Quantitative Estimation of the Nutrient Uptake Requirements of Peanut

: Understanding the characteristics of the balanced nutrient requirements for peanut to achieve target yields is paramount when formulating fertilizer management strategies to increase yields and avoid fertilizer loss. Nutritional requirement estimation models can provide e ﬀ ective alternatives for the estimation of the optimum crop balanced nutrient requirements under varied agricultural conditions which are less time consuming and expensive. In the present study, the quantitative estimation of the optimum crop balanced nutrient requirements of peanut in China were obtained using quantitative evaluation of fertility of tropical soils (QUEFTS) model. The database covered the main agro-ecological region for peanut crops in China between 1993 and 2018. The predicted results of the QUEFTS model indicated that nutrient uptake requirements increased linearly with increasing pod yields until the yields had reached approximately 60% to 70% of the potential pod yields. It was found that with the increasing pod yields during the nutrient linear absorption stage, the plants had required 38.4 kg N, 4.3 kg P, and 14.0 kg K in total to produce 1000 kg of pods, and the corresponding internal e ﬃ ciencies were 26.0 kg N / kg, 235.0 kg P / kg, and 71.6 kg K / kg, respectively. In addition, the balance rates of the removal nutrient in the pods were determined to be 29.4 kg N, 2.9 kg P, and 4.9 kg K per 1000 kg of pod yield, or approximately 76.5%, 67.4%, and 34.7% of N, P, and K in the total plants, respectively. This study’s ﬁeld validation experiments veriﬁed the applicability and accuracy of the QUEFTS model. Therefore, it was considered to be an e ﬀ ective alternative for the estimation of the optimal balance N, P, and K uptake requirements for peanut crops. These ﬁndings will potentially be helpful when making future decisions regarding fertilizer recommendations for peanut crops in China.


Introduction
Peanut (Arachis hypogaea L.) is one of the major global food legume and oilseed crops, and offers multiple benefits to meet human nutritional needs throughout the world [1,2]. As stated by FAO (2017), peanut crops are grown in 118 countries (or regions) around the world using approximately 28 million ha of land [3]. China is currently one of the main peanut producing countries. China has allotted 4.63 million ha of land to peanut production, producing 17.15 million tons of peanut pods, which approximately correspond to 17% and 36% of the global peanut production rates [3,4]. However,

Model Development
The QUEFTS model was designed as a promising new tool for evaluating the fertility levels of native and tropical soil. It was first developed for maize crops under rain-fed conditions by Janssen et al. [12]. In addition, further improvements were made by Smaling and Janssen [13] and Sattari et al. [14], in which the relationships between grain yields and nutrient accumulations in the total aboveground plant dry matter of crops at maturity were described using a large database. Peanut belong the category of underground crops. During the harvesting processes, the total nutrient of plants (including roots, pods, and haulms) will be removed [25]. In the present study, the nutrient data of the roots were accumulated with the haulm data.
The following were the key steps in this research experiment: (1) The available appropriate data were selected in order to realize the boundary conditions of QUEFTs model, in which the HI less than 0.4 were excluded due to the effects of abiotic or biotic stress conditions, rather than nutrient deficiency conditions; (2) The two boundary lines of N, P, and K for the crop yields at maximum nutrient accumulation (YA) and maximum nutrient dilution (YD) were defined by the maximum nutrient accumulation value (a) and maximum nutrient dilution value (d), which were calculated by the lower and upper 2.5th (Set I), 5th (Set II), and 7.5th (Set III) percentiles of all the nutrient internal efficiencies (IEs) data; (3) The appropriate a and d values for the peanut database were then acquired; (4) Then, the optimal N, P, and K uptake curves for peanut crops in China were simulated; (5) Finally, the applicability of the QUEFTS model for simulating the nutrient uptake requirements in China were validating using field experiments.
The related parameters IE, that was represented the amount of pod yield produced per kg nutrient uptake in total plant matter [15,20]. The reciprocal internal efficiency (RIE) was represented nutrient uptake requirements per 1000 kg of pod yield [15,20]. IE (Equation (1)) and RIE (Equation (2)) using the following formula:

Model Development
The QUEFTS model was designed as a promising new tool for evaluating the fertility levels of native and tropical soil. It was first developed for maize crops under rain-fed conditions by Janssen et al. [12]. In addition, further improvements were made by Smaling and Janssen [13] and Sattari et al. [14], in which the relationships between grain yields and nutrient accumulations in the total above-ground plant dry matter of crops at maturity were described using a large database. Peanut belong the category of underground crops. During the harvesting processes, the total nutrient of plants (including roots, pods, and haulms) will be removed [25]. In the present study, the nutrient data of the roots were accumulated with the haulm data.
The following were the key steps in this research experiment: (1) The available appropriate data were selected in order to realize the boundary conditions of QUEFTs model, in which the HI less than 0.4 were excluded due to the effects of abiotic or biotic stress conditions, rather than nutrient deficiency conditions; (2) The two boundary lines of N, P, and K for the crop yields at maximum nutrient accumulation (YA) and maximum nutrient dilution (YD) were defined by the maximum nutrient accumulation value (a) and maximum nutrient dilution value (d), which were calculated by the lower and upper 2.5th (Set I), 5th (Set II), and 7.5th (Set III) percentiles of all the nutrient internal efficiencies (IEs) data; (3) The appropriate a and d values for the peanut database were then acquired; (4) Then, the optimal N, P, and K uptake curves for peanut crops in China were simulated; (5) Finally, the applicability of the QUEFTS model for simulating the nutrient uptake requirements in China were validating using field experiments.
The related parameters IE, that was represented the amount of pod yield produced per kg nutrient uptake in total plant matter [15,20]. The reciprocal internal efficiency (RIE) was represented nutrient uptake requirements per 1000 kg of pod yield [15,20]. IE (Equation (1)) and RIE (Equation (2)) using the following formula: Agronomy 2020, 10, 119 where Y pod is pod yield and U plant indicates the nutrient uptake of the total plants.

Model Validation
This study's field validations were conducted in 2018 across five agro-ecological zones in China (including the provinces of Jilin, Liaoning, Shandong, and Hebei). A completely randomized block design with three replications was applied to all of the experiments, and each plot size measured 30 m 2 . The peanut crops were sown in May and harvested in September. The fertilizer recommendations were provided by the Nutrient Expert System (NE). The NE was a computer software developed by IPNI based on a SSNM strategy, which included the QUEFTS model's predicted nutrient uptake data for determining the optimal nutrient requirements of the peanut crops [26]. Adequate measurements were taken in the field experiments in order to control the effects of pests, disease, and weeds.
At maturity, the peanut plants were manually harvested and total weights of the pods and haulm were measured separately. All of the plant parts were oven-dried to a constant weight at 70 • C and then analyzed in order to determine the N, P, and K concentrations. The individual plant parts were digested with H 2 SO 4 -H 2 O 2 , and the concentrations of N, P, and K were determined adopting the Kjeldahl approach, vanado-molybdate yellow color approach, and flame spectrophotometers, respectively. And, the accumulation of N, P, and K in total plant was calculated by multiplying the matter dry weight by the nutrient concentrations.
The performance of the QUEFTS model was evaluated using coefficient of determination (R 2 ), root-mean-square error (RMSE), and normalized-RMSE (n-RMSE) between the measured and simulated data. The R 2 estimates the combined dispersion against the single dispersion of the measured and simulated series [27]. The RMSE measures the mean discrepancy between the simulated and measured data with the same unit and the n-RMSE removed the unit and allowed for comparisons between values with different units [21].
where M i and M represent the values of the measured and mean of the measured nutrient uptakes (kg/ha), respectively; S i represents the values of the simulated nutrient uptake (kg/ha); and n indicates the number of values.

