Abstract
Studying the canopy spectral reflection characteristics of different N-efficient maize varieties and analyzing the relationship between their growth indicators and spectral vegetation indices can help the breeding and application of N-efficient maize varieties. To achieve the optimal management of N fertilizer resources, developing N-efficient maize varieties is necessary. In this research, maize varieties, i.e., the low-N-efficient (Zhengdan 958, ZD958), the high-N efficient (Xianyu 335, XY335), the double-high varieties (Qiule 368, QL368), and the double inefficient-type varieties (Yudan 606 YD606), were used as materials. Results indicate that nitrogen fertilization significantly increased the vegetation indices NDVI, GNDVI, GOSAVI, and RVI of maize varieties with different nitrogen efficiencies. These findings were consistent with the performance of yield, dry matter mass, and leaf nitrogen content and were also found highest under both medium and high nitrogen conditions in the double-high variety QL368. The correlations of dry matter quality, leaf nitrogen content, yield, and vegetation indices (NDVI, GNDVI, RVI, and GOSAVI) at the filling stage of different N-efficient maize varieties were all highly significant and positive. In this relationship, the best effect was found at the filling stages, with correlation coefficients reaching 0.772–0.942, 0.774–0.970, 0754–0.960, and 0.800–0.960. The results showed that the yield, dry matter weight, and leaf nitrogen content of maize varieties with different nitrogen efficiencies increased first and then stabilized with the increase in the nitrogen application level in different periods, and the highest nitrogen application level of maize yield should be between 270 and 360 kg/hm2. At the filling stage, canopy vegetation index of maize varieties with different nitrogen efficiencies was positively correlated with yield, dry matter weight, and leaf nitrogen content, especially GNDVI and GOSAVI on the leaf nitrogen content. It can be used as a means to predict its growth index.
Keywords: maize, nitrogen efficiency, vegetation indices, leaf nitrogen content, dry matter content
1. Introduction
Among the essential nutrients of plants, nitrogen is the primary chemical element that limits the biological production and yield. So nitrogen fertilizer plays an important role in crop growth. Its abundance and deficiency can affect crop metabolism, physiological characteristics, and nutrient absorption and utilization directly [1]. There were significant differences in nitrogen absorption and utilization between different N-efficient varieties of maize crop [2,3]. According to the nitrogen use efficiency (NUE) of different maize varieties, it can be divided into four types: higher grain yield under high nitrogen and low nitrogen levels, higher grain yield only under low nitrogen level, higher grain yield only under high nitrogen level, and lower than average grain yield under high nitrogen and low nitrogen levels [4,5], namely, double efficient type, low nitrogen efficient type, high nitrogen efficient type, and double inefficient type. Accurate and timely evaluation of crop growth and development as well as leaf analysis is the basis for reasonable fertilization, and it is also necessary to monitor the growth status of crops, along with [6]. Studies have shown that in wheat, rice, and other crops, leaf area, biomass, and leaf color are used to characterize the growth of crops [7,8,9]. A number of results have been achieved in the use of spectral imaging for crop growth and nutrient monitoring. The optimal spectral variables of key growth periods were used as parameters to model the N nutrition of winter wheat [7,8,9]. When using multispectral and spectral data to invert the adaptability of the maize leaf area index, it was found that the vegetation index could accurately reflect the maize leaf area index in the study area [10,11]. The results show that the spectral vegetation index (GNDVI, PSSRc, NDVI4, and DI) has higher fitting effects with the amount of aboveground dry matter in maize at different growth stages [12]. The optimal spectral vegetation index for summer maize LAC estimation at different growth stages had good results [13]. However, in the process of agricultural production, due to the complexity of the background factors for maize growth, the use of spectral vegetation index to monitor the growth status of maize varieties with different nitrogen efficiencies is rarely studied. Especially for maize hybrids with significant differences in genetic background, the single spectral vegetation index is difficult to reflect the differences in nitrogen absorption and utilization [14]. In this study, the relationship between yield, biomass, and leaf nitrogen content under different nitrogen levels was demonstrated, which provided a possibility for the optimal management of nitrogen resources in the production of maize varieties with different nitrogen efficiencies. Through the correlation between canopy vegetation index and growth index, the possibility of UAV spectroscopy to predict maize varieties and growth indexes under different nitrogen efficiencies was explored.
The trials for the present experiment were conducted in 2021 in Yuzhou, Xuchang, Henan (113°56′E, 34°09′N), China. The test sites were flat with uniform ground force. In the individual test field, the drainage and irrigation were convenient, and the planting system was two ripe wheat-maize rotations per year. The soil pH was 7.5 and the organic matter content was 24.8 g/kg. The alkaline nitrogen and the available potassium content were 98.34 and 152.22 mg/kg, respectively. While the content of available phosphorus was 30.76 mg/kg. In this experiment, four representative maize varieties with different nitrogen efficiencies were selected as the research objects. the low-N efficient varieties (Zhengdan 958, ZD958), the high-N efficient varieties (Xianyu 335, XY335), the efficient-efficient varieties (Qiule 368, QL368), and the nonefficient-nonefficient varieties (Yudan 606, YD606) were used as test materials [5,15].
1.1. Field trial design and management
A split-plot experimental design was used in this study. The five nitrogen application levels in the main zone are 0, 90, 180, 270 and 360 kg/hm2, namely N0, N90, N180, N270, and N360, respectively, for 0, 90, 180, 270, and 360 kg/hm2 N application rates. Four different nitrogen efficiency maize varieties in the sub-zone are the low-N-efficient (Zhengdan 958, ZD958), the high-N efficient (Xianyu 335, XY335), the double-high varieties (Qiule 368, QL368) and the double inefficient-type varieties (Yudan 606 YD606). The experimental planting density was 67,500 plants/hm2 with three replicates and for a total of 60 plots. The area of each plot was 6.0 m × 3.6 m (6 rows) in size. The amount of potassium and phosphorus fertilizer applied to each treatment was 150 kg/hm2. All fertilizers were applied at one time during the jointing period by artificial inter-row trenching. The trial was sown on June 15, 2021, and harvested on September 29, 2021, with other field management being the same as local fields (from Henan Xinlianxin Chemical Industry Group Co).
