akaike, 1974) For the response variables plant height, plant diameter and leaf number the ANOVAs show significant interactions between the factors chromium and nutrient, between nutrient and time and between chromium and time Since the factor levels of nutrient and chromium are quantitative, it is challenging to capture this response pattern by nonlinear regression models Here, we have to distinguish between the variables plant height and diameter on the one hand and on leaf number and leaf area on the other, since the former increases with time whereas the latter may decrease with time after a period of increasing under high chromium concentrations Therefore, two models were developed, which are both based on an exponential growth law (Eq (1)) The models are specified by the form of the dependence of the growth rate μ on the chromium (Cr) and nutrient (Nu) levels and on time f Nu; Cr; t ị ẳ a expẵ Nu; Cr; t ị t and 1000 mg/L and the variable Nu codes the amounts of nutrient and 15 g per pot respectively If all factor levels take the value of (no chromium, no fertilizer), the growth rate is equal to k0, the natural growth rate The parameters k1, k2 and k3 can take on negative or positive values according to the influence of the factor on the growth The first and the second term describe the effect of chromium and nutrient respectively and the third term takes into account the interaction between chromium and nutrient As we have seen in the foregoing section, the interaction might be positive or negative This model fits satisfactorily to the growth data as shown in Fig for the tree diameter To demonstrate the negative interaction between chromium and nutrient, we have chosen the dataset obtained for the soil AN Under high chromium level, the growth of the diameter in the fertilized plants is lower than the growth without fertilizer (Fig 6a) If no chromium is given, tree growth is stimulated by the fertilizer as has to be expected (Fig 6b) Application of the model to the height data gives similar results ð1Þ In the case of plant height and diameter, the growth rate μ(Nu,Cr) is made dependent on nutrient level Nu and chromium level Cr by a multiple linear approach with an interaction between nutrient and chromium factor (Eq (2)) Nu; Crị ẳ k0 ỵ k1 Nu1ị þ k2 ðCr−1Þ þ k3 ðCr−1ÞðNu−1Þ ð2Þ Note that there is no dependence on time of the growth rate The three levels of the variable Cr code chromium concentrations of 0, 500 Fig Increment of tree diameter at different chromium levels for nutrient level (no nutrient added) and soil (AN) The points are data; the response surface is generated by the model defined by Eqs (1) and (2) with the parameters given in Table B.1, Appendix B K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 Because of the complex response pattern of the leaf number, a more complicated form of the growth rate had to be devised The growth curves observed under the different chromium levels (cf Fig 3) run parallel for a certain time and separate after about months Under the highest chromium level, the mean leaf number decreases after attaining a maximum value This behavior cannot be described by a growth rate independent of time, because the sign of the growth rate μ(Nu, Cr)determines the growth behavior from the beginning There is either an increase or a decrease Therefore, a time dependent growth rate is introduced based on the assumption that the buildup of toxic concentrations within the trees takes a certain time This explains why in the first three months the curves run almost parallel This effect is captured by the first factor of the third term of Eq (3) Toxic effects develop only if a critical time tc is approached and surpassed Furthermore, a nonlinear response of the growth rate to chromium was introduced taking into account a threshold effect, i.e chromium toxicity is effective 377 only if a certain threshold level thr is surpassed (Eq (3)) This effect is captured by the second factor of the third term of Eq (3) "   #! "   # t SðCr; NuÞ ð1‐Exp Nu; Cr; tị ẳ k0 ỵ k1 Nu1ị 1Exp tc thr 3ị with SCr; Nuị ẳ Cr1ị þ k2 ðCr−1ÞðNu−1Þ ð4Þ To allow for