Merit winner for the IES Publication Awards (Student Category) 2003 School of Mechanical and Production Engineering Nanyang Technological University Singapore Title: Nutrient Stress Analyser For Hydrophonic Plants Name: Wong Chee Beng Supervisor: Professor Dr. Anand Krishna Asundi Introduction Nutrient management is a major concern for farmers and agriculturalists. Besides water and sunlight, plant growth and survival is very dependent on available nutrients. Once the plant is exposed to severe nutrient deficiency, it is very difficult for the plant to fully recover. Therefore it is important to develop a device that is able to detect nutrient stress early, allowing for early corrective action to be taken. Existing devices for nutrient management have limitations, are expensive and destructive to plants in some cases. This project investigates an alternative method to build a device which is cheap, reliable, user friendly and able to do large samplings to detect nutrient stress. Principles of Plant Nutrition
There are 16 essential nutrient elements required by plants. These are subdivided into micronutrient and macronutrient. Each of these nutrient elements has its own deficiency symptoms and functional role within the plants. Analysis of these visual symptoms, such as chlorosis, curling of leaf and stunted growth, are the basis of this research. The acquired results are carefully studied to determine distinct pattern for each nutrient deficiency. The 16 nutrient elements are further categorized into mobile and immobile nutrient element. Mobile nutrient elements are nutrients that can be transferred from one leaf to another and the first apparent symptoms of deficiency are found in the old leaves. Priority is given to the young leaf to ensure survival and continued growth of the plant. For immobile elements, the deficiency symptoms are first apparent in the young leaf. The young leaves have much faster metabolic growth than the old plants, so the young leaf would exhaust its own nutrient supply much earlier. Thus, the young leaves would show deficiency symptoms first.
Crop Yield Criticalnutrient range Hidden Hunger Sufficiency level Toxic Visual deficiency symptoms The sufficiency yields may have a wide range of nutrient concentration before the nutrient becomes toxic (thus the broken bar) Nutrient concentration in the plant increase Figure 1 Plant Growth, Nutrient Concentration and Its Effect on State of the Plant (Munson and Nelson 1990)
Figure 1 shows the relationship between crop yield and nutrient concentration. The parabolic graph has a maximum at the nutrient sufficiency level, corresponding to the highest crop yield. A decrease from this level of concentration puts the plant in the critical nutrient range and a slight decrease in nutrient concentration from this level will cause the plant to drop to the hidden hunger stage. At this stage the crop yield starts to drop significantly. A further drop of nutrient concentration will force the plant into the visual deficiency symptoms stage and at this stage the plants will exhibit severe deficiency symptoms such as yellowing of leaf and stunted growth. At this level, the crop yield is the lowest. If the nutrient concentration becomes larger than the sufficiency level, the crop yield drops because the nutrient has become too concentrated and too toxic for the plant. Thus a balance in nutrient concentration and a need for good nutrition management is very important to ensure healthy and abundant harvest. Experimental Details The flow chart in figure 2 summarizes the experimental procedure. The test plant for this experiment was Chinese Caixin (Brassica Chinesis). The caixin seedlings were germinated for 3 to 4 days to ensure that the seedlings stem harden and would not break during the transplanting procedure. The germinated plants were transferred into the Deep Flow System to grow for about 17 days. 50 healthy plants were carefully selected and divided into 5 batches of 10 plants. Each batch was exposed to either nitrogen deficiency, or iron deficiency, or calcium deficiency, or phosphorus deficiency or no deficiency (complete solution) for 7 days. The spectra (figure3) for old leaf and young leaf of all the
plants were taken every day for 7 days. The whole procedure was repeated for the recovery process for another 7 days whereby all 50 plants were exposed to complete solutions. In the experiment, the young leaf is predefined as the second youngest leaf with observable stem and fully opened foliage whilst old leaf is predefined as the third oldest leaf. Preparation of Nutrient Solution Germination Hardening of germinate 3-4 days Transplanting germinated seedlings to Deep Flow System About 17 days Duration 7 days Exposing plants to different deficiencies- 10 plants Recovery exposing the plant to complete solution Duration 7 days Figure 2 Flow Chart of Test
Young leaf Old Leaf Figure 3 Typical Spectra of Young and Old Leaves Spectrometer Tungsten Halogen Light Source Light Source Transmitting fiber optics Bifurcated cable Probe Output Capture fiber optics Laptop Plant Figure 4 Experimental Setup The spectra are measured using reflectance spectroscopy schematically shown in figure 4. It consists of a light source, a spectrometer, a bifurcated fiber cable with a probe on the distal end and a laptop for data analysis. Tungsten halogen lamp is used as light source because of its white light characteristics. The bifurcated fiber cable has two bundles, one with six fibers was connected to the light source, and the remaining single fiber received light reflected from the leaf and transmitted it to the spectrometer.