1.2. Ground-truth data collection
Twenty representative plants with uniform growth were selected from each plot at the jointing stage for listing, and then, three representative above-ground parts of maize plants were selected from each plot at the trumpet stage, the filling stage, and the maturity stage, including stem, sheath, leaf, bract, stem axis, and grain. After sampling, the samples were divided into two parts: stem sheath and leaf at the large bell-mouth stage; into three parts: stem sheath (stem and leaf sheath), leaf blade, and ear at the filling stage; and into four parts: stem sheath (stem and leaf sheath), leaf blade, bract shaft (bract and cob shaft), and grain at the maturity stage. The samples were packed into kraft paper bags. Then, they were put in the oven (Yamato-dec912c, made in Japan) for 30 min at 105°C, dried at 80°C to constant weight, cooled to room temperature, and weighed with a one-tenth balance (Mettler Toledo-me303e, made in China). The weighed samples were crushed by Taiste-fw100, made in China, and then, 0.1 g of the crushed and screened samples was weighed in a deboiling tube, with 5 mL of concentrated sulfuric acid for deboiling, H2O2 as catalyst, in a continuous flow analyser (AA3, SEAL-Analytical System, made in Germany) to determine the plant and grain nitrogen content [16]. In the mature stage of maize, the middle 2 rows of each plot were selected to harvest all the ears, and 10 representative ears were selected for the determination of ear length, ear diameter, row number of ears, row number of grains, 100 grain weight, grain number per ear, and grain weight per ear. Then, all the ears were threshed and the grain yield was converted according to 14% water. The data from this part were used to calculate the indicators related to dry matter and nitrogen accumulation.
1.3. The UAV platform, image acquisition and processing
The multispectral remote sensing system of the UAV was used to obtain spectral information. The system consists of a six-rotor drone, a DJI Matrice 600 Pro UAV (SZ DJI Technology Co., Shenzhen, China) (Figure 1b). This system is equipped with a MicaSense Rededge-Altum (Figure 1a). The MicaSense Rededge-Altum did collect five discontinuous spectral bands of red, green, blue, red edge, near infrared, and thermal infrared with pixel size of 3.45 μm, resolution of 2,064 × 1,544, and field of view of 48 × 36.8°.
Figure 1.
(a) A MicaSense Rededge-Altum. (b) A DJI Matrice 600 Pro UAV platform (SZ DJI Technology Co., Shenzhen, China).
In the large trumpet stage, grouting period, and mature stage, the UAV multispectral remote sensing system was used to carry out information collection. For the purpose, clear, windless weather, and a working time between 10:00 and 14:00 h were chosen. Before the flight of the drone, the route was demarcated, the gray board was used for correction, and the aircraft flight direction was set east–west, with 80% for both the heading overlap and side overlap. The acquired multispectral photos were imported into Pi×4D software in sequence for image processing and the stitched images were corrected geometrically with high-definition digital images. Finally, the band operation was performed using ENVI 5.3 to extract the parameters. Microsoft Excel was used for data sorting and SPSS 19.0 software was used for statistical analysis, inter-processing significance test (Duncan’s), and correlation.
2. Results
2.1. Effects of nitrogen supply levels on yield
It can be seen from Table 1 that the yield of maize varieties with different nitrogen efficiencies increases first with the increase in nitrogen application and then tends to be stable. Under the conditions of N0, N90, N180, N270, and N360, the coefficients of variation between grain yield varieties were 24.94, 9.37, 11.29, 11.44, and 12.91%, respectively, and the rate of nitrogen application increased first and then decreased. Under the same nitrogen fertilizer conditions, except for N90, there were significant differences in grain yield between varieties, and the yield of YD606 was the lowest. And under medium and high nitrogen conditions, the yield of QL368 was the highest. Under the treatment of N0, N90, N270, and N360, the grain yield of ZD958 was 83.15, 15.89, 11.34, and 16.83% higher, respectively, than that of YD606, the grain yield of XY335 was 28.55, 0.19, 1.33, and 7.40% higher than that of YD606, and the grain yield of QL368 was 46.70, 19.29, 27.26, and 34.26% higher than that of YD606.
Table 1.
Effects of nitrogen supply level on yield of maize varieties with different nitrogen efficiencies
| ZD958 | XY335 | YD606 | QL368 | |
|---|---|---|---|---|
| N0 | 8721a | 8327a | 7220b | 8887a |
| N90 | 9431a | 9964a | 8866a | 10362a |
| N180 | 10153b | 10364ab | 10065b | 10889a |
| N270 | 11521a | 11974a | 10724b | 11916a |
| N360 | 11454a | 11772a | 10480b | 11717a |
| Nitrogenous fertilizer (F) | 61.84* | |||
| Cultivars (C) | 19.12* | |||
Note: Different lowercase letters of different varieties indicate significant differences among varieties (P < 0.05). “*” indicates the significance at the 0.01 level.