an interaction between chromium and nutrient level, the term k2 (Cr − 1) (Nu − 1) was added to the chromium level Eqs (3) and (4) imply the following effects Fig (a) Under high chromium level, fertilizing has a negative effect on the growth (b) If no chromium is given, tree growth is enhanced by adding fertilizer 378 K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 i) If chromium level equals 1, i.e no chromium added to the water, and nutrient level equals 1, i.e no nutrient given, the growth rate reduces to the natural saline soil specific growth rate k0, which increases to k0 + k2(Nu − 1) if nutrient is added ii) If chromium is added to the water, plant growth is affected only if the chromium level approaches a threshold value thr iii) Toxic effects slowly build up iv) Toxic effects are amplified if a nutrient is added Fig shows the response of leaf number under different combinations of Cr and nutrients on the three soils Discussion Considering the diversity of the detailed analyses given above, it might be helpful to summarize the results to get the overall picture We have therefore devised the following table, where the outcomes of the ANOVA of each variable are put together 5.1 Response of R apiculata to different soil types There is a highly significant effect of soil type (S) on all variables Post hoc tests (Bonferroni) show that soil WN (salt-leached soil) has the highest rates and the soil NN (natural saline soil) has the lowest rates Fig Time course of leaf numbers for soil types Blue data points: no fertilizer, red data points: addition of 15 g fertilizer per pot The response surfaces are generated by the model defined by Eqs (1), (3) and (4) Note that leaf numbers decrease after months under the highest chromium load (level 3) Fertilizing (red data points) amplifies the toxic effect of chromium Parameters for these models are given in Tables B.7–B.9 in Appendix B K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 for the parameters diameter, height, number of leaves, leaf area, root cell density, root length, root number In contrast, stomata density is highest in the soil NN (natural saline soil) Clearly, salinity in the natural saline soil (NN) can be considered as an additional toxic effect influencing uptake of nutrients and thus plant growth This might be due to competitive inhibition of ammonium uptake by sodium, as discussed by Naidoo (1987) Reef et al (2010) discussed nutrient availability of mangroves by the redox state of the soil, which may be influenced by salinity This explains why we found that growth of R apiculata is fast in the soil salt-leached with fresh water (WN) and low in the natural saline soil (NN) A limitation of microorganism population in the soil (as in the treatment AN) is a disadvantage for the growth and chromium tolerance of R apiculata, since microorganisms stimulate the development of the root system and are capable of rendering chromium species inactive in the soil, many microbes were reported to reduce Cr under aerobic and anaerobic conditions (Jeyasingh and Philip, 2005; Shukla et al., 2009) In our experiments, growth variables in the soil AN are smaller than those in the soil WN since it lacked of a population of chromium retention microorganisms 5.2 Response of R apiculata to different levels of nutrient and chromium The nutrient factor (Nu) has a highly significant effect on all variables except root length and stomata number Chromium has a significant effect on all variables except the root number Post hoc tests reveal negative effects on root length, shoot cell number, leaf number, leaf area, height and diameter However, stomata density is highest under the highest chromium load These results are in agreement with the findings of other authors Root length is more affected by chromium than by other heavy metals (Shanker et al., 2005) Results from Panda and Patra (2000) showed that at a chromium concentration of μM, the root length of seedlings of wheat (cv Sonalika) cultivated on different N fertilizer levels increased However, under higher chromium concentrations, root length decreased regardless of