Effects of Reducing Chemical Fertilisers Application on Tea Production and Soils Quality: An In Situ Field Experiment in Jiangsu, China

Abstract

In order to achieve sustainable development of the tea industry in China, it is necessary to reduce the use of chemical fertiliser rationally. With conventional fertilisation (CF) treatment as the control, five different chemical fertiliser-reduced regimes, including tea-specific formula fertiliser (T1), T1 + acidification amendment (T2), organic substitution based on T1 (T3), urea formaldehyde slow-release fertiliser (T4) and carbon-based organic fertiliser (T5), were conducted and evaluated on a green tea plantation from 2018 to 2021. The results showed that the spring tea yield of T1–T5 increased by 4.65–28.67%, while the free amino acids, tea polyphenols and sensory evaluation scores did not remarkably decrease. In addition, the T1–T5 treatments had a slight effect on soil acidification mitigation (except T2) and maintained the essential nutrients for tea production. Nutrient use efficiency improved, with agronomic efficiency (AE) increasing by 0.01–0.08 kg kg⁻¹, shoot nutrient use efficiency (NUE) by 0.14–0.70% and partial factor productivity (PFP) by 0.05–0.18 kg kg⁻¹. The net economic benefits also improved, with T1 showing a 135.28% increase, followed by T3 (67.53%), T2 (48.65%), T4 (38.07%) and T5 (33.35%). Overall, our results indicated that the T1 treatment could maintain the tea yield and quality while reducing the chemical fertiliser input and maximising the net economic benefit and AE.

1. Introduction

Tea (Camellia sinensis L.) is one of the most important economic and beverage crops in the world. The high abundance of health-promoting and pleasant flavour metabolites (including catechins, caffeine, aroma and amino acids) in tea makes it one of the most popular beverages [1]. To obtain high tea yields and economic benefits, excessive and unbalanced fertilisers have been applied in most tea production areas. The published literature has reported that 30% of tea plantations are over-fertilised with nitrogen (N) fertilisers, and 80% of tea plantations are fertilised with unreasonable N-P₂O₅-K₂O ratios [2,3], which have caused many adverse effects, including soil acidification, nitrate leaching risk, greenhouse gas nitrous oxide emissions and tea quality reduction [4,5,6,7]. Therefore, it is essential and urgent to reduce the chemical fertiliser input to achieve the goal of a sustainable tea industry.

Related studies have well demonstrated that N, P and K fertilisers play important roles in tea yields and in the formation of flavour metabolites. For instance, N plus K fertilisation can increase the chlorophyll content, as well as root and shoot biomass [8,9,10]. N fertilisation mainly contributes to the increase in the content of free amino acids (AA), total polyphenols (TP) and aroma [11,12,13,14], and with chemical N fertilisation rates rising, the TP/AA ratio is reduced [9,15]. Moreover, the genes and enzyme activity related to AA and catechins biosynthesis that have been reported can be regulated by the N input rate and different N forms [9,16]. P and K applications also exert positive effects on improving the growth and quality of tea plants. In a pot experiment, an exogenous K supply increased the TP content significantly [17]. K application in field experiments increased the tea yield significantly and improved the AA and TP contents in shoots [18]. High P and K fertiliser inputs were beneficial for the accumulation of sugars and catechins in tea shoots [19].

Nevertheless, excessive use of single-element fertilisers or the application of N, P and K fertilisers in inappropriate ratios could also result in adverse effects, including soil degradation, acidification, increase in environmental risk and economic benefits decline. Ruan et al. [20] revealed that soil pH decreases with the increase of N fertiliser input, and the decrease in soil pH is unfavourable to the growth and nitrogen absorption of tea plants, because inappropriate root zone pH could result in the declining absorption of N, and in the meantime, the accumulation of nitrate under low pH conditions will also limit the root growth. In addition, long-term field experiments have found that excessive N input has adverse effects, e.g., declining in the diversity of soil microorganisms, shifting in the community composition and decreasing in the soil quality [21,22,23], though tea growth and quality are improved [23]. Meanwhile, the tea yield is not always directly proportional to the input of the fertiliser rates. Many studies have reported that excessive N supply may lead to a decrease in tea yield [24,25]. A recent study also showed that overapplication of P and K fertilisers could inhibit AA accumulation but promote the metabolism of flavonoid [19]. Therefore, exploring the appropriate ratios of N, P and K inputs has become an urgent issue that must be solved for reaching the goal of sustainable development of the tea industry.

