Predicting leaf arsenic concentration in hydroponically grown rice and spinach leaves using narrow-band leaf reflectance and stereological measurements.

摘要

Contamination of soil and water with inorganic arsenic (As) is a serious threat to human and ecological health because plants grown in As-contaminated soils can accumulate high levels of As into shoots and leaves. The presence of As in edible portions of plants allows for potentially dangerous ingestion by humans and animals. The ability to detect As in plants is an important tool to minimize such risks. The objectives of this research were: (1) to use leaf reflectance, and mathematical transformations of leaf reflectance, to predict As concentrations in leaves of hydroponically-grown rice ( Oryza sativa) and spinach (Spinacia oleracea), and (2) to use stereological techniques to quantify leaf structural changes in spinach leaves caused by As.;The hydroponic study was conducted at the USDA-ARS Beltsville Agricultural Research Center in Beltsville, Maryland. Experiments were arranged as a completely randomized design with four treatments and either five replications (rice) or four replications (spinach). Treatments consisted of one control and three levels of As (added as Na2HAsO4): 5, 10, and 20 mumol As L-1. Biophysical and biochemical data include plant dry weight, elemental analyses, and leaf chlorophyll content. Hyperspectral reflectance data were acquired over the 350 to 2500 nm range using an ASD spectroradiometer. Only normally-distributed data were used for statistical analyses. Leaf structural data were obtained for spinach using confocal and light microscopy. Stereological measurements were used to calculate structural components such as leaf thickness, ratio of palisade to spongy mesophyll, mean mesophyll cell surface area, and ratio of intercellular air spaces surface area to external surface area. Spectral transformations included vegetative indices (VIs) and first derivative reflectance (FDR).;Results from the rice study showed that leaf dry weight and leaf chlorophyll content decreased significantly with increased solution As concentration. In addition, visible symptoms such as chlorosis along the leaf margin and reddish brown discoloration of roots were observed. Leaf reflectance increased in visible wavelengths (400 to 700 nm) and decreased above 750 nm as solution As increased. Correlation analysis showed that leaf As concentration was correlated with leaf reflectance and first derivative reflectance at different wavelengths. Reflectance values in the red edge region (690-730 nm) were well correlated with leaf As concentration, indicating that changes in leaf reflectance were due, in part, to plant stress. Regression analyses found that FDR was a better predictor of leaf As concentration than were leaf reflectance or VIs. While all of the models tested gave encouraging results for predicting As concentration, the FDR ratio of 706/723 nm generated the highest coefficient of determination (R2 = 0.75).;Spinach plants were less responsive to As treatments, and showed no visible signs of toxicity in the leaves as leaf As concentration increased. Leaf dry weight was significantly reduced with solution As concentration, as was leaf chlorophyll content, although differences in chlorophyll content between the three rates of added As were not significant. Quantitative analysis of leaf structure showed that total leaf thickness and intercellular spaces in spongy mesophyll cells decreased with As treatment. Changes in leaf reflectance in visible wavelengths were not well-correlated with leaf As concentration. However, leaf reflectance in near infrared wavelengths was strongly correlated with leaf As concentration and leaf structural changes. Multi linear regression of leaf reflectance values at the highest correlated wavelengths (1048, 1098, 1081, and 1080 nm) generated an R2 value of 0.68. Mathematical transformations, except for FDR, did not improve prediction of spinach leaf As concentration.;Results from this research support the use of remote sensing techniques to predict leaf As concentration in contaminated rice and spinach. However, in rice the best results were in visible wavelengths while the best results in spinach were in the near infrared spectrum. This research was conducted under controlled conditions; further research is necessary to verify these results in natural conditions and at large spatial scales.