2.2. Effects of nitrogen supply levels on the leaf nitrogen content
It can be seen that the effect of nitrogen fertilizer treatment on the leaf nitrogen content of maize at different growth stages reached a very significant level, and there were significant differences in the leaf nitrogen content of maize cultivars with different nitrogen efficiencies at the filling stage (Table 2). Under the conditions of N0, N90, N180, N270, and N360, the coefficients of variation between cultivars with the leaf nitrogen content at different growth stages were 21.63, 9.74, 5.91, 17.11, 7.10, 26.21, 13.66, 13.75, 8.62, 10.54, 18.71, 8.63, 10.70, 17.19, and 5.88%, respectively, which decreased first and then increased with the increase in nitrogen application. With the increase in nitrogen application, the leaf nitrogen content of maize varieties with different nitrogen efficiencies increased first and then was stable. Under the condition of no nitrogen application, there was no significant difference in the leaf nitrogen content of different N-efficient maize cultivars at the filling stage and the mature stage. Under the same nitrogen fertilizer condition, YD606 had the lowest nitrogen content in leaves at both the trumpet stage and the filling stage. Under high nitrogen conditions, the leaf nitrogen content of QL368 at the silking stage was significantly higher than that of ZD958, XY335, and YD606.
Table 2.
Effects of nitrogen supply level on the leaf nitrogen content of maize varieties with different nitrogen efficiencies
| N level | Cultivar | Trumpet stage (kg/hm2) | Filling stage (kg/hm2) | Maturity (kg/hm2) |
|---|---|---|---|---|
| N0 | ZD958 | 24.81bc | 27.15a | 17.97a |
| XY335 | 33.31a | 24.92a | 12.87a | |
| YD606 | 19.81c | 14.70b | 11.89a | |
| QL368 | 29.12ab | 28.44a | 14.26a | |
| F | 7.02* | 7.05* | 1.78 ns | |
| N90 | ZD958 | 48.56a | 44.64b | 27.85a |
| XY335 | 53.57a | 57.41a | 23.63a | |
| YD606 | 43.29a | 43.87b | 25.88a | |
| QL368 | 44.58a | 44.71b | 23.15a | |
| F | 1.45 ns | 6.70* | 1.00 ns | |
| N180 | ZD958 | 64.31a | 65.61ab | 29.30a |
| XY335 | 56.67a | 58.05bc | 23.79a | |
| YD606 | 57.55a | 50.93c | 30.11a | |
| QL368 | 61.43a | 69.99a | 26.10a | |
| F | 1.13 ns | 6.94* | 0.75 ns | |
| N270 | ZD958 | 57.21a | 62.67a | 35.57a |
| XY335 | 62.41a | 67.74a | 26.27bc | |
| YD606 | 69.11a | 59.95a | 29.90b | |
| QL368 | 84.12a | 72.78a | 24.22c | |
| F | 1.40 ns | 1.20 ns | 9.92** | |
| N360 | ZD958 | 60.20a | 65.22a | 29.47a |
| XY335 | 71.09a | 66.64a | 33.47a | |
| YD606 | 66.25a | 56.25a | 31.18a | |
| QL368 | 69.06a | 72.93a | 29.79a | |
| F | 1.65 ns | 1.70 ns | 1.50 ns | |
| ANOVA | C | 1.88 ns | 9.56** | 2.34 ns |
| F | 41.69** | 82.11** | 19.23** | |
| C × F | 1.52 ns | 1.38 ns | 0.77 ns |
Note: Different lowercase letters of different varieties indicate significant differences among varieties (P < 0.05). “*” indicates the significance at the 0.05 level. “**” indicates the significance at the 0.01 level.
2.3. Effects of nitrogen supply levels on dry matter accumulation
It can be seen that the effect of nitrogen fertilizer treatment on the dry matter accumulation of maize at different growth stages reached a highly significant level, and the influence of varieties on the dry matter accumulation of maize at the filling stage and the maturity stage also reached a highly significant level (Table 3). Under the conditions of N0, N90, N180, N270, and N360, the coefficients of variation between biomass cultivars at different growth stages were 17.89, 2.80, 7.88, 7.01, 11.33, 23.00, 11.79, 13.09, 5.06, 7.95, 24.25, 7.10, 13.46, 7.43, and 14.43%. Under the condition of N0, the differences among varieties at different growth stages were the greatest. With the increase in nitrogen application, the dry matter accumulation of maize varieties with different nitrogen efficiencies increased first and then decreased. Under the same nitrogen application condition, YD606 had the lowest dry matter accumulation at different growth stages, and QL368 had the highest dry matter accumulation at the filling stage and the maturity stage.
Table 3.
Effects of nitrogen supply level on dry matter content of maize varieties with different nitrogen efficiencies
| N level | Cultivar | Trumpet stage (kg/hm2) | Filling stage (kg/hm2) | Maturity (kg/hm2) |
|---|---|---|---|---|
| N0 | ZD958 | 3445.66a | 6814.35b | 9867.43ab |
| XY335 | 2811.26a | 6567.04b | 8598.79bc | |
| YD606 | 2329.08a | 4565.73c | 6476.94c | |
| QL368 | 3426.53a | 8218.65a | 11778.76a | |
| F | 2.22 ns | 16.94** | 6.64* | |
| N90 | ZD958 | 3671.97a | 10370.38ab | 13515.83a |
| XY335 | 3732.19b | 12207.77a | 14867.34a | |
| YD606 | 3507.45a | 9702.00ab | 14012.74a | |
| QL368 | 3565.53a | 9504.18b | 15872.52a | |
| F | 0.88 ns | 2.73 ns | 0.67 ns | |
| N180 | ZD958 | 4202.19a | 11171.13ab | 15204.54b |
| XY335 | 4252.36a | 11014.22ab | 13577.34b | |
| YD606 | 3587.39a | 9322.75b | 15548.42b | |
| QL368 | 3851.00a | 12878.27a | 18649.97a | |
| F | 1.10 ns | 3.32 ns | 6.164* | |
| N270 | ZD958 | 4964.10a | 11497.98a | 16391.80a |
| XY335 | 4348.66a | 12263.84a | 17278.87a | |
| YD606 | 4702.03a | 11661.05a | 15363.65a | |
| QL368 | 4275.75a | 12835.29a | 18297.20a | |
| F | 0.84 ns | 0.31 ns | 0.88 ns | |
| N360 | ZD958 | 4545.51ab | 11026.16a | 14814.16a |
| XY335 | 3991.02b | 12317.56a | 17667.53a | |
| YD606 | 4431.75b | 11703.94a | 14246.56a | |
| QL368 | 5236.37a | 13288.74a | 19278.06a | |
| F | 4.93* | 1.46 ns | 2.38 ns | |
| ANOVA | C | 2.17 ns | 5.75** | 8.45** |
| F | 17.16** | 33.21** | 26.90** | |
| C × F | 1.33 ns | 1.41 ns | 0.95 ns |
Note: Different lowercase letters of different varieties indicate significant differences among varieties (P < 0.05). “*” indicates the significance at the 0.05 level. “**” indicates the significance at the 0.01 level.