Field investigations have suggested the optimal ratio of N, P and K application for premium green tea production, i.e., 200–300 kg N ha⁻¹, 60–90 kg P₂O₅ ha⁻¹ and 60–120 kg K₂O ha⁻¹ [3,23,26]. Also, tea-specific formula fertiliser (N-P₂O₅-K₂O = 18-8-12) (T1) has been developed based on the database of the tea plantation soil nutrient status and the nutrient requirement of tea plants [3,27]. Based on these foundation, an in situ field experiment with seven different fertilisation regimes (i.e., no fertilisation (CK), conventional fertilisation (CF), tea-specific formula fertiliser (T1), T1 + acidification amendment (T2), organic substitution based on T1 (T3), urea formaldehyde slow-release fertiliser (T4) and carbon-based organic fertiliser (T5)) was carried out to address the following questions: (1) whether chemical fertiliser reduction will affect the tea yield, quality, soil fertility and economic benefit compared to conventional farmer practices and (2) exploring the optimal fertilisation regime for premium green tea production on the basis of the yield, quality, economic benefit, cost efficiency and environmental safety.

2. Materials and Methods

2.1. Site Description and the Field Experiment

The experiment field was located in Jiangsu Tea Expo Garden of Jurong City, Jiangsu Province (31°55′23″ N, 119°16′17″ E, ~17 m above sea level). The area was dominated by plain land, and the soil type was characterised as Alfisol. The annual mean temperature is 15.2 °C; the annual average precipitation is 1058.8 mm, which belongs to the central north subtropical climate zone. The tea tree variety in the plantation used in this experiment was ‘Fudingdahao’, about 30 years old. Prior to the experiment, the basic properties of the surface soil (0–20 cm) were soil pH 5.42, soil organic matter 22.71 g kg⁻¹, soil total N 0.88 g N kg⁻¹, available P 176.94 mg P kg⁻¹ and available K 163.46 mg K kg⁻¹.

The field experiment was initiated in 2018 (September) and ended in 2021 (April). The specific fertiliser input of each treatment is displayed in

Table 1. Integrated scheme of fertiliser reduction technology for the tea garden in Jiangsu Province.

Treatment	Organic Fertiliser (kg ha−1)			Chemical Fertiliser (kg ha−1)				Total NPK	Chemical Fertilisation Reduction Ratio (%)
	N	P2O5	K2O	N	P2O5	K2O	Sum
CK	0	0	0	0	0	0	0	0
CF	64	38	26	675	345	345	1365	1493
T1	0	0	0	303	60	90	453	453	66.8
T2	0	0	0	303	60	114	477	477	65.1
T3	90	270	150	212	60	90	362	871	73.5
T4	0	0	0	300	80	110	490	490	64.1
T5	0	0	0	270	135	180	585	585	57.1

CK, no fertiliser; CF, conventional fertiliser; T1, tea-specific formula fertiliser; T2, T1 + acidification amendment; T3, organic substitution based on T1; T4, urea formaldehyde slow-release fertiliser; T5, carbon-based organic fertiliser.

. The tea trees were planted in single rows of 1.5 m wide and 51.0 m long. The total area of the experimental tea garden was about 0.83 ha, with the area of the trial plot ranging from 0.09 to 0.15 ha. The nutrient contents for the fertilisers were as follows: tea-specific formula fertiliser (N-P₂O₅-K₂O = 18-8-12), soil acidification ameliorant (K₂O = 4%), microbial organic fertiliser (N-P₂O₅-K₂O = 3-9-5), urea formaldehyde slow-release fertiliser (N-P₂O₅-K₂O = 30-5-6), carbon-based compound fertiliser (N-P₂O₅-K₂O = 6-3-4), urea (N = 46%), rapeseed cake (N-P₂O₅-K₂O = 5-3-2) and compound fertiliser (N-P₂O₅-K₂O = 15-15-15). All the base fertilisers were applied in late September or early October, and the topdressing was applied about one month before the spring tea plucking. Topdressing was only applied in CF, T1, T2 and T3, with 330, 168, 168 and 77 kg N ha⁻¹ in the form of urea. Rapeseed cake and microbial organic fertiliser were used for CF and T3, respectively.

