sas analytics experience 2016

#analyticsx

Presenter: Reuben Hilliard

Faculty Advisors: Dr. Paula Jackson & Dr. Brad Barney

Kennesaw State University

INTRODUCTION

Surviving Climate ChangeComparing drought & fungicide response in two riparian tree species for use in ecological restoration

Riparian zones fulfill many ecosystem functions and occur at all elevations near rivers, streams, and in floodplains. They function as a natural buffer against erosion in river and stream banks, filter downstream pollution, and provide increased habitat complexity (Wildlife, 2006). Due to development, logging, and expanding agriculture, many riparian zones have been destroyed or depleted. This has a profound ecological effect that leads to increased sedimentation and pollution in natural water systems (Hernandez-Santana, 2011). These zones often undergo rehabilitation to restore health back into the surrounding environment. Salix nigra (Black Willow) and Platanus occidentalis (American Sycamore) are two common riparian species of trees (Conger, 1996). Of these, Salix nigra is frequently used to restore these areas, however not much information exists on the ability to use Platanus occidentalisfor this purpose. This research is part of a larger study looking at the ecology and physiology of both of these species, with the overarching aim of comparing the behavior of Platanus occidentalis to the more widely studied Salix nigra, and determining the feasibility of using Platanus occidentalis in restoration processes. Of additional importance is the fungal biota which inhabit the soil beneath these trees. Mycorrhizal fungi have been reported to improve plant growth in many crops through enhanced root growth and function (Westphat et al, 2008). It also improves early plant establishment and increased the most valuable early fruit yield under some environmental stress conditions. This is of ecological importance and will be incorporated in this study.

Salix nigra leaves (left) and Salix nigra tree (right)

Platanus occidentalis leaves (left) and Platanus occidentalistree (right)

(Image: bioimages.vanderbilt.edu/baskauf/15370.htm)



(Image: http://bioimages.vanderbilt.edu/baskauf/13574.htm)

STATISTICAL MODELS

Randomized Complete Block Design

yij = µ + αi + bj + εij

Mixed Effects Model

y = Xβ + Zu + ε

Unstructured Covariance Matrix

Autoregressive Heterogeneous Covariance Matrix

One-way ANOVA

yij = µ + τi + εij

Sum of MeanSquares df Square F

Treatment SSR / dfF = MSR MSR/MSEError SSE / dfE = MSETotal SST dfT

Both species can be found in overlapping geographical regions of the Southeast.

#analyticsx





One cause for data errors in greenhouse experiments can be due to microclimate differences, such as light, airflow andheat among saplings, within different planters. As stated by Brien et al. (2003), sound statistical design and analysis isbetter than rearranging the position of plants during the experiment itself. In our experiment, Platanus occidentalis andSalix nigra saplings were planted using a Randomized Complete Block Design, as seen in Table I. It involved a completeexperimental treatment within each planter series, allowing for homogenous growing conditions within a single tray,regardless of differences among planters themselves. To account for variability in sapling size, each tray had a similardistribution of size classes. All the saplings were tagged with unique identifiers, such as PO-01 or SN-02. 15-17individuals of each species were subjected to a control, inundation, or drought condition; and drought with andwithout the addition of mycorrhizal spores, a fungal biota, which was controlled with the addition of a fungicide,Benomyl. In total, 31 Platanus and 34 Salix cuttings received sufficient nutrients in the form of a slow-release fertilizerand after taking baseline measurements, were allowed to grow in planters through the spring of 2015. From May 18thuntil August 23rd, a team of myself and 3 undergraduate research assistants, took anatomical and physiologicalmeasurements.