2.4. The normalized vegetation index (NDVI)
The NDVI of maize varieties with different nitrogen efficiencies decreased with the growth process, and the NDVI decreased more in the later growth period than in the early growth period (Figure 2). Under the conditions of N0, N90, N180, N270, and N360, the coefficients of variation of biomass between varieties of maize cultivars with different nitrogen efficiencies were 7.95, 4.21, 4.88, 3.49, 4.72, 14.18, 5.19, 5.30, 5.6, 5.78, 16.70, 36.81, 52.66, 75.92, and 106.64%, respectively. In addition to the maturity period, the coefficient of variation between varieties was largest under the condition of N0 and then showed a decreasing trend and finally an upward trend. Under the same nitrogen application conditions, the low-N efficient cultivar (ZD958) was the lowest. In addition to the maturity stage, the canopy vegetation index (NDVI) of maize cultivars with different nitrogen efficiencies increased first and then stabilized with the increase of nitrogen application. In addition to QL368 at the maturity stage, the maize canopy vegetation index NDVI decreased with the increase in nitrogen application.
Figure 2.
Effects of nitrogen supply level on canopy vegetation index (NDVI) of maize varieties with different nitrogen efficiencies. Note: Different lowercase letters of different varieties indicate significant differences among varieties (P < 0.05). The same as below.
2.5. The green-normalized vegetation index (GNDVI)
It can be seen that nitrogen treatments and varieties have significant effects on the GNDVI of maize varieties with different nitrogen efficiencies (Figure 3). With the progress of growth, the GNDVI of maize canopy vegetation index showed a decreasing trend, and the decreasing range was larger in the late growth period. Under the conditions of N0, N90, N180, N270, and N360, the variation coefficients of maize varieties with different nitrogen efficiencies at different growth stages were 14.39, 2.07, 5.20, and 2.71%; 2.26, 36.14, 6.71, 9.14, 9.04, and 6.62%; and 19.81, 30.43, 59.55, 72.80, and 93.03%. The GNDVI of maize canopy vegetation index increased significantly with increasing N fertilizer application compared with no N fertilizer. With the increase in the nitrogen application rate, GNDVI increased first and then stabilized with the increase in the nitrogen application rate except at the maturity stage. Under medium and high nitrogen conditions, the GNDVI of the efficient-efficient cultivar (QL368) at all growth stages was significantly higher than that of ZD958, XY335, and YD606. Under high nitrogen conditions, the GNDVI of QL368 was the highest. Except for the N0 treatment, the GNDVI of the low-N efficient cultivar (ZD958) was lower than that of other varieties at the maturity stage, and the GNDVI of QL368 was the highest. However, there was no significant difference between XY335 and YD606.
Figure 3.
Effects of nitrogen supply level on canopy vegetation index (GNDVI) of maize varieties with different nitrogen efficiencies.
2.6. The green-optimized soil conditioning vegetation index (GOSAVI)
The effect of nitrogen supply levels on the GOSAVI of different nitrogen efficiency maize varieties reached a significant level (Figure 4). With the growth period, the GOSAVI of different nitrogen efficiency maize varieties decreased, and the lower rate was greater in the late growth period. Except for the mature stage, the coefficient of variation among varieties was the largest under the condition of N0, which increased first and then decreased with the increase in the nitrogen application rate. With the increase in the nitrogen application rate, GOSAVI of maize varieties with different nitrogen efficiencies increased continuously and then tended to be stable. At maturity, GOSAVI of ZD958, XY335, and YD606 decreased with the increase in the nitrogen application rate. Under medium and high nitrogen conditions, the GOSAVI of QL368 was significantly higher than that of other cultivars at different growth stages. Under high nitrogen condition, GOSAVI of the high-N efficient varieties XY335 was significantly lower than that of other varieties during the filling stage, and there was no significant difference between ZD958 and YD606.
Figure 4.
Effects of nitrogen supply level on canopy vegetation index (GOSAVI) of maize varieties with different nitrogen efficiencies.
2.7. The ratio of vegetation index (RVI)
The effect of nitrogen supply levels on the RVI of maize varieties with different nitrogen efficiencies reached a significant level (Figure 5). During the growth process, the RVI of different nitrogen efficiency maize varieties decreased continuously, and the decline in the late growth period of maize was smaller than that in the early growth stage. In addition to the maturity period, the coefficient of variation of maize varieties decreased first and then increased with the increase in nitrogen application. The RVI of maize varieties with different N efficiencies increased with the increase in the N application rate at the big trumpet stage and the filling stage. At the maturity stage, the RVI of maize decreased with the increase in nitrogen application, except for the efficient varieties QL368. At the filling stage and the maturity stage, the RVI of QL368 was significantly higher than that of ZD958, XY335, and YD606 under medium and high nitrogen conditions, while the RVI of XY335 was significantly lower than that of other varieties. There was no significant difference in RVI between ZD958 and YD606.