2.2. Tea Yield and Quality Determination

In the spring of 2019–2021, single tea buds were collected for yield measurement. Germination density and 100-bud weight were also investigated for yield measurement. In brief, six rows (2 m long) were selected randomly and investigated in each treatment in the spring. The germination density was counted in 6 sampling boxes (33 cm × 33 cm), according to the method of Li et al. [28]. The fresh buds in each row were weighed immediately and then fixed by microwave. The fixed buds were dried to a constant weight in the oven and weighed using an analytical balance, and the 100-bud weight of the dried buds was counted and weighed.

In the tea-plucking season, all batches of tea buds were picked. The yield of the fresh tea buds was determined by accumulating all batches of plucked fresh buds in 2-m-long tea rows. Then, the first batch of samples was ground into powder for a quality analysis. The later batches (~400 g) were processed into green tea for the sensory evaluation. According to the method specified in China National Standard GB/T 23776-2018 [29], the processed tea was evaluated and scored by appearance, colour of the tea water, fragrance, flavour and infused leaf, and the final score of each processed tea was calculated by the weighted average, in which 25% accounted for the dry tea colour, 10% for the infusion colour, 25% for aroma, 30% for taste and 10% for infused leaves.

For the TP and caffeine content analysis, 0.2 g tea bud powders were added into a 10 mL centrifuge tube with 5 mL 70% methanol under a 70 °C water bath for 5 min, then the mixtures were centrifuged at 3500× g for 5 min, and the above supernatant was collected and diluted to measure the TP content by ferric tartrate colourimetry.

For the AA content determination, 0.10 g tea bud powders were extracted by 5 mL distilled water for 5 min under a 100 °C water bath, then the mixtures were centrifuged at 3500× g for 5 min. The above supernatant was collected and used for AA content detection by the ninhydrin colourimetry method.

Caffeine content was analysed using HPLC (1260 Infinity II, Agilent, Palo Alto, CA, USA) by GB/T 8312-2013 [30]. Water extract content was analysed by the method in GB/T 8305-2013 [31].

The tea bud powders were digested with HNO₃:HClO₄ (5:1), and the P, K and Mg contents were detected by ICP-AES. The total N (TN) and carbon (TC) of each sample were measured by the element analyser (Vario Max, Elementar, Langenselbold, Germany).

2.3. Soil Sampling and Measurements

Before base fertilisation, the 0–20 cm depth soil samples were taken from four points on each treatment using soil augers. Then, the soils were mixed to make a composite sample for later analysis. The collected soil samples were sieved through a 2 mm mesh, then dried and ground for the following soil properties analysis.

Soil pH was determined in 1:2.5 (w/v) soil:deionised water suspension with a pH meter (ORION 3 STAR, Thermo Fisher, Waltham, MA, USA). Soil organic carbon (SOC) and TN were measured by the element analyser (Vario Max, Elementar, Germany). Soil organic matter (SOM) was calculated from the SOC by a conversion factor of 1.724. Soil available P (AP) and available K (AK) were extracted using the Mehlich 3 method [32] and then measured using inductive coupled plasma emission spectrometer (ICAP6300, Thermo Fisher, USA). The soil chemical properties mentioned in this section were conducted following the methods in the soil analysis handbook [33]. Determination of heavy metals in the soil was completed by reference to the Chinese National Standard procedures GB/T 22105.2-2008 (arsenic, As) [34], GB/T 22105.3-2008 (lead, Pb) [35], GB/T 17141 (cadmium, Cd) [36], GB/T 17137 (chromium, Cr) [37] and GB/T 17138 (copper, Cu) [38].

The soil fertility index (SFI) in this study was calculated according to the method of Guo et al. [39]. First, the weight of each soil factor was calculated, and then, the score of each soil factor was calculated using the standard score function. Then, the fertility index was calculated by using the following equation:

$$\mathrm{S}\mathrm{F}\mathrm{I}=\sum _{i=1}^{n}Wi \ast Si$$

where W is the weight value of each soil parameter, S is the score of each parameter and n is the number of parameters [40].

2.4. Fertiliser Use Efficiency and Economic Benefit

Fertiliser NPK utilisation efficiency was represented by agronomic efficiency (AE), nutrient utilisation efficiency (NUE) of tea buds and partial factor productivity (PFP), which were calculated using the following formulas:

$$AE={\displaystyle \frac{Y-Y0}{NPK rate}}$$

$$NUE={\displaystyle \frac{NPK-NPK0}{NPK rate}} \ast 100\%$$

$$PFP={\displaystyle \frac{Y}{NPK rate}}$$

Y and Y0 are the tea yields of the fertilisation treatments and CK, NPK and NPK0 are the NPK uptake in the tea buds of the fertilisation treatments and CK and NPK rate is the fertiliser NPK application rate.