Analysis Method Analysis is done by observing two changes: color change and spectra physical change. Each of the deficiency has its own characteristic symptoms. An understanding on observable symptoms of each deficiency is crucial in predicting the outcome of the result. The first method uses principal of CIE color system to analyze the spectra. RGB values are deduced from the spectra using prescribed algorithms. These RGB values provide precise description of the color changes of the tested leaf caused by deficiency. The second approach is the physical changes in the spectrum curve. Some of the physical in the spectra include intensity and wavelength shift, and changes in intensity ratios or gradients. This approach is useful to relate deficiency symptoms that exhibit physical changes on the leaf (instead of color), for example, curling of leaf or withered edge leaf. Results Iron Deficiency Iron is an immobile nutrient element and its deficiency symptoms would first be apparent in young leaves. Therefore to analyze iron deficiency plants, the results would be obtained from the young leaves. From figure 5, it is observed that iron had the lowest BLUE component which starts to drop drastically by day 3. This trend is distinctive for iron deficiency and very different from other nutrient deficiencies and the complete solution. The first YELLOWING was apparent to the naked eye by day 5. The nutrient stress analyzer is able to detect this deficiency by day 2.
0.205 Ratio of BLUE in RGB 0.2 0.195 0.19 0.185 0.18 0.175 0.17 0.165 0 2 4 6 8 10 Day Calcium Complete Iron Nitrogen Phosphorus Figure 5 Bluing of Young Leaves Nitrogen Deficiency Nitrogen deficiency has the same symptoms as that of iron deficiency that is YELLOWING of leaves (chlorosis) but for nitrogen, the symptoms appear in the old leaves. Since nitrogen deficiency exhibits the same symptoms as that of iron deficiency, the same pattern found in iron deficiency is apparent as in figure 6. Ratio of BLUE in RG 0.191 0.19 0.189 0.188 0.187 0.186 0.185 0.184 0.183 0.182 0.181 0 2 4 6 8 10 Day complete nitrogen phosphorus Figure 6 Bluing of Old Leaves
From the graph, the BLUE component starts to drop from day 7. Thus the nitrogen deficiency symptoms appear much later than iron deficiency. However, after careful study it was found that by looking at only the BLUE component for nitrogen deficiency was inadequate and led to false indication. It was observed that the RED component also varies with Nitrogen deficiency. Thus monitoring of both blue and red components permits proper detection for Nitrogen deficiency. Phosphorus Deficiency Phosphorus is an immobile nutrient. From figure 6, phosphorus deficiency shows higher values for the BLUE component as compared to the complete solution. It was found that for phosphorus deficiency exhibited higher BLUE, higher RED and lower GREEN components compared to the control. This pattern starts by day 3 and continues till day 8. This pattern agrees with the symptoms of phosphorous; leaves turning to dark greenish purple. A high percentage of BLUE and RED in phosphorus deficiency changed the leaves to dark greenish purple. This trend only occurred in phosphorus deficiency; therefore, these findings could be used to differentiate phosphorus deficiency from the rest of the solutions.
