Precision Horticulture: Application of Optical Sensor Technology for Nitrogen Monitoring Status in Cocoplum, a Native Landscaping Plant

Cocoplum (Chrysobalanus icaco) is an ecologically significant native species to Southern Florida. Application of precision agriculture technologies such as optical sensors reduces the cost of over-fertilization and nutrient runoff. The aim of this work was to establish a base line sensor value for fertilizer treatment in cocoplum by monitoring chlorophyll content using the Soil Plant Analytical Development (SPAD), atLEAF, and Normalized Difference Vegetation Index (NDVI) sensors. Initial slow-released fertilizer treatment 8N-3P-9K was used at 15 g (control), 15 g (supplemented with +15 g × 2; T1), 15 g (+15 g; T2), 30 g (+15 g × 2; T3), 30 g (+15 g; T4), and 45 g (+15 g × 2; T5). Evaluations were conducted at 0 (base reading), 30, 60, 90, 120, 150, and 180 days after treatment. Growth parameters, optical non-destructive chlorophyll meters, leaf and soil total nitrogen and total carbon, and total nitrogen of leachate were analyzed. The results demonstrated that the treatment using 30 g slow-released fertilizer (8N-3P-9K) supplemented twice with 15 g in November and March after the first fertilization in October provided the least contamination through runoff while still providing adequate nutrients for plant growth compared to higher fertilizer concentrations. These results demonstrate that the highest treatment of nitrogen can cause considerable losses of N, causing extra costs to producers and environmental damage due to the flow of nutrients. Thus, techniques that help in N monitoring to avoid the excessive use of nitrogen fertilization are necessary. This study can serve as a basis for future research and for nurseries and farms, since it demonstrated from the monitoring of the chlorophyll content by optical sensors and by foliar and substrate analysis that lower treatments of nitrogen fertilization are sufficient to provide nutrients suitable for the growth of cocoplum plants.


Introduction
The total crop value of floriculture production in the United States topped $4.80 billion in 2020 with Florida representing 24% of the wholesale value for operations. Within the state of Florida, foliage was by far the largest produced category, evaluated to be $520 million in 2020 with garden plants representing the second highest at $245 million [1].
Chrysobalanus icaco, colloquially known as cocoplum or paradise plum, is a woody shrub native to South Florida. Cocoplum is comprised of two distinct ecotypes, inland, and coastal, differentiated by growing conditions and growth habits. Cocoplum is both ecologically and economically significant to South Florida [2].
Nitrogen (N) is one of the most influential macronutrients and is crucial in plant development especially in landscaping plants, being a limiting element of production. Due to this characteristic, it is intensively used in productive crops, aiming to get the crop to reach its maximum potential [3]. Generally, only a minor part of the N applied is recovered by crops, and the excess N is susceptible to loss to the environment where it is The aim of this study is to establish a base line sensor value for fertilizer treatment in cocoplum, a woody shrub native to South Florida, by monitoring chlorophyll content using SPAD and atLEAF sensors, and Normalized Difference Vegetation Index (NDVI), by a precision horticulture point of view. At the end of the treatments for six months, growth parameters, total nitrogen in the soil and in the leaf, and a leachate sample were also carried out in order to measure the nutrient runoff to find the best management practices among fertilizer treatment. The results of this study are extremely important and can serve as a basis for future research and for nurses and farms for best management practices among fertilizer treatment for cocoplum.

Results
The Table 1 shows the acronyms, sensors, and measures of each sensor used in this study.

Growth Characteristics Relative Chlorophyll Content (atLEAF), and Normalized Difference Vegetation Index (NDVI)
The growth characteristics, relative chlorophyll content (atLEAF), and NDVI did not differ significantly for the interaction between fertilization rate and evaluation period represented by days after fertilization (DAF). Therefore, these factors were evaluated separately. Fertilizer treatments were not significantly different in plant height and NDVI. However, atLEAF values and number of leaves were significantly different (p ≤ 0.05). The atLEAF values (66.22) and leaves number (215.09) were significantly (p ≤ 0.05) higher with using 45 g supplemented with +15 g (November and March; T5) compared to 15 g (control) with 61.97 and 182.91, respectfully (Table 2). The highest leaf number (227) was recorded after 90 DAF which was significantly (p ≤ 0.05) higher, compared to 0, 30, and 60 DAF (107.20, 171.53, and 187.10), respectively. The highest plant height was recorded at 180 DAF with 55.73 cm. These results show the plant growth and the increase in the relative chlorophyll content (atLEAF) and NDVI over the months, during six months of evaluation, showing the normal growth behavior of the plants (Table 3).

Relative Chlorophyll Content (SPAD)
There was significant interaction (p ≤ 0.05) between fertilization rate and days after fertilization for relative chlorophyll content (SPAD). An increase in relative chlorophyll content (SPAD) (67.32) was observed in the treatment 30 g (+15 g November; T4) at 90 days after fertilization ( Figure 1) by Tukey's test (p ≤ 0.05).