The economic benefits were evaluated by input cost, gross income and net income, and they were calculated using the following formulas:

$$Gross income=Tea price \ast Yield$$

$$Input cost=Fertilizer cost+Labor cost$$

$$Net income=Gross income-Input cost$$

Labor cost mainly includes picking, trimming, ploughing and fertilisation costs. Tea price was determined according to the local market price.

2.5. Statistical Analysis

Microsoft Excel 2016 (Microsoft Corporation, Redmond, WA, USA) and SPSS 18.0 (SPSS Corp., Chicago, IL, USA) were used for statistical analysis of the measured sample data. Normal distribution of the data was tested using the Kolmogorov–Smirnov test (p = 0.05). The significant difference between the NPK rates treatments was tested with SPSS 18.0 using one-way analysis of variance (ANOVA), with Duncan’s post hoc test (p < 0.05). Regression analysis of the yield and bud density with NPK rates, box plots and radar map were made by Excel 2016. The radar map was made based on six parameters, including fertiliser reduction, tea yield, tea quality, SFI, AE and economic benefit. The data were standardised by the parameter/total score of each item.

3. Results

3.1. Tea Yields and Quality Response to Different Fertilisation Regimes

During the experiment, CK treatment revealed the lowest yield compared to the other treatments, and fertilisation can increase the tea yield by 19.90–54.28% compared to CK (Figure 1A). All the fertilisation reduction treatments, i.e., T1–T5, showed an obvious improve in tea yield compared to CF; specifically, the yields of T1, T2, T3 and T5 increased by 28.67%, 14.10%, 22.58% and 23.86%, respectively, compared to that of CF (Figure 1A). For every single year, the yield of CF increased by 2.44–42.66% when compared to the CK treatment, while the yields of the T1–T5 treatments increased by 3.42–51.65% when compared to the CF treatment (except the yield of T4 in 2020) (Figure S1). As for the average bud density, T1–T5 produced more numbers of buds than CK and CF in 2019–2021, and T1 and T3 performed better than the other treatments (Figure 1B). In 2019 and 2021, the bud density of T3 and T5 was higher than that of CF (p < 0.05), whereas T1–T5 performed better than CK and CF (p < 0.05) in 2020 (Figure S1). Moreover, no significant difference was observed in 100-bud weight between the CF and fertiliser reduction regimes T1–T5 (Table S1). In addition, the yield (R² = 0.07, p < 0.05) and bud density (R² = 0.26, p < 0.001) showed a quadratic response to increasing the NPK rates, and when the NPK rates were 770.00 kg ha⁻¹ and 955.50 kg ha⁻¹, the tea yield and bud density reached the peak values, respectively (Figure 1C,D).

For the tea quality parameters (

Table 2. Average tea quality parameters under different fertilisation regimes in from 2019 to 2021.

Treatment	AA (%)	TP (%)	TP/AA	Water Extract (%)
CK	3.31 ± 1.25	23.02 ± 2.15a	7.13 ± 2.30a	42.61 ± 2.81
CF	4.25 ± 1.30	21.32 ± 1.28b	5.46 ± 1.66b	41.47 ± 3.22
T1	3.82 ± 1.31	22.42 ± 1.93ab	6.49 ± 2.06ab	43.95 ± 3.53
T2	3.55 ± 1.26	23.00 ± 2.05a	7.35 ± 2.66a	43.98 ± 3.36
T3	3.59 ± 1.21	22.96 ± 2.00a	7.06 ± 2.17ab	42.97 ± 3.89
T4	3.89 ± 1.14	22.66 ± 1.98a	6.39 ± 2.18ab	42.56 ± 4.64
T5	3.90 ± 1.41	22.81 ± 1.91a	6.54 ± 2.12ab	41.80 ± 4.45

CK, no fertiliser; CF, conventional fertiliser; T1, tea-specific formula fertiliser; T2, T1 + acidification amendment; T3, organic substitution based on T1; T4, urea formaldehyde slow-release fertiliser; T5, carbon-based organic fertiliser. AA, free amino acids; TP, total polyphenols. Values shown are the means ± SD (n = 18). Different letters in the same column represent significant differences among treatments at a significance level of p < 0.05.