Block 1 Block 2

A B B A

D C C D

TreatmentsA: Control; No Fungi (Benomyl added) & WetB: Fungi & DroughtC: Fungi & WetD: No Fungi (Benomyl added) & Drought

Saplings planted; fertilizedSaplings measured; tagged

METHODS

For the anatomical measures, an indicator of growth rate, both the circumference and the height were taken for eachplant on a weekly basis. For the physiological data, a LICOR LI-6400 Infrared Gas Analyzer (IRGA) was used to measurephotosynthetic rate from leaves repeatedly over the period of weeks, systematically moving through the plants,selecting a predefined leaf from a randomly selected plant from each treatment and block. This was a tedious process,with each leaf taking up to 16min for a full measurement run. The results were used to build light response curves. NetPhotosynthetic rate, or CO2 assimilation (µmol CO2 m-2 leaf area s-1) from several trials were plotted against lightintensity, or Absorbed Photosynthetically Active Radiation (αPAR, µmol photons m-2 leaf area s-1). The slope of thelinear phase of the response curve is a measure of "photosynthetic efficiency" of the plant, or how efficiently solarenergy is converted into chemical energy. Different plants show differences in the shape of their light response curves,which reveals characteristics of the underlying photosynthesis processes, including the efficiency at which light isutilized by photosynthesis and the rate of O2 uptake.In this longitudinal study, both the anatomical and the physiological data were analyzed using SAS 9.4 with the MIXEDProcedure, which models mixed effects over time.Because a drought stress condition increases root to shoot ratio (Xu et al, 2015), to test this, 16 trees were randomlyselected from each of the treatment and species groups. After drying in the oven overnight, the dry weight of eachtree was used to calculate the root to shoot ratio. The cumulative total leaf area was taken using the LI-COR 3100C. Todetect the presence of arbuscular mycorrhizal root colonization, 160 root samples were collected from eight trees ofeach species. Roots were stained using a 0.05% Trypan Blue solution and the presence of mycorrhizal structures wasquantified using the root piece method.

The LI-COR LI-6400 (left) is a portable Photosynthesis System and a major tool for ecological researchers in the field (LI-COR, 2013). The function of the LI-6400 is based on detecting differences in CO2 concentration of air before (reference) and after (sample) it comes in contact with the plant leaf. Differences in CO2concentrations are detected through the use of infrared gas analyzers. The LI-COR allows for in situ measurements of photosynthetic rates and for independent control of the leaf chamber CO2, H2O, temperature, and light . Results from the LI-COR may be downloaded into an Excel spreadsheet for further analysis.

Physiological measurements (top left); Planters in grid – trial start date (top mid); Anatomical measurements (top right); LI-COR 3100C Leaf Area Meter (bottom left); Staining of greenhouse samples in 0.05% Trypan Blue solution (bottom mid); Colonized root after staining (bottom right)

#analyticsx





RESULTS

Graphs, Boxplots and Bar Chart output designed in SAS 9.4 and Tableau 9.2

Anatomical Results from the Mixed Effects Model. Drought treatment highly significant (** p < 0.0001)

#analyticsx




CONCLUSIONS


The anatomical results indicate that Salix nigra and Platanus occidentalis do respond differently to droughtconditions. In fact, the interaction of the days count and the drought condition was highly significant (p < .0001).This meant that as the experiment proceeded the drought condition became more pronounced. As seen in theanatomical results, for the drought condition, Platanus was able to outperform Salix in the linear rate of growth(corrected for errors), 0.86cm/Day and 0.74cm/Day, respectively. The fungicide treatment did not have a significanteffect in either species.The physiological results indicated that the PAR level, or light intensity given to the leaf, was significant (p < 0.05). Thisanalysis had to be stratified by species, as none of the other factors were significant initially. When this was done,Platanus was right on the cusp of being significant in the drought condition during period 2 (last 5 weeks of theexperiment). Even though Salix didn’t respond to the drought or fungicide treatments, this information is still ofbiological interest to researchers, as it indicates that Salix can be stressed and still perform unhindered, with minimalinterruptions.Of particular note is how well Platanus performed in the ‘No Fungicide/Drought’ (Treatment B), which can be observedin both the light curve and the mean maxima photosynthetic output figures. But when an ANOVA was performed toconfirm the results, there was no significant difference at PAR 800 by Species and Treatment.The ANOVAs from the Root to Shoot Ratio and Leaf Area showed no significant differences in means between Speciesand Drought treatment, but of interest was the much larger variability among Salix plants. Larger Root to Shoot Ratiosamong younger trees and greater variability in measurements, indicate that compared to Platanus, when Salixscavenged for water, it struggled far more in the drought treatment.Because Platanus performed as well or slightly better than Salix in this study, the overall results were quite positiveand allow future research to focus specifically on Platanus as a species to use in restoration of Southeastern USriparian ecosystems.