the N fertilizer level Reduction in root growth caused by chromium toxicity could be due to the inhibition of root cell division, root elongation or to the extension of cell cycle in the roots (Shanker et al., 2005) Shanker et al (2005) showed that the reduction of plant height in chromium stress conditions is mainly due to the reduction in root growth leading to lesser nutrients and water transport to the above parts of the plant In addition, chromium transport to the aerial parts of the plant can have a direct impact on the cellular metabolism of shoots contributing to the retardation of plant height Following Ghani (2011), typical symptoms observed in chromium intoxication are chlorosis, necrosis and red brownish discoloration of the leaves In our experiments, beside typical symptom reported, wilting of leaves at the top occurred and ring spots appeared on leaf surfaces of R apiculata These symptoms are similar to those in iron deficiency It could be that the more chromium plants accumulated the less iron they uptake as these two elements compete in binding with the same acceptor at the root cell membrane The increment of stomata density as response to heavy-metal toxic stress has been reported in the literature for other plants It has been found that stomata density in tobacco leaves increased with elevated chromium intoxication (Orcen et al., 2013) Stomata are by far the most influential components in gas exchange The regulation of wateruse efficiency in plants occurs, in part, by changes in stomatal density Stomatal density is regulated by regulators Factors that increase stomatal density might be working through the positive regulators (basic helix-loop-helix; bHLH SPCH, MUTE and FAMA) By contrast, those that decrease the number of stomata might be working through the negative ones (SDD1, TMM, ER-family, YDA and the mitogen-activated protein kinase MAPK module) To cope with stress, plant produces more enzymes to catalyze the process of phosphorylation, enhancing 379 photosynthesis to supply energy for plant Thus the processes of gas exchange and evaporation increase In this case, plant has to produce more stomata 5.3 Response of R apiculata to interactions of environmental factors The second and third order interactions of the factors Cr and nutrient (Cr Nu), Cr and time (Cr T) and Cr, nutrient and time (Cr Nu T) are highly significant for the variable root cell density This means the effects of chromium and nutrient are dependent on each other and also dependent on time Cell number decreases with the presence of Cr In the treatments with Cr loads, cell number is higher when adding nutrient Root cell development at the tips of the main root (root tip meristem) represents the elongation of the root This value is also an indicator for the growth of the whole plant, when the number of cells in this region reduced, root growth is inhibited If this situation lasts long, roots will be weakened This will affect the uptake of water and nutrient to the plant For stem growth, at the highest chromium load of 1000 mg/L (Cr level 3), growth stops after to months With the exception of soil AN, the addition of nutrients mitigates the chromium's growth inhibition which is manifested by the significant interaction between the nutrient and chromium factor (Nu Cr) The interactions between the chromium factor and the time factor (Cr T) refer to the different shapes of the growth curves in dependence of chromium level as shown in the regression model (Eq (2)) and in Fig The interaction between the nutrient and the time factor (Nu T) has the same interpretation For leaf growth, as observed from ANOVA table and the regression model (Eq (3)), the response pattern of the leaf number to the treatments is most complex and shows one surprising result: the amplification of the inhibitory effect of chromium by the addition of nutrient The most striking feature is that in the soils AN and NN the decrease in leaf number under high chromium levels is enhanced by the nutrient, whereas under no chromium, the development of leaf number is promoted by the nutrient All these effects are shown in