Figure 5.
Effects of nitrogen supply level on canopy vegetation index (RVI) of maize varieties with different nitrogen efficiencies.
2.8. Relationships between canopy vegetation index and yield
There was a significant positive correlation between canopy vegetation index and maize yield in the trumpet opening stage and the filling stage, and the correlation was significantly higher than that in the mature stage (Table 4). There was a significant positive correlation between the variety mixing at the bell-mouth stage and the filling stage, but the correlation coefficient was lower than that of single variety. In the trumpet stage, the correlation coefficient between NDVI and yield was 0.748–0.827. The correlation coefficient between GNDVI and yield was 0.809–0.895. The correlation coefficient between RVI and yield was 0.567–0.840, and that between GOSAVI and yield was 0.848–0.937. In the filling stage, the correlation coefficient between NDVI and yield was 0.775–0.963, that between GNDVI and yield was 0.809–0.970, that between RVI and yield was 0.754–0.927, and that between GOSAVI and yield was 0.809–0.960 (Table 4).
Table 4.
Correlation between canopy vegetation index NDVI, GNDVI, RVI, GOSAVI, and yield at different growth stages
| Stage | Cultivar | NDVI | GNDVI | RVI | GOSAVI |
|---|---|---|---|---|---|
| Trumpet stage | ZD958 | 0.823** | 0.809** | 0.796** | 0.848** |
| XY335 | 0.827** | 0.835** | 0.840** | 0.886** | |
| YD606 | 0.748** | 0.895** | 0.564* | 0.937** | |
| QL368 | 0.777** | 0.857** | 0.786** | 0.857** | |
| Mixture | 0.681** | 0.810** | 0.617** | 0.794** | |
| Filling stage | ZD958 | 0.775** | 0.809** | 0.754** | 0.809** |
| XY335 | 0.829** | 0.844** | 0.826** | 0.839** | |
| YD606 | 0.963** | 0.970** | 0.927** | 0.960** | |
| QL368 | 0.917** | 0.904** | 0.877** | 0.897** | |
| Mixture | 0.655** | 0.853** | 0.787** | 0.848** | |
| Maturity stage | ZD958 | −0.806** | −0.813** | −0.823** | −0.658** |
| XY335 | −0.763** | −0.763** | −0.763** | −0.763** | |
| YD606 | −0.512 | −0.177 | −0.223 | −0.182 | |
| QL368 | 0.623* | 0.836** | 0.376 | 0.833** | |
| Mixture | 0.014 | 0.299* | 0.090 | 0.236 |
“*” indicates the significance at the 0.05 level. “**” indicates the significance at the 0.01 level.
2.9. Relationship between canopy vegetation index and leaf nitrogen content
There was a significant positive correlation between canopy vegetation index and leaf nitrogen content at the filling stage, followed by the big trumpet stage, and the lowest correlation at the maturity stage (Table 5). In the big trumpet stage and the filling stage, the correlation leaf of variety mixing showed extremely significant positive correlation, and the correlation coefficient was lower than that of single variety. At the filling stage, the correlation coefficient between NDVI and leaf nitrogen content was 0.839–0.979, that between GNDVI and leaf nitrogen content was 0.871–0.958, and that between RVI and leaf nitrogen content was 0.851–0.960. The correlation coefficient between GOSAVI and leaf nitrogen content was 0.873–0.957.
Table 5.
Correlation between canopy vegetation index NDVI, GNDVI, RVI, GOSAVI, and leaf nitrogen content in maize at different growth stages
| Stage | Cultivar | NDVI | GNDVI | RVI | GOSAVI |
|---|---|---|---|---|---|
| Trumpet stage | ZD958 | 0.827** | 0.808** | 0.814** | 0.826** |
| XY335 | 0.820** | 0.871** | 0.653** | 0.860** | |
| YD606 | 0.647** | 0.882** | 0.551* | 0.918** | |
| QL368 | 0.513 | 0.704** | 0.532* | 0.704** | |
| Mixture | 0.611** | 0.770** | 0.580** | 0.775** | |
| Filling stage | ZD958 | 0.908** | 0.939** | 0.960** | 0.939** |
| XY335 | 0.853** | 0.877** | 0.851** | 0.874** | |
| YD606 | 0.979** | 0.958** | 0.931** | 0.957** | |
| QL368 | 0.839** | 0.871** | 0.909** | 0.873** | |
| Mixture | 0.777** | 0.869** | 0.817** | 0.869** | |
| Maturity | ZD958 | 0.779** | 0.777** | 0.746** | 0.812** |
| XY335 | 0.617* | 0.733** | 0.650** | 0.733** | |
| YD606 | 0.567* | 0.761** | 0.386 | 0.876** | |
| QL368 | 0.418 | 0.622* | 0.301 | 0.613* | |
| Mixture | 0.124 | 0.314* | 0.160 | 0.297* |
“*” indicates the significance at the 0.05 level. “**” indicates the significance at the 0.01 level.
2.10. Relationship between canopy vegetation index and biomass
There was a significant positive correlation between canopy vegetation index and maize yield at the grain-filling stage, and the correlation was significantly higher than that at the trumpet stage and the maturity stage (Table 6). At the filling stage, the correlation of variety mixing was significantly positive, but the correlation coefficient was lower than that of single variety. At the filling stage, the correlation coefficient between NDVI and dry matter quality was 0.772–0.942. The correlation coefficient between GNDVI and dry matter quality was 0.774–0.927. The correlation coefficient between RVI and dry matter mass was 0.787–0.878, and that between GOSAVI and dry matter mass was 0.777–0.927.
Table 6.