Total Nitrogen (TN) and Total Carbon (TC) of Leaf and Substrate Samples
There was significant interaction (p ≤ 0.05) between fertilization rate and days after fertilization for total nitrogen and total carbon of leaf samples. An increase in total nitrogen (2.51) was observed using 30 g (+15 g November and March; T3) at 150 days after fertilization, and an increase in total carbon (48.27) was observed in the 15 g (+15 g November and March; T1) at 150 days after fertilization ( Figure 2) by Tukey's test (p ≤ 0.05).

Total Nitrogen (TN) and Total Carbon (TC) of Leaf and Substrate Samples
There was significant interaction (p ≤ 0.05) between fertilization rate and days after fertilization for total nitrogen and total carbon of leaf samples. An increase in total nitrogen (2.51) was observed using 30 g (+15 g November and March; T3) at 150 days after fertilization, and an increase in total carbon (48.27) was observed in the 15 g (+15 g November and March; T1) at 150 days after fertilization ( Figure 2) by Tukey's test (p ≤ 0.05).
There was significant interaction (p ≤ 0.05) between fertilization rate and days after fertilization for total nitrogen and total carbon of leaf samples. An increase in total nitrogen (2.51) was observed using 30 g (+15 g November and March; T3) at 150 days after fertilization, and an increase in total carbon (48.27) was observed in the 15 g (+15 g November and March; T1) at 150 days after fertilization ( Figure 2) by Tukey's test (p ≤ 0.05). There was a significant (p ≤ 0.05) interaction between fertilization rate and days after fertilization, contributing to an increase in total nitrogen and total carbon of soil samples in 180 days after fertilization. An increase in total nitrogen (1.67) was observed using 30 g (+15 g November and March; T3) at 180 days after fertilization, while using 30 g (+15 g November; T4) provided an increase (38.29) in total carbon at 180 days after fertilization ( Table 4). The treatment 30 g (+15 g November and March; T3) provided higher values of total nitrogen in the leaf (2.51) and in the substrate (1.67). Nitrogen is one of the most influential nutrients of plant development, being a limiting element of production [3].  There was a significant (p ≤ 0.05) interaction between fertilization rate and days after fertilization, contributing to an increase in total nitrogen and total carbon of soil samples in 180 days after fertilization. An increase in total nitrogen (1.67) was observed using 30 g (+15 g November and March; T3) at 180 days after fertilization, while using 30 g (+15 g November; T4) provided an increase (38.29) in total carbon at 180 days after fertilization ( Table 4). The treatment 30 g (+15 g November and March; T3) provided higher values of total nitrogen in the leaf (2.51) and in the substrate (1.67). Nitrogen is one of the most influential nutrients of plant development, being a limiting element of production [3].

Salt, Electric Conductivity (EC), and Total Nitrogen (TN) of Leachate Samples
There was a significant (p ≤ 0.05) interaction between fertilization rate and days after fertilization for salt, electric conductivity, pH, and total nitrogen. For leachate samples at days 60, 90, 150, and 180 DAF, the treatment 45 g (+15 g November and March; 5) provided a higher value for salt (1622,2194,1966 (Table 5).
These results demonstrate that the highest treatment of nitrogen can cause considerable losses of N, causing extra costs to producers and environmental damage due to the flow of nutrients. Thus, techniques that help in N monitoring to avoid the excessive use of nitrogen fertilization are necessary. This study can serve as a basis for future research and for nurseries and farms, since it demonstrated from the monitoring of the chlorophyll content by optical sensors and by foliar and substrate analysis that lower treatments of nitrogen fertilization are sufficient to provide nutrients suitable for the growth of cocoplum plants (Figure 3). In addition, by providing less contamination by runoff, environmental hazard is avoided

Correlation Coefficient between Sensor Parameters, Number of Leaves (NL), and Total Nitrogen (TN) and Total Carbon (TC) of Leaf Samples
SPAD and atLEAF values were significantly (p ≤ 0.05) correlated at 90 DAF. Also, NDVI values were significantly (p ≤ 0.05) correlated with NL at 150 DAF. There were not significant correlations observed for any parameters at 30 and 180 DAF ( Table 6).
The correlation analysis evidenced a negative significant and high correlation (−0.857) between NDVI and number of leaves at 60 DAF. Also, negative significant and high correlations (−0.849, and −0.811) between NDVI and atLEAF, and total carbon and number of leaves, respectively, at 120 DAF, were observed (Table 6).
to the flow of nutrients. Thus, techniques that help in N monitoring to avoid the excessive use of nitrogen fertilization are necessary.
This study can serve as a basis for future research and for nurseries and farms, since it demonstrated from the monitoring of the chlorophyll content by optical sensors and by foliar and substrate analysis that lower treatments of nitrogen fertilization are sufficient to provide nutrients suitable for the growth of cocoplum plants (Figure 3). In addition, by providing less contamination by runoff, environmental hazard is avoided