Fig 7, where the response surfaces for the two nutrient levels are superimposed It is clearly that the toxic effect of chromium on the leaf number is amplified under the addition of nutrients (red data points) One also recognizes the nonlinear effect of chromium The growth curves not differ much under chromium levels (no chromium) and (500 mg/L) At chromium level (1000 mg/L) leaf numbers are drastically reduced after about months The effect of chromium on plant growth and the interaction between chromium and nutrient uptake have been well documented in the literature According to Oliveira (2012), chromium, being structurally similar to other essential elements, may affect plant mineral nutrition Uptake of both macro nutrient (N, P, K) and micro nutrient decreased with increasing of chromium concentration in irrigation of paddy (Sundaramoorthy et al., 2010) High concentration of chromium may displace the nutrients from physiological binding sites and consequently decrease in uptake and translocation of essential elements (Oliveira, 2012) Turner and Rust (1971) found that nutrient solution with 9.6 μM Cr decreased the uptake of K, Mg, P, Fe and Mn in root of soybean Barcelo et al (1985) described the inhibition of P, K, Zn, Cu, Fe translocation within the plant parts when bean plants were exposed to Cr Excess chromium interfered with the uptake of Fe, Mo, P, and N and effected the translocation of P, S, N, Zn and Cu from roots to tops (Chatterjee and Chatterjee, 2000) The reduction in N, P, K and other elements could be due to the reduced root growth and impaired penetration of roots into the soil due to chromium toxicity (Shanker et al., 2005) On the other hand, chromium stress may affect the photosynthesis in terms of CO2 fixation, electron transport, photophosphorylation and enzyme activity, leading to a decrease in productivity and ultimately to death Disorganization of the chloroplast ultrastructure and the inhibition of electron transport processes due to chromium and a diversion of electrons from the electron-donating side of Photosystem 380 K.L Nguyen et al / Science of the Total Environment 574 (2017) 369–380 I (PSI) to chromium is a possible explanation for chromium-induced decrease in photosynthetic rate Besides, the alteration of photosynthetic pigments by chromium may be a further explanation of the inhibition of photosynthesis under chromium toxic stress (Shanker et al., 2005) Conclusion The results of the ANOVA and of the regression model allowed us to identify toxicity thresholds and growth responses of young mangrove trees as a function of chromium, fertilizer and soil type Growth response patterns show that the toxic effect of chromium is augmented in the natural soil Most surprising is the fact that in the natural soil fertilization amplifies the effect of chromium Furthermore, adverse effects of chromium contaminations become manifest only after a time delay of several months The deactivation of microorganisms in the soil proved to be a disadvantage for the growth and chromium tolerance of R apiculata compared to the growth on salt-leached soil and natural soil The microorganism's activities stimulated the development of the root system These findings are important for the planning of mangrove restoration and replantation in heavily contaminated environments The nonlinear regression model derived from the experiments can be used to assess the feasibility of such measures Based on the model results the following caveats apply for reforestation with young mangrove plants: i) Chromium contamination has a long term effect on plant growth Initial success of replantation may be followed by a complete breakdown of the plant population ii) Replanting in heavily contaminated soils with high salinity is likely to fail In this situation addition of fertilizer may have adverse effects Although we were able to capture the interaction between chromium stress and fertilizing in a comprehensive nonlinear regression model, the physiological mechanism behind the amplification of the chromium effect needs to be further studied Acknowledgements This work is part of