), CF exhibited the highest value of AA content and lowest value of TP/AA ratio. The AA contents of CF and T1–T5 were elevated by 2.54–21.35% when compared to the CK treatment, while the TP/AA ratio decreased by 7.57–16.95%. Moreover, all the fertilisation treatments showed a slight decrease in TP content compared to the CK treatment. There was no obvious change in water extracts among the different treatments, as well as the sensory evaluation score of processed tea (Table S2).

The elemental composition of the tea buds was analysed across different fertilisation regimes in 2019 and 2020 (Table S3). The TC and P contents remained consistent across treatments. Fertilisation generally increased the TN content, with treatments T1, T4 and T5 showing significant increases compared to CF in 2019 (p < 0.05). The K content decreased in CF and T1–T5 compared to the CK treatment in 2019 (p < 0.05), with similar trends observed in 2020, particularly in T2 and T5. The Mg content exhibited variable responses: in 2019, it decreased significantly in T1 compared to CK but increased in T3 and T5 relative to CF (p < 0.05). In 2020, the Mg content increased across all fertilisation treatments, with significant elevations in T2–T5 (p < 0.05). No significant differences were observed between fertiliser reduction regimes.

3.2. Soil Properties Variations under Different Fertilisation Regimes

In order to assess the effects of fertilisation on soil fertility, the soil properties of different fertilisation regimes were detected in 2020. The CK and T2 treatments showed the highest soil pH and were significantly higher than in the other treatments (p < 0.05) (

Table 3. Soil properties in different fertilisation regimes.

Treatment	pH	TN (g kg−1)	SOM (g kg−1)	AP (mg kg−1)	AK (mg kg−1)	SFI
CK	5.31 ± 0.39a	0.96 ± 0.14c	15.36 ± 1.75c	45.09 ± 25.51c	118.25 ± 25.79b	0.50 ± 0.14b
CF	4.49 ± 0.16b	1.59 ± 0.18a	27.03 ± 3.33a	149.32 ± 27.63a	208.75 ± 10.72a	0.88 ± 0.08a
T1	4.86 ± 0.22b	1.28 ± 0.15abc	21.30 ± 3.04ab	98.08 ± 26.98b	212.5 ± 29.44a	0.74 ± 0.07a
T2	5.32 ± 0.31a	1.37 ± 0.34ab	22.22 ± 5.28ab	79.84 ± 44.91bc	235.25 ± 49.92a	0.77 ± 0.14a
T3	4.64 ± 0.07b	1.35 ± 0.22ab	23.03 ± 3.35ab	62.13 ± 28.28bc	179.00 ± 6.93a	0.76 ± 0.12a
T4	4.65 ± 0.16b	1.37 ± 0.23ab	22.53 ± 3.44ab	73.18 ± 24.10bc	184.75 ± 50.71a	0.77 ± 0.09a
T5	4.82 ± 0.07b	1.18 ± 0.19bc	19.64 ± 3.36bc	100.34 ± 23.93b	196.00 ± 32.03a	0.70 ± 0.08a

CK, no fertiliser; CF, conventional fertiliser; T1, tea-specific formula fertiliser; T2, T1 + acidification amendment; T3, organic substitution based on T1; T4, urea formaldehyde slow-release fertiliser; T5, carbon-based organic fertiliser. TN, total N; SOM, soil organic matter; AP, available P; AK, available K; SFI, soil fertility index. AP and AK are represented in the form of the pure element. Values shown are the means ± SD (n = 4). Different letters in the same column represent significant differences among treatments at a significance level of p < 0.05.

). Moreover, soil TN in the CK was diminished to lower than 1.0 g kg⁻¹, and soil SOM of the CK was also the lowest in all treatments (p < 0.05). Although fertilisation led to high levels of soil TN and SOM, there were no significant differences in the TN and SOM contents between the CF and T1–T4 treatments. The soil TN and SOM in T5 were lower than that of CF (p < 0.05). The lowest soil AP and AK contents were observed in CK. Fertilisation treatments can elevate the content of soil AK and AP. The soil AP content in T1–T5 was higher than that in CF (p < 0.05). However, the AK content of T1–T5 was at the same level as in CF (Table 3).

As for the SFI, different fertilisation regimes promoted the SFI significantly compared to CK, although there was no significant difference between the different fertilisation regimes (Table 3). In addition, the soil heavy metal contents, including Cu, Cs, Cd, Cr and Pb, were not beyond the threshold in all treatments with the input of organic fertilisers (Table S4).