Discussion
In a study conducted by Freidenreich et al. [26], application of precise amount of fertilizer at the right time is the most crucial task for horticultural nursery producers/managers. Therefore, this study used optical sensor reading and plant growth parameters to determine the sustainable ideal fertilizer rate for cocoplum plants as a base guideline in nursery production; furthermore, information collected can be used to determine if SPAD, atLEAF, and GreenSeeker TM are appropriate devices to estimate fertilizer need of the potted plants.
Nitrogen is an essential element for plant growth and development. It is a major component of chlorophyll in plant leaves. Several sensors have been designed to measure either the reflectance or the absorbance of the green color present in the leaves. The greenness of the leaves represents the amount of chlorophyll found in the chloroplasts, which can be used as an indirect indicator for the photosynthetic processes of the plant to determine plant health and vigor [18]. In fact, portable sensors have opened a new approach to acquire crop growth information rapidly and in a non-invasive manner [27].
In this study, the highest fertilizer concentration showed a higher number of leaves when compared to the control; also, the values of relative chlorophyll content (atLEAF) increased with higher fertilizer concentrations, showing the relationship between plant growth and N status. Similar results to this study were reported; Khoddamzadeh and Dunn [28] observed higher values for atLEAF in the highest fertilizer concentration of 15 g and 20 g N treatments in Chrysanthemum. In another study, Dunn et al. [29] reported that atLEAF readings increased with increasing N content in Salvia, and Swearengin et al. [30] observed that the atLEAF values increased with greater N rates in 'Helene Von Stein'.
Chlorophyll is the most important pigment of the leaf and one of the most important of the plant since it is through it that plants manage to capture sunlight and use it as an energy source. By means of sensors it is possible to estimate the amount of chlorophyll in the leaf, and thus be able to evaluate the deficiency of nitrogen in the plant, indicating the necessity of nitrogen fertilizer [31].
The highest value for the relative chlorophyll content (SPAD) was provided by treatment 30 g (+15 g November; T4) at 90 DAF (67.32). Other authors observed SPAD readings increased with increasing N content, such as Dunn et al. [32] in Gaillardia, Khoddamzadeh and Dunn [28] in Chrysanthemum, Dunn et al. [29] in Salvia, and Swearengin et al. [30] in 'Helene Von Stein'. However, Freidenreich et al. [26] observed a higher SPAD value for the 20 g fertilizer rate at eight weeks after a top-dressed treatment (WAT), compared to the highest fertilizer rate of 30, 40, and 50 g, the control, and the 10 g in Justicia brandegeana.
Leaf tissue N analysis refers to the measurement of total N content in leaf blades of the most recently fully expanded leaves. It is a long-established method for monitoring crop N status [33,34]. Although tissue analysis is limited as a N monitoring approach, multi-element tissue analysis is useful for diagnosis of possible nutritional problems [35].
Khoddamzadeh and Dunn [28] reported that the Leaf N increased with increasing fertilizer rates through 38 days after a top-dressed (DAT) pretreatment. The results were different from our findings with usage of 30 g with higher Leaf N compared to the highest concentration of 45 g at 90, 120, 150, and 180 DAF. The results proved N monitoring is very important to avoid over-fertilization and at the same time provide adequate nutrients for plant growth. In addition, these findings could serve as a base guideline for cocoplum fertilization in nurseries and landscapes in South Florida.
Carbon-nitrogen metabolism is the most basic and important nutrient metabolism of plants, and its dynamic changes in the plant directly affect the absorption, transformation of mineral nutrition, formation of protein, and so on [36,37]. Therefore, carbon and nitrogen metabolism and their harmony affect plant growth and development [38].
The correlation analysis demonstrated the association between SPAD and atLEAF sensors at 90 DAF; both sensors can be used to monitor the fertilizer status of the potted cocoplums. Another important observation was the association between NDVI and number of leaves at 150 DAF; this correlation demonstrates that the normalized difference vegetation index is positively related to the number of leaves; that is, when there is a linear increase in one parameter, the same thing occurs for the other parameter. Significant negative correlations were also observed between NDVI and number of leaves at 60 DAF, for NDVI and atLEAF, and between total carbon and number of leaves at 120 DAF. For negative correlations, as there is a linear increase in one parameter, the other parameter decreases.

Growth Analyses
Five plants per treatment were evaluated monthly to the number of leaves (unit) by a counter, and plant height (cm) by a tape measure. Two branches of each plant, one larger and one smaller, were marked and measured, and the average of these two branches represented the plant height.