Correlation between maize canopy vegetation index NDVI, GNDVI, RVI, GOSAVI, and dry matter quality in different growth stages
| Stage | Cultivar | NDVI | GNDVI | RVI | GOSAVI |
|---|---|---|---|---|---|
| Trumpet stage | ZD958 | 0.843** | 0.823** | 0.806** | 0.847** |
| XY335 | 0.774** | 0.802** | 0.787** | 0.800** | |
| YD606 | 0.907** | 0.882** | 0.878** | 0.884** | |
| QL368 | 0.404 | 0.736** | 0.462 | 0.736** | |
| Mixture | 0.546** | 0.675** | 0.500** | 0.666** | |
| Filling stage | ZD958 | 0.942** | 0.927** | 0.828** | 0.927** |
| XY335 | 0.774** | 0.802** | 0.787** | 0.800** | |
| YD606 | 0.907** | 0.882** | 0.878** | 0.884** | |
| QL368 | 0.772** | 0.774** | 0.854** | 0.770** | |
| Mixture | 0.742** | 0.793** | 0.743** | 0.792** | |
| Maturity | ZD958 | −0.516* | −0.705** | −0.702** | −0.378 |
| XY335 | −0.681** | −0.328 | −0.681** | −0.280 | |
| YD606 | −0.397 | −0.055 | −0.172 | −0.062 | |
| QL368 | 0.481 | 0.662** | 0.203 | 0.654** | |
| Mixture | 0.124 | 0.314* | 0.161 | 0.297* |
“*” indicates the significance at the 0.05 level. “**” indicates the significance at the 0.01 level.
3. Discussion
With the increase in nitrogen application, the yield, dry matter accumulation, and nitrogen accumulation of maize showed a trend of first increasing and then stabilizing, and the high-efficiency varieties were significantly higher than the nitrogen-inefficient varieties [17]. This conclusion is similar to the results of this study. Previous studies showed that both nitrogen application level and variety affected crop canopy reflectance [18,19]. The application of nitrogen fertilizer increased the near-infrared band reflectance of the rice canopy and reduced the reflectivity of the visible light band [20,21]. Increasing N fertilizer application could increase NDVI, GNDVI, RVI, and GOSAVI of rice (Table 7). Fu et al. also found that leaf N content of rice was positively correlated with canopy vegetation indices ndvi, rvi, dvi, and gndvi [22]. The results of this study showed that the differences of dry matter weight and leaf nitrogen concentration of maize varieties with different nitrogen efficiencies at different nitrogen application levels were consistent with their own variety characteristics. With the increase in the nitrogen application level, the yield, dry matter weight, and leaf nitrogen content of maize varieties with different nitrogen efficiencies increased first and then stabilized. The canopy vegetation index (NDVI, GNDVI, GOSAVI, RVI) and growth index of maize varieties at different stages and with different nitrogen efficiencies were significantly different, and the canopy vegetation index (NDVI, GNDVI, GOSAVI, RVI) and growth index increased with the increase in the nitrogen application rate, which indicated the possibility of canopy vegetation index predicting growth index. Ryckewaert et al. analyzed the spectral data of different genotypes of maize varieties under water stress and screened out the more drought-tolerant maize varieties [23]. Fan et al. showed that nitrogen application rate had different effects on canopy vegetation index of different maize varieties [24]. The results showed that there was no significant difference in canopy vegetation index of maize varieties with different N efficiencies under low N conditions. With the increase in the nitrogen application rate, the difference of canopy vegetation index of maize varieties with different nitrogen efficiencies increased. QL368 had the highest canopy vegetation indices (NDVI, GNDVI, RVI, and GOSAVI) under medium and high nitrogen conditions, and its yield, dry matter weight, and leaf nitrogen content were also higher than those of ZD958, XY335, and YD606. The results indicated that maize varieties with high nitrogen efficiency had strong greenness, higher contents of chlorophyll A, chlorophyll B, and carotenoid in leaves, and higher absorption rates of visible light, especially red, blue, and violet light, resulting in higher canopy vegetation indices NDVI, GNDVI, RVI, and GOSAVI, and higher response to nitrogen supply, strong material production capacity, and higher yield [25,26]. This indicates that index does not have the ability to screen high-efficiency maize varieties under low nitrogen condition and canopy vegetation, while canopy vegetation index provides the possibility to screen high-efficiency maize varieties under non-low nitrogen condition. There is a correlation between canopy vegetation index and crop yield, dry matter quality, and leaf nitrogen content. Phyu et al. found that NDVI was significantly associated with rice yields [27]. Niu et al. found that the vegetation indices derived from UAV RGB imagery had great potential to estimate maize above-ground biomass with R 2 values ranging from 0.63 to 0.73. When estimating the above-ground biomass of maize by using multivariable linear regression based on vegetation indices, a higher correlation was obtained with an R 2 value of 0.82 [28]. Panek et al. found that early spring NDVI in Croatia, the Czech Republic, Germany, Hungary, Latvia, Lithuania, Poland, and Slovakia was very closely related to cereal yields [29]. The results of this study showed that maize yield, dry matter mass, and leaf nitrogen content had similar trends with canopy vegetation indices (NDVI, GNDVI, RVI, and GOSAVI) among different nitrogen levels or maize varieties. Therefore, this experiment analyzed the correlation between yield (Table 4), dry matter mass (Table 6), leaf nitrogen content (Table 5), and canopy vegetation index at the trumpet stage, the filling stage, and the maturity stage. The results showed that maize yield, dry matter mass, and leaf nitrogen content were significantly positively correlated with canopy vegetation index at the filling stage, and the correlation coefficient was higher than that at the trumpet stage and the mature stage. The correlation coefficient between leaf nitrogen content and four canopy vegetation indices was higher than that of yield and dry matter mass at the filling stage (Table 5). However, the correlation between GNDVI and GOSAVI and leaf nitrogen content, yield, and dry matter quality in the canopy vegetation index at the filling stage was stronger than that of other vegetation indexes. Therefore, it provided the possibility for the canopy vegetation index of maize varieties with different nitrogen efficiencies to predict the leaf nitrogen content, yield, and dry matter quality at the filling stage, and the prediction of the leaf nitrogen content might be more accurate.