3.3. Fertiliser Use Efficiency and Economic Benefits under Different Fertilisation Regimes

AE was calculated in different fertilisation regimes based on the data of 2019 and 2020. Compared to CF, there were higher AE, NUE and PFP in T1–T5, among which T1 performed better than CF and the other chemical fertiliser reduction regimes. T1 showed the highest AE (0.10 kg kg⁻¹), followed by T5 (0.06 kg kg⁻¹), T2 (0.05 kg kg⁻¹) and T4 (0.04 kg kg⁻¹). The shoot NUE of T1–T5 was increased by 0.20–0.79% compared to that of CF, and NUE performed better in T1 (0.87%), T5 (0.62%), and subsequently, T2 (0.42%) and T4 (0.41%). The PFP was increased by 0.06–0.21 kg kg⁻¹ compared to CF and was relatively higher in the T1 (0.21 kg kg⁻¹) and T4 (0.19 kg kg⁻¹) treatments (

Table 4. Average nutrient utility efficiency in different fertilisation regimes from 2019 to 2020.

Treatment	AE (kg kg−1)	NUE (%)	PFP (kg kg−1)
CF	0.01 ± 0.01	0.08 ± 0.09	0.06 ± 0.01
T1	0.10 ± 0.01	0.87 ± 0.06	0.27 ± 0.01
T2	0.05 ± 0.03	0.42 ± 0.22	0.21 ± 0.01
T3	0.03 ± 0.01	0.28 ± 0.16	0.11 ± 0
T4	0.04 ± 0.02	0.41 ± 0.18	0.19 ± 0
T5	0.06 ± 0.03	0.62 ± 0.30	0.19 ± 0.02

CK, no fertiliser; CF, conventional fertiliser; T1, tea-specific formula fertiliser; T2, T1 + acidification amendment; T3, organic substitution based on T1; T4, urea formaldehyde slow-release fertiliser; T5, carbon-based organic fertiliser. AE, agronomic efficiency; NUE, shoot nutrient use efficiency; PFP, partial factor productivity. Values shown are the means ± SD (n = 2).

).

The input, output and benefit in the experimental tea garden were also assessed based on the values from 2019 to 2021. The total labour cost included fertilisation, trimming and plucking; the total output value was varied in the yield and price of fresh tea and fertiliser cost was varied in the fertiliser types and dosages.

The net income of CF was the lowest in all the treatments, even lower than that of CK (except in 2020, when CK was the lowest), and T1 was the highest in all treatments. The average net income of the three years (2019–2021) was also consistent with the above trend. Meanwhile, the increased proportion in the average net income compared to CF was the highest in T1 (135.28%), and the performance of T3 (67.53%), T2 (48.65%), T4 (38.07%) and T5 (33.35%) was lower than that of T1 (

Table 5. Average of the economic benefit of different fertilisation regimes from 2019 to 2021.

Treatment	Gross Income (CNY ha−1)	Labor Cost (CNY ha−1)	Fertiliser Cost (CNY ha−1)	Net Income (CNY ha−1)	Net Income Increment (%)
CK	6372.69 ± 2557.58	5027.43 ± 2020.23	0 ± 0	1345.23 ± 543.34ab
CF	7692.04 ± 3035.11	5886.47 ± 2377.26	614.3 ± 0	1191.3 ± 699.87b
T1	10,346.71 ± 3365.02	7370.73 ± 2637.86	173.1 ± 0	2802.87 ± 728.71a	197.39 ± 182.77
T2	8781.43 ± 3770.73	6616.43 ± 2923.46	394.1 ± 0	1770.87 ± 847.92ab	68.24 ± 67.99
T3	9578.14 ± 4923.42	7055.27 ± 3641.22	527.1 ± 0	1995.77 ± 1282.2ab	80.63 ± 71.23
T4	8102.89 ± 2921.53	6127.27 ± 2382.37	330.8 ± 0	1644.77 ± 548.36ab	72.02 ± 104.8
T5	9729.77 ± 3356.3	7121.77 ± 2858.07	1019.4 ± 0	1588.60 ± 498.52ab	60.22 ± 78.52

CK, no fertiliser; CF, conventional fertiliser; T1, tea-specific formula fertiliser; T2, T1 + acidification amendment; T3, organic substitution based on T1; T4, urea formaldehyde slow-release fertiliser; T5, carbon-based organic fertiliser. Values shown are the means ± SD (n = 3). Different letters in the same column represent significant differences among treatments at a significance level of p < 0.05. Net income increment was the lift ratio compared to that of the CF.