Discussions The nutrient stress analyzer is able to detect nutrient stress early. Some questions still need to be resolved. The nutrient stress analyzer was able to pick up nutrient deficiency by day 3. However not all deficient plants were identified. This was due to the inequality setting which distinguished a deficient plant from the control plant; in other words, the component of red and blue which separates a healthy plant from a deficient plant. By varying the setting, it would be possible to identify more deficient plants, but the occasional control plant may also be identified as deficient. By setting the false detection to less than 5%, it is possible to identify more than 80% of deficient plants. These numbers can be further improved by exercising greater control on the experimental process. Secondly it is observed that yellowing of leaves is an indication of a nutrient deficient plant. Thus instead of RGB, the CMY (Cyan, Magenta, Yellow) color space was considered to see if it gave any advantages. It was found that by using CMY it was a lot easier to conclude the result for iron deficiency since the anticipated color is yellow (Figure 7). However the lack of other nutrients was not noticeable. For the older leaves (please refer to figure 8), the YELLOW in nitrogen deficiency which exhibits the same symptoms (yellowing of leaves), as iron in young leaves, no such trend is observed. The result obtained showed that CMY color space did not ease the analyzing of result. This proved that there was still a need to look at the color combination as a whole and whether RGB or CMY it would constitute to the same result
Ratio of Yellow in CMY 0.835 0.83 0.825 0.82 0.815 0.81 0.805 0.8 0.795 0 2 4 6 8 10 Day Calcium Complete Iron Nitrogen Phosphorus Figure 7 Yellowing of Young Leaves. Ratio of Yellow in CMY 0.819 0.818 0.817 0.816 0.815 0.814 0.813 0.812 0.811 0.81 0.809 0 2 4 6 8 10 Day complete nitrogen phosphorus Figure 8 Yellowing of Old Leaves
Conclusion The project shows that the nutrient stress analyzer has great potential in detecting nutrient stress despite some shortcomings. The potential of the device could be extended to the other nutrient elements and if this proves successful the nutrient stress analyzer could greatly help by providing a cheaper alternative and reliable source into nutrient management. References 1. Caltech (2000) Common Symptoms of Nutrient Deficiency in Plant http://www.cco.caltech.edu/ aquaria/krib/plants/fertilizer/nutrient-deficiency.html 2. E.J.Hewitt. (1963). The essential nutrient elements. pp 137-360. London : Her Majesty s Stationary Office 3. E.J.Hewitt. (1983 ). The Effects of Mineral Deficiencies and Excesses on Growth and Composition. Vol. 1 London : Her Majesty s Stationary Office 4. John Walker (1996). Colour Rendering Spectra http://www.fourmilab.ch/ 5. Lasse L. (1998). Spectra and Color http://www.it.lut.fi/opetus/9900/010588000/exercises/06/spectra/ 6. Maurice E. Watson, (1998). Analytical Instruments for the Determination of Elements in Plant Tissue. Soil and Plant Analysis Concil,nc. CRC Press LLC 7. Newfoundland and Labrador Agricultura (2002)l. Plant Nutrition. http://www.gov.nf.c /agric/pubfact/fertility/nutrition.htm#1
Acknowledgement The author wishes to express his sincere appreciation to the following people and institutions: 1. Professor Dr. Anand Krishna Asundi, the supervisor of this project, for his invaluable advice, support and guidance throughout the project; 2. Dr Liew Oi Wah, the supervisor of this project, for her invaluable guidance, patience and assistance throughout the project; 3. Miss Li Bingqing, the author s mentor and friend, for her invaluable help and friendship; 4. Mr. William S.L. Boey, the technical staff of the Solarium Laboratory, for his assistance and friendship. 5. NTU and Singapore Poly for funding, supporting and making this project possible.