the VNUHCM-BMBF Project EWATEC-COAST, supported by the Bundesministerium für Bildung und Forschung of Germany (BMBF) Grant No O2WCL1217A and by the Vietnam National University of Ho Chi Minh City (VNU-HCM) Grant No NDT2012-2401/HD-KHCN Supplementary data Supplementary data to this article can be found online at http://dx doi.org/10.1016/j.scitotenv.2016.09.063 References Acharya, G., 2002 Life at the margins: the social, economic and ecological importance of mangroves Madera y Bosques Número especial 53–60 Akaike, H., 1974 A new look at the statistical model identification IEEE Trans Autom Control 19 (6), 16–723 Akhand, A., Chanda, A., Dutta, S., Hazra, S., Sanyal, P., 2011 Comparative study of heavy metals in selected mangroves of Sundarban ecosystem J Environ Biol 33, 1045–1049 Alongi, D., 2011 Early growth responses of mangroves to different rates of nitrogen and phosphorus supply J Exp Mar Biol Ecol 397, 85–93 Barcelo, J., Poschenrieder, C., Gunse, B., 1985 Effect of chromium (VI) on mineral element composition of bush bean J Plant Nutr 8, 211–217 Chatterjee, J., Chatterjee, C., 2000 Phytotoxicity of cobalt, chromium and copper in cauliflower Environ Pollut 109, 69–74 Chen, Y., Ye, Y., 2014 Effects of salinity and nutrient addition on mangrove Excoecaria agallocha PLoS One (4) Dangremond, E., Feller, I., Sousa, W., 2015 Environmental Tolerances of Rare and Common Mangroves along Light and Salinity Gradients Oecologia Ghani, A., 2011 Effect of chromium toxicity on growth, chlorophyll and some mineral nutrients of Brassica juncea L Sci Technol 4, 197–202 Glime, J.M., Wagner, D.H., 2013 Laboratory techniques: slide preparation and stains In: Glime, J.M (Ed.), Bryophyte Ecology (Vol 3) Ebook sponsored by Michigan Technological University and the International Association of Bryologists Hawkins, A., Smith, R., Tan, S., Yasin, Z., 1998 Suspension-feeding behaviour in tropical bivalve molluscs: Perna viridis, Crassostrea belcheri, Crassostrea iradelei, Saccostrea cucculata and Pinctada margarifera Mar Ecol Prog Ser 166, 173–185 Jayatissa, L., Wickramasinghe, W., Dahdouh-Guebas, F., Huxham, M., 2008 Interspecific variations in responses of mangrove seedlings to two contrasting salinities Int Rev Hydrobiol 93 (6), 700–710 Jeyasingh, J., Philip, L., 2005 Bioremediation of chromium contaminated soil: optimization of operating parameters under laboratory conditions J Hazard Mater B118, 113–120 Ke, L., Tam, N.F., 2012 Phytoremeditaion Using Constructed Mangrove Wetlands: Mechanisms and Application Potential Nova Science Publishers, Inc Keshavarz, M., Mohammadikia, D., Gharibpour, F., Dabbagh, A.-R., 2012 Accumulation of heavy metals (Pb, Cd, V) in sediment, roots and leaves of mangrove species in Sirik creek along the sea coasts of Oman Iran J Appl Sci Environ Manage 16 (4), 323–326 Lee, S., Primavera, J., Dahdouh-Guebas, F., McKee, K., Bosire, J., Cannicci, S., Record, S., 2014 Ecological role and services of tropical mangrove ecosystems: a reassessment Global Ecology and Biogeography, (Global Ecol Biogeogr.) 23, 726–743 Lewis, M., Pryor, R., Wilking, L., 2011 Fate and effects of anthropogenic chemicals in mangrove ecosystems: a review Environ Pollut 159, 2328–2346 MacFarlane, G.R., Burchett, M.D., 2002 Toxicity, growth and accumulation relationships of copper, lead and zinc in the grey mangrove Avicennia marina (Forsk.) Vierh Mar Environ Res 54, 65–84 Mahdavi, E., Rahimi, A.E., Amini, H., 2012 Pb and Cd accumulation in Avicennia marina from Qeshm Island, Persian Gulf Iran J Fish Sci 11 (4), 867–875 Naidoo, G., 1987 Effects of salinity and nitrogen on growth and water relations in the mangrove Avicennia marina (Forsk.) Vierh New Phytol 107, 317–325 Naidoo, G., Hiralal, T., Naidoo, Y., 2014 Ecophysiological responses of the mangrove Avicennia marina to trace metal contamination Flora 209, 63–72 Nazli, M., Hashim, N., 2010 Heavy metal concentrations in an important mangrove species, Sonneratia caseolaris, in Peninsular Malaysia Environment Asia (Special Issue) 3, 50–55 Nguyen, H.A., Richter, O., Huynh, D.H., Nguyen, K.L., Kolb, M., Nguyen, V.P., Tran, B.T., 2014 Accumulation