RELEVANT SAS CODE

REFERENCES

*i) Coding the explanatory variables;DATA research.anatomical_data;

SET together;daysc = date - mdy(5,18,2015); dayscat = daysc;tag = species; species = substr(tag,1,2);fungicide = scan(treatment,1,"/"); drought = substr(scan(treatment,2,"/"),1,1);tmt=1; if species="PO" then tmt=tmt+4;if drought="D" then tmt=tmt+2;if fungicide="Fungicide" then tmt=tmt+1;

RUN;

*ii) Optimal Anatomical model. Unstructured Covariance Matrix;PROC MIXED data = research.anatomical_data;CLASS species drought fungicide tag blocknum dayscat;MODEL height = species drought species|daysc drought|daysc

/solution ddfm = kr;REPEATED dayscat/ subject = tag(blocknum) type = un R RCORR ; RUN;

*iii) Optimal Physiological model, stratified by species. Autoregressive Heterogeneous Covariance Matrix;PROC SORT data = research.master_photo; BY Species Tag PAR; run;PROC MIXED data = research.master_photo;BY Species;CLASS tag treatment blocknum PAR;MODEL photo = treatment PAR ;RANDOM blocknum;REPEATED PAR /subject = tag type = arh(1) ;LSMEANS treatment /pdiff tdiff ; RUN;

*iv) Data output for Mean Maxima Photosynthetic Rate and Standard Errors;

ODS GRAPHICS ON;PROC MIXED data = research.master_photo;WHERE PAR = 800;CLASS Tag blocknum fungicide drought species PAR period;MODEL photo = species*fungicide*drought

/ noint solution ddfm=kr;RANDOM blocknum;RANDOM int / subject=tag;

RUN; ODS GRAPHICS OFF;

*iv) One-way ANOVA for Root to Shoot Ratio and Total Leaf Area;TITLE "Root to Shoot Ratio";PROC ANOVA data=mass_area;CLASS Fungi Drought Species ID Block drought_sps ;MODEL Root_Shoot_Ratio = drought_sps Block;MEANS drought_sps/Tukey;

RUN;

TITLE "Total Leaf Area";PROC ANOVA data=mass_area;CLASS Fungi Drought Species ID Block drought_sps ;MODEL Leaf_Area = drought_sps Block;MEANS drought_sps /Tukey;

RUN;

Oregon Department of Fish and Wildlife. 2006. Oregon Conservation Strategy. Oregon Department of Fish and Wildlife, Salem, Oregon.

Hernandez-Santana, V., Asbjornsen, H., Sauer, T., Isenhart, T., Schilling, K., & Schultz, R. 2011. Enhanced transpiration by riparian buffer trees in response to advection in a humid temperate agricultural landscape. Forest Ecology and Management, 261(8), 1415-1427.

Conger, RM. 1996. Black willow (Salix nigra ) use in phytoremediation techniques to remove the herbicide bentazonfrom shallow groundwater. Master’s thesis, Louisiana State University

Brien, C. J., Berger, B., Rabie, H., & Tester, M. 2013. Accounting for variation in designing greenhouse experiments with special reference to greenhouses containing plants on conveyor systems. Plant Methods, 9(5), 1746-4811

Westphal, A., Snyder, N., Xing, L. 2008. Effects of Inoculations with Mycorrhizal Fungi of Soilless Potting Mixes During Transplant Production on Watermelon Growth and Early Fruit Yield. HortScience, 43(2), 354-360

LI-COR Biosciences. (2013). The LI-6400 Portable Photosynthesis System. Retrieved from http://envsupport.licor.com/index.jsp?m=Current&spec=LI6400,Brochures&menu= Photosynthesis%20Systems

Xu, W., Cui, K. , Xu, A., Nie, L., Huang, J., & Peng, S. 2015. Drought stress condition increases root to shoot ratio via alteration of carbohydrate partitioning and enzymatic activity in rice seedlings. Acta Physiologiae Plantarum, 37:9

sas analytics experience 2016

Documents