Relative Chlorophyll Content and NDVI
Individual plants were scanned from five pots per treatment using a SPAD-502 chlorophyll meter (SPAD-502, Konica Minolta, Japan), an atLEAF chlorophyll meter (FT Green LLC, Wilmington, DE, USA), and a GreenSeeker TM Normalized Difference Vegetation Index (NDVI) sensor (Trimble Agriculture, Sunnyvale, CA, USA). During measurements, the NDVI (Figure 5a) sensor was placed 45 cm above the plant canopy. For the SPAD and atLEAF (Figure 5b,c), measurements were collected from four mature leaves from the middle area of the plant.

Leachate Samples
This analysis was performed from individual containers to determine nutrient runoff rates. Each plant was irrigated until a saturated state was reached. Once containers reached the saturation point, a tray was placed underneath, serving as a collection reservoir. The plants were further irrigated with 350 mL of water, allowing the collection of 50 mL leachate. Samples were stored in 50 mL conical tubes that were immediately refrigerated at 4 • C until laboratory analysis in the CAChE Nutrient Analysis Core Facility at Florida International University; the test was performed for total nitrogen (ppm). The electric conductivity (EC), and salt of each leachate sample was measured in situ.

Relative Chlorophyll Content and NDVI
Individual plants were scanned from five pots per treatment using a SPAD-502 chlorophyll meter (SPAD-502, Konica Minolta, Japan), an atLEAF chlorophyll meter (FT Green LLC, Wilmington, DE, USA), and a GreenSeeker TM Normalized Difference Vegetation Index (NDVI) sensor (Trimble Agriculture, Sunnyvale, CA, USA). During measurements, the NDVI (Figure 5a) sensor was placed 45 cm above the plant canopy. For the SPAD and atLEAF (Figure 5b,c), measurements were collected from four mature leaves from the middle area of the plant.

Leachate Samples
This analysis was performed from individual containers to determine nutrient runoff rates. Each plant was irrigated until a saturated state was reached. Once containers reached the saturation point, a tray was placed underneath, serving as a collection reservoir. The plants were further irrigated with 350 mL of water, allowing the collection of 50 mL leachate. Samples were stored in 50 mL conical tubes that were immediately refrigerated at 4 °C until laboratory analysis in the CAChE Nutrient Analysis Core Facility at Florida International University; the test was performed for total nitrogen (ppm). The electric conductivity (EC), and salt of each leachate sample was measured in situ.

Leaf and Substrate N and C Content
Five plants per treatment were used for leaf samples and they were collected monthly. For the substrate samples, five plants per treatment were used and they were collected at the beginning and at the end of the experiment. The leaf and substrate samples were dried at 70 °C for 48 h, ground, and then analyzed for the total nitrogen (%) and total carbon (%). These analyses were performed at the CAChE Nutrient Analysis Core Facility at Florida International University.

Leaf and Substrate N and C Content
Five plants per treatment were used for leaf samples and they were collected monthly. For the substrate samples, five plants per treatment were used and they were collected at the beginning and at the end of the experiment. The leaf and substrate samples were dried at 70 • C for 48 h, ground, and then analyzed for the total nitrogen (%) and total carbon (%). These analyses were performed at the CAChE Nutrient Analysis Core Facility at Florida International University.

Statistical Analysis
The experiment was established in a completely randomized design with six treatments that were replicated five times with single pot replications (one plant in each pot), totaling 30 plants. Data were subjected to analysis of variance (ANOVA), and the means were compared by Tukey's test (p ≤ 0.05) using the SISVAR statistical program [39]. The correlation analysis was performed between means of sensor parameters, number of leaves, and total nitrogen and total carbon of leaf samples was performed using the GraphPad Prism version 9.4.1 for Windows, GraphPad Software, San Diego, CA USA, (www.graphpad.com, accessed on 27 July 2022).

Conclusions
This study was performed by monitoring the chlorophyll content using handheld non-destructive optical sensors to demonstrate a lower amount of fertilizer with providing adequate nutrients for plant health and growth in cocoplum as well as less pollution via runoff, thus reducing environmental damage. The treatment using 30 g slow-released fertilizer (8N-3P-9K) supplemented twice with 15 g in November and March after the first fertilization in October provided the least contamination through runoff while still providing adequate nutrients for plant growth compared to higher fertilizer concentrations. The results of this study could serve as a guideline for nursery producers and landscape personnel as a fast and non-destructive tool for sustainable fertilizer management practices within the ornamental plant industry.
Funding: Funding for project 14 was made possible by the U.S. Department of Agriculture's (USDA) Agricultural Marketing Service through grant AM200100XXXXG037. Its contents are solely the responsibility of the author and do not necessarily represent the official views of the USDA.

Data Availability Statement:
The dataset is available upon reasonable request to the corresponding author.