of Contaminants in Mangrove Species Rhizophora apiculata Along ThiVai River in the South of Vietnam EWATEC-COAST: Technologies for Environmental and Water Protection of Coastal Zones in Vietnam Contributions to 4th VNU – HCM International Conference for Environment and Natural Resources, ICENR Cuvillier, Göttingen, Germany (ISSN: 2363-7218 ISBN: 978-3-95404-852-6) Oliveira, H., 2012 Chromium as an environmental pollutant: insights on induced plant toxicity Journal of Botany 375843 (8 pages) Orcen, N., Nazarian, G., Gharibkhani, M., 2013 The responses of stomatal parameters and SPAD value in Asia tobacco exposed to chromium Pol J Environ Stud 22 (5), 1441–1447 Ouyang, X., Guo, F., 2016 Paradigms of mangroves in treatment of anthropogenic wastewater pollution Sci Total Environ 544, 971–979 Panda, S., Patra, H., 2000 Nitrate and ammonium ions effect on the chromium toxicity in developing wheat seedlings Proceedings of the National Academy of Sciences India Section B, Biological Sciences 70 (1), 75–80 (ISSN 0369–8211) Reef, R., Feller, I.C., Lovelock, C.E., 2010 Nutrition of mangroves Tree Physiol 30, 1148–1160 http://dx.doi.org/10.1093/treephys/tpq048 Richter, O., Nguyen, H.A., Nguyen, K.L., Nguyen, V.P., Biester, H., Schmidt, P., 2016 Phytoremediation by mangrove trees: experimental studies and model Chem Eng J 294, 389–399 Shanker, A., Cervantes, C., Loza-Tavera, H., Avudainayagam, S., 2005 Chromium toxicity in plants Environ Int 31, 739–753 Shukla, O.P., Rai, U., Dubey, S., 2009 Involvement and interaction of microbial communities in the transformation and stabilization of chromium during the composting of tannery effluent treated biomass of Vallisneria spiralis L Bioresour Technol 100, 2198–2203 Sodré, V., Caetano, V.S., Rocha, R.M., Carmo, F.L., Medici, L.O., Peixoto, R.S., Reinert, F., 2013 Physiological aspects of mangrove (Laguncularia racemosa) grown in microcosms with oil-degrading bacteria and oil contaminated sediment Environ Pollut 172, 243–249 Sundaramoorthy, P., Chidambaram, A., Ganesh, K., Unnikannan, P., Baskaran, L., 2010 Chromium stress in paddy: (i) nutrient status of paddy under chromium stress; (ii) phytoremediation of chromium by aquatic and terrestrial weeds Comptes Rendus Biologies 333, 597–607 Takemura, T., Hanagata, N., Sugihara, K., Baba, S., Karube, I., Dubinsky, Z., 2000 Physiological and biochemical responses to salt stress in the mangrove, Bruguiera gymnorrhiza Aquat Bot 68, 15–28 Tansel, B., Lee, M., Tansel, D.Z., 2013 Comparison of fate profiles of PAHs in soil, sediments and mangrove leaves after oil spills by QSAR and QSPR Mar Pollut Bull 73, 258–262 Turner, M., Rust, R., 1971 Effect of chromium on growth and mineral nutrition of soybeans Soil Sci Soc Am Proc 35, 55–758 Usman, R.A., Alkredaa, R.S., Al-Wabel, M.I., 2013 Heavy metal contamination in sediments and mangroves from the coast of Red Sea: Avicennia marina as potential metal bioaccumulator Ecotoxicol Environ Saf 97, 263–270 Ye, Y., Tam, N.F.-Y., Lu, C.-Y., Wong, Y.-S., 2005 Effects of salinity on germination, seedling growth and physiology of three salt-secreting mangrove species Aquat Bot 83, 193–205 Yim, M., Tam, N., 1999 Effects of wastewater-borne heavy metals on mangrove plants and soil microbial activities Mar Pollut Bull 39, 179–186 Zhang, F.-Q., Wang, Y.-S., Lou, Z.-P., Dong, J.-D., 2007 Effect of heavy metal stress on antioxidative enzymes and lipid peroxidation in leaves and roots of two mangrove plant seedlings (Kandelia candel and Bruguiera gymnorrhiza) Chemosphere 67, 44–50  ... thresholds of mangrove vegetation to the concentrations of pollutants need to be investigated This motivated our investigation on the physiological responses of mangroves under elevated contamination... standard deviation of the root length and root number of young R apiculata for each treatment combination The data shows the negative effect of the chromium treatment on root elongation, this... exchange The regulation of wateruse efficiency in plants occurs, in part, by changes in stomatal density Stomatal density is regulated by regulators Factors that increase stomatal density might