). The high net income of T1 was attributed to the highest yield of tea buds and the lowest fertiliser cost. The high fertiliser cost in T3 and T5 contributed to a lower net income.

CK, no fertiliser; CF, conventional fertiliser; T1, tea-specific formula fertiliser; T2, T1 + acidification amendment; T3, organic substitution based on T1; T4, urea formaldehyde slow-release fertiliser; T5, carbon-based organic fertiliser. The radar map was based on the normalised percentage of the fertiliser reduction ratio, tea yield, tea quality, SFI, AE and economic benefits. Among the above parameters, the SFI was the result of 2020; AE was the average result of 2019 and 2020 and tea yield, tea quality and economic benefit were the average results of 2019–2021.

The overall evaluation of each fertilisation regime revealed that the T1 treatment performed better than the other fertilisation regimes, especially with more yield, economic benefit, AE and chemical fertiliser reduction rates, whereas CF had the worst effect (Figure 2).

4. Discussion

Our research findings revealed that fertilisation can increase the tea yield, and excessive fertilisation resulted in a yield decrease compared to the fertiliser reduction regimes (T1–T5). These results are consistent with other researchers’ findings [14,41] that, at low N levels, the tea yield increases with the rising N dosage but then maintains or even decreases under conditions of N excess. Similarly, Ma et al. [23] also found that the tea yield demonstrated a quadratic response to increasing N rates (0–569 kg ha⁻¹). They found that annual N application rates of 119–285 kg ha⁻¹ were the most optimal for achieving a high tea yield and quality. Chen et al. [41] also suggested that an appropriate amount of N fertiliser (225 kg ha⁻¹) balanced the tea yield and quality. These results indicates that the marginal yield decreased under high N application conditions. In our study, excess NPK application in CF also led to a lower tea yield and bud density. Meanwhile, the tea yield and bud density had strong correlations with the NPK rates; when the NPK rates were 777.00 kg ha⁻¹ and 955.50 kg ha⁻¹, the tea yield and bud density reached the peak values, respectively. This explained higher yields in the fertilisation reduction regimes (T1–T5), which were attributed to a higher bud density rather than 100-bud weight.

In addition, the fertilisation reduction regimes also affected the tea quality. Overall, high N application was beneficial for improving the tea quality. High N application contributes to higher AA and chlorophyll contents while decreasing the TP and TP/AA ratio [15,41]. Ma et al. [23] found that long-term N application in a field experiment resulted in a quadratic increase in the AA content but a quadratic decrease in the TP content. Our results also revealed that fertilisation reduction regimes (T1–T5) could decrease the content of AA and increase the content of TP and the ratio of TP/AA in tea buds. Considering the balance between tea yield and quality in our study, fertilisation reduction regimes were acceptable in the tea production. Nevertheless, Xie et al. [42] reported that organic substitution in tea plantations can promote the tea yield by 4.4–17.4% and decrease the TP/AA ratio [36], which is inconsistent with our results, i.e., T3 (organic substitution treatment). Wei et al. [19] suggested that high P and K fertiliser input could inhibit AA accumulation in tea plants. Irrational P application also reduced the sensory quality of tea by inhibiting polyphenol accumulation and inducing the accumulation of certain anthocyanins [43]. Chen et al. [44] found that the application of low-phosphorus fertilisers on a tea plantation improved the taste and aroma of the tea. In our study, excessive P (330 kg ha⁻¹) and K (240 kg ha⁻¹) rates mainly input by organic fertiliser in T3 may contribute to the increase in the TP/AA ratio. These results indicated that an appropriate P and K ratio in organic fertiliser substitution is essential to tea quality formation. Furthermore, the sensory evaluation score of processed tea for each treatment did not show remarkable differences. This suggests that reducing the appropriate fertilisers can increase the tea yield without compromising the quality or with a negligible decline in tea quality.