Table 7.
Vegetation indices involved in this study
4. Conclusion
The results showed that the yield, dry matter weight, and leaf nitrogen content of maize varieties with different nitrogen efficiencies increased first and then stabilized with the increase in the nitrogen application level in different periods, and the highest nitrogen application level of maize yield should be between 270 and 360 kg/hm2. At the filling stage, canopy vegetation index of maize varieties with different nitrogen efficiencies was positively correlated with yield, dry matter weight, and leaf nitrogen content, especially GNDVI and GOSAVI on the leaf nitrogen content. It can be used as a means to predict its growth index.
Footnotes
Funding information: This work was financially supported by the National Key Research and Development Program of China (2022YFD2300805-2) and Basic Scientific Research project of Henan Academy of Agricultural Sciences (2022JC09) and Project of the Shanghai Cooperation Organization Academy of Modern Agriculture (SCO21B003), supported by the Cooperation project of Henan Province.
Author contributions: Xia Zhao: conceptualization and data curation and formal analysis and project administration and supervision and writing – review & editing; Shuaili Wang: writing – original draft; Tao Wen: writing – review & editing; Jiamin Xu & Gangqiang Gao: data curation and writing – review & editing; Bao Huang & Jiangfang Qiao: formal analysis; Shufeng Yan: data curation and writing – review & editing; Jinliang Yang & Tianxue Liu: writing – review & editing; Lianhai Wu: data curation. Yali Zhao & Hongping Li: data curation; Hongwei Wang: application of results; Xinyuan Mu: data curation and formal analysis; writing – review & editing.
Conflict of interest: Authors state no conflict of interest.
Data availability statement: The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Contributor Information
Xia Zhao, Email: zhaoxia1007@126.com.
ShuaiLi Wang, Email: 18438616519@163.com.
Tao Wen, Email: wentao18317765003@163.com.
Tianxue Liu, Email: tianxueliu2005@163.com.
Xinyuan Mu, Email: muxinyuan@163.com.
References
- [1].Liu BF, Li J, He J, Shi ZJ. Estimation models of above-ground dry matter accumulation of summer maize based on hyperspectral remote sensing vegetation indexes. Trans Chin Soc Agric Machinery. 2016;47(3):254–62. 10.6041/j.issn.1000-1298.2016.03.036. [DOI]
- [2].Chen YL, Zhang J, Li Q, He XL, Su XP, Chen FJ, et al. Effects of nitrogen application on post-silking root senescence and yield of maize. Agron J. 2015;107(3):835–42. 10.2134/agronj14.0509. [DOI]
- [3].Mi GH. Nutrient use efficiency in crops and its genetic improvemen. J Plant Nutr Fertilizers. 2017;23(6):1525–35. 10.11674/zwyf.17283. [DOI]
- [4].Chen F, Fang Z, Gao Q, Ye Y, Jia L, Yuan L, et al. Evaluation of the yield and nitrogen use efficiency of the dominant maize hybrids grown in North and Northeast China. Sci China Life Sci. 2013;56(6):552–60. 10.1007/s11427-013-4462-8. [DOI] [PubMed]
- [5].Yang YL, Zhao X, Wang SL, Xu JM, Mu XY, Xing RR, et al. Screening and yield character analysis of maize nitrogen efficient varieties in south and central Huang-Huai-Hai Area. J Maize Sci. 2022;30(1):23–32. 10.13597/j.cnki.maize.science.20220104. [DOI]
- [6].Kross A, Mcnairn H, Lapen D, Sunohara M, Champagne C. Assessment of rapideye vegetation indices for estimation of leaf area index and biomass in corn and soybean crops. Int J Appl Earth Obs Geoinf. 2015;34(1):235–48. 10.1016/j.jag.2014.08.002. [DOI]
- [7].Donga T, Liua J, Qiana B, Heb L, Liub J, Wangb R, et al. Estimating crop biomass using leaf area index derived from landsat 8 and sentinel-2 data. ISPRS J Photogramm Remote Sens. 2020;168:236–50. 10.1016/j.isprsjprs.2020.08.003. [DOI]
- [8].Liu CH, Wang Z, Chen ZC, Zhou L, Rui XZ, Miao YX. Nitrogen monitoring of winter wheat based on unmanned aerial vehicle remote sensing image. Trans Chin Soc Agric Machinery. 2018;49(6):207–14. 10.6041/j.issn.1000-1298.2018.06.024. [DOI]
- [9].Zhou X, Zheng HB, Xu XQ, He JY, Ge XK, Yao X, et al. Predicting grain yield in rice using multi-temporal vegetation indices from uav-based multispectral and digital imagery. ISPRS J Photogramm Remote Sens. 2017;130:246–55. 10.1016/j.isprsjprs.2017.05.003. [DOI]
- [10].Mananze S, Pôças I, Cunha M. Retrieval of maize leaf area index using hyperspectral and multispectral data. Remote Sens. 2018;10(12):1942. 10.3390/rs10121942. [DOI]
- [11].Wang LG, Xu JH, He J, Li B, Yang XZ, Wang LJ, et al. Estimating leaf area index and yield of maize based on remote sensing by unmanned aerial vehicle. J Maize Sci. 2020;28(6):88–93. 10.13597/j.cnki.maize.science.20200613. [DOI]