Pod Yields and Nutrient Uptake Data
The average value of the pod yields (10.0% moisture content) was determined to be 3918 kg/ha in the present peanut database and had ranged from 500 to 8602 kg/ha (Table 1). In addition, it was determined that 95.1% of the pod yield had a centralized distribution in the range from the 2.5th to 97.5th percentile for the present peanut database ( Figure 2). The significant variabilities in the peanut pod yields were possibly due to the fact that differences had existed between the different field experiments, including climate conditions, agronomic practices, fertilized treatments, and peanut varieties. Average peanut pod yield in the current database was found to be higher than the 3709 kg/ha in China's mainland and the 1685 kg/ha for the world [3]. The harvest index (HI) for the entire database ranged from 0.31 to 0.82, with an average value of 0.51. The distributions of the HI and pod yields are shown in Figure 2. It can be seen in the figure that 95.5% of data were centralized distributed in the range of the 2.5th to 97.5th percentiles for the HI of the present peanut database. , and 75% Q represent the 25th, 50th, and 75th percentiles of the database, respectively; 3 N-HI, nutrient harvest index of N; P-HI, nutrient harvest index of P; K-HI, nutrient harvest index of K.
Agronomy 2020, 10, x FOR PEER REVIEW 5 of 12 yields are shown in Figure 2. It can be seen in the figure that 95.5% of data were centralized distributed in the range of the 2.5th to 97.5th percentiles for the HI of the present peanut database.   The concentrations of N and P in the peanut pods (35.6 g N/kg and 3.6 g P/kg, ranging from 13.9 to 64.5 g N/kg and 1.6 to 8.2 g P/kg, respectively) were found to be higher than the haulm (14.2 g N/kg and 1.7 g P/kg, ranging from 6.9 to 38.4 g N/kg and 0.31 to 4.0 g P/kg, respectively). Meanwhile, the K content in the peanut pods (7.1 g K/kg, ranging from 2.2 to 17.3 g K/kg) was observed to be lower than the haulm (12.9 g K/kg, ranging from 2.1 to 35.8 g K/kg), as detailed in Table 1. The average total plant accumulations were 163.6 kg N/ha, 17.8 kg P/ha, and 59.2 kg K/ha, with accumulation ranges of 30.1 to 456.7 kg N/ha, 2.8 to 51.2 kg P/ha and 7.6 to 228.1 kg K/ha, respectively. The variability of the nutrient distributions reflected the diversity in the environmental conditions and tillage treatments. However, the current database was lower than the N accumulation of 180 to 228 kg/ha under the various tillage treatments [28]. This may be due to that plant macronutrient amounts in underground peanut parts have significant effects by soil bulk density, and tends to be affected by tillage practices [25].
Nutrient harvest indexes (N-HI, P-HI, and K-HI) is defined as kg nutrient in pod per kg nutrient in total plant dry matter at maturity. Average N-HI, P-HI, and K-HI were 0.72, 0.69, and 0.39 kg/kg, respectively. Peanut pod absorbed corresponding proportion of the nutrient, that approximately 72%, 69%, and 39% of the N, P and K distribution in the pods, respectively.

Internal Efficiency and Reciprocal Internal Efficiency
In the current database, the IE values of N were found to have relatively small variations, and ranged from 14.3 to 49.8 kg pod/kg plant N, with an average value of 24.7 kg/kg. However, the IE values of P and K were observed to widely vary, ranging from 89.8 to 483.1 kg pod/kg plant P (average value of 238.5 kg/kg), and 17.9 to 231.5 kg/kg plant K (average value of 85.1 kg/kg), as detailed in Table 2. It was observed that the IE values were greatly affected by the nutrient management practices. The large variations in the observed IE values in the current peanut database reflected the variations in the field experiments, including season and site specific differences in environmental conditions (for example, temperature), nutrient imbalances, irrigation, weed and pest control measures, continuous cropping practices [29], tillage treatments [25], and so on. The average RIE (reciprocal internal efficiency; nutrient uptake requirements per 1000 kg of pod yield) values of the N, P, and K were 42.2, 4.5, and 15.3 kg, with ranges from 20.1 to 69.9 kg for N; 2.1 to 11.1 kg for P; and 4.3 to 55.9 kg for K, respectively observed, as detailed in Table 2.

Selection of the Data for Adapting the QUEFTS Model
In the present study, the data with HI lower 0.4 was removed due to the fact that the crop yields had been subjected to different constraints, such as disease, pests, and drought, with the exception of N, P, or K supply limitations [23]. The parameters of a and d were calculated using data with HI ranging from 0.4 to 0.82 in order to ensure that the peanut growth had been mainly limited by nutrient constraints rather than biotic or abiotic stress conditions. The constant a and d values of the nutrient were calculated by excluding the upper and lower 2.5th (Set I), 5th (Set II) or 7.5th (Set III) percentiles of the IEs (Table 3), which were indicated by the relationships between the pod yields and nutrient accumulations in the dry plant matter at maturity. The relationship between the pod yields and the nutrient requirements of the total dry plant matter was simulated by the QUEFTS model using three groups of a and d values under a potential yield of 10 t/ha, respectively. As shown in Figure 3, the N, P, and K uptake requirements (YU) which were simulated by the QUEFTS model were similar under three a and d sets for the different pod yield targets, with the exception being when the yield target approached a potential yield 10 t/ha. This study suggested that Set I should be adopted as the standard parameter set in the QUEFTS model