Soil pH also plays an important role in tea plant quality parameters formation [20]. According to the literature reported, the rhizosphere soil pH of tea plants was affected by the N application rates, N form and root exudates [8,45,46]. Excessive application of N fertilisers exacerbated the soil acidification in tea plantations [4,45]. Therefore, chemical fertilisers reduction, organic fertilisers, biochar and acidification amendments are recommended to ameliorate soil acidification [47,48]. In our study, chemical fertiliser reduction treatments can alleviate soil acidification compared to excessive fertilisation. Although soil acidification in T2 was completely resolved, this did not contribute to a higher yield and better tea quality. Yan et al. reported that tea plant growth will be restricted by over-liming as a result of the high soil pH and Ca concentration inhibiting the K and Al uptake [49]. This accounted for the reason that the yield of T2 was lower than that of the T1 and T3 treatments, since the T2 treatment contained soil acidification amendments with a high Ca content (CaO ≥ 30.0%). These results also suggest that pH acidification amendment is only necessary and helpful to tea production when the soil pH < 4.5.

With regards to the fertiliser use efficiency, the chemical fertiliser reduction regimes (T1–T5) increased the AE, NUE and PFP in 2019 and 2020 compared to CF. This suggested that excessive fertilisation could decrease the fertiliser utilisation efficiency. The decreased AE, NUE and PFP in CF was mainly explained by the excessive NPK rates, which were much higher than the threshold recommended by Ni et al. [2]. In other related experiments, chemical fertiliser reduction also improved the fertiliser utilisation efficiency in a tea garden [50,51]. These results were consistent with our study. Tang et al. [51] found that the average AEs for N, P and K were 2.63, 5.56 and 6.32 kg kg⁻¹. However, in our study, the AE, NUE and PFP were far lower than the above-mentioned values, and this may be caused by different tea yields, which were determined by different picking standards. Between the fertiliser reduction regimes, T1 performed better than the other treatments in agronomic use efficiency. This can be attributed to the lowest fertiliser rate and highest tea yield in T1. AE, NUE and PFP of T3 were the lowest in all the fertiliser reduction treatments, which can be also explained by the high fertilisation rates and imbalance of the N-P₂O₅-K₂O ratio, which were was only second to that of CF.

Slow-release fertiliser could significantly decrease the soil nitrification, increase the N use efficiency and improve tea growth, thus reducing environmental pollution [52]. A former study found that slow-release N fertiliser blending with conventional N fertiliser can lead to a higher total net income increment when compared to the common chemical N fertiliser [53]. In our study, the net income of T4 (urea formaldehyde slow-release fertiliser) was only higher than that of T5 in all fertiliser reduction regimes. This difference may be caused by different types of fertilisers and the high price of the slow-release fertilisers used in T4.

According to the average net benefit of 2019–2021, the CF resulted in the lowest net benefit in all treatments, even unexpectedly lower than that of CK; this result suggested that overuse of fertilisation had negative effects on the net benefit. Due to the higher fertilisation rates, the fertiliser cost of CF was higher than that of T1–T4. However, the total cost of labour was lower than that of T1–T5; this was caused by the lower yield and corresponding lower picking cost in CF. Among the T1–T5 treatments, T1 produced the highest net benefit and ratio of the net income increment compared to that of the other treatments. This was mainly caused by the highest yield and lowest fertiliser cost in T1.

5. Conclusions

Over a 3-year field experiment, the chemical fertiliser reduction treatments significantly altered tea production, including soil quality, tea yield and quality and economic benefits. The chemical fertiliser nutrient input was reduced by 57.14–73.51%, and the actual total nutrient was diminished by 41.62–69.68% in T1–T5. Meanwhile, compared to the CF, the spring tea yield increased by 4.65–28.67%; among which, T1 (28.67%), T5 (23.86%), T3 (22.58%) and T2 (14.10%) had good effects, while AA, TP and the other quality parameters were not significantly decreased. The nutrient use efficiency also increased to varying degrees, with the AE increasing by 0.02–0.10 kg kg⁻¹, NUE 0.15–0.89% and PFP 0.05–0.18 kg kg⁻¹ in T1–T5 compared to the CF. Moreover, the T1 treatment showed the best economic benefits, which increased by 135.28% compared to the CF treatment, and the T3 (67.53%), T2 (48.65%), T4 (38.07%) and T5 (33.35%) treatments also achieved good results. In summary, the T1 treatment, i.e., fertilisation reduction and NPK ratio adjustment treatment (N = 300 kg ha⁻¹, P₂O₅ = 60 kg ha⁻¹ and K₂O = 90 kg ha⁻¹), has good application prospects in tea plantations.