- [12].Liu L, Peng ZO, Zhang BZ, Wei Z, Han NN, Lin SZ, et al. Canopy nitrogen concentration monitoring techniques of summer corn based on canopy spectral information. Sensors. 2019;19(19):4123. 10.3390/s19194123. [DOI] [PMC free article] [PubMed]
- [13].Wei TF, Xu XG, Li ZY, Yang GJ, Li ZH, Feng HK, et al. Remote sensing estimation of nitrogen content in summer maize leaves based on multispectral images of UAV. Trans Chin Soc Agric Eng. 2019;35(8):126–33. 10.11975/j.issn.1002-6819.2019.08.015. [DOI]
- [14].Zhou LL, Feng HY, Yan ZM, Liu, K, Zhou SL. Hyperspectral diagnosis of leaf N concentration of maize and variety difference. Trans Chin Soc Agric Eng. 2010;26(8):195–9. 10.3969/j.issn.1002-6819.2010.08.033. [DOI]
- [15].Zhao X, Wang HR, Zheng YZ, Mu XY, Yaang YL, Fu J, et al. Study on the uptake and utilization Efficiency of nitrogen,phosphorus and potassium of maize in lime concretion black soils. J Maize Sci. 2021;29(6):144–50. 10.13597/j.cnki.maize.science.20210620. [DOI]
- [16].Ding XP, Li GH, Zhang JW, Liu P, Ren BC, Zhao B. Effects of base application depths of controlled release urea on yield and nitrogen utilization of summer maize. Sci Agric Sin. 2020;53(21):4342–54. 10.3864/j.issn.0578-1752.2020.21.004. [DOI]
- [17].Chang X, Wang XB, Wu M, Yang ZS, Li J. Effects of nitrogen fertilizer on the yield,dry matter and nitrogen accumulation of maize inbred lines with different nitrogen use efficiency types. Soil Fert Sci China. 2021;2:221–7. 10.11838/sfsc.1673-6257.20075. [DOI]
- [18].Li YX, Zhu Y, Cao WX. Characterizing Canopy Hyperspectral and Multispectral Reflectance under Different N-application Conditions in Wheat. J Triticeae Crop. 2006;26(2):103–8. 10.3969/j.issn.1009-1041.2006.02.022. [DOI]
- [19].Liu TJ, Pan XZ, Tang JJ, Zhu B, Hu WJ, Zhang XD, et al. An analysis on hyper-spectral characteristics of rice under different nitrogen levels. Acta Agric Univ Jiangxiensis. 2014;36(4):699–704. 10.3969/j.issn.1000-2286.2014.04.001. [DOI]
- [20].Xue LH, Cao WX, Luo WH. Canopy spectral reflectance characteristics of rice at elongating stage under different cultivated practice and their fuzzy cluster Analysis. Chinese. J Rice Sci. 2004;18(2):65–9. 10.3321/j.issn:1001-7216.2004.02.013. [DOI]
- [21].Buchaillot ML, Gracia-Romero A, Vergara-Diaz O, Zaman-Allah M, Tarekegne A, Cairns J, et al. Evaluating maize genotype performance under low nitrogen conditions using rgb uav phenotyping techniques. Sensors. 2019;19(8):1815. 10.3390/s19081815. [DOI] [PMC free article] [PubMed]
- [22].Fu YQ, Zhong XH, Huang NR, Pan JF, Hu XY, Hu R, et al.. Research on relationship between canopy spectral characteristic and nitrogen nutrient of rice cultivars based on UAV multispectral remote sensing. Guangdong Agricultural Science. 2021;48(10):121–31. 10.16768/j.issn.1004-874X.2021.10.013. [DOI]
- [23].Ryckewaert M, Gorretta N, Henriot F, Gobrecht A, Hran D, Moura D, et al. Potential of high-spectral resolution for field phenotyping in plant breeding: Application to maize under water stress. Comput Electron Agriculture. 2021;189:106385. 10.1016/j.compag.2021.106385. [DOI]
- [24].Fan HY, Li YY, Lu XJ, Gu SH, Guo XY, Liu YH. Comparative analysis of lAI and above-ground biomass estimation models based on UAV multispectral remote sensing. J Plant Nutr Fertilizers. 2021;23(9):112–20. 10.13304/j.nykjdb.2020.0173. [DOI]
- [25].Chen JG, Feng YS, Yang ZC, Zhou YJ, Chen YL. The main work is to study the characteristics of light spectrum of objects. J Jilin Agric Univ. 1984;3:68–74. 10.13327/j.jjlau.1984.03.014. [DOI]
- [26].Echarte L, Rothstein S, Tollenaar M. The response of leaf photosynthesis and dry matter accumulation to nitrogen supply in an older and a newer maize hybrid. Crop Sci. 2009;48(2):656–65. 10.2135/cropsci2007.06.0366. [DOI]
- [27].Phyu P, Islam M, Sta Cruz P, Collard BC, Kato Y. Use of ndvi for indirect selection of high yield in tropical rice breeding. Euphytica. 2020;216(5):1–9. 10.1007/s10681-020-02598-7. [DOI]
- [28].Niu YX, Zhang LY, Zhang HH, Han WT, Peng XS. Estimating above-bround biomass of maize using features derived from uav-based rgb imagery. Remote Sens. 2019;11(11):1261. 10.3390/rs11111261. [DOI]
- [29].Panek E, Gozdowski D, Acutis M. Relationship between modis derived ndvi and yield of cereals for selected european countries. Agronomy. 2021;11(2):340. 10.3390/agronomy11020340. [DOI]
- [30].Tůma L, Kumhálová J, Kumhála F, Krepl V. The noise-reduction potential of radar vegetation index for crop management in the czech republic. Precis Agric. 2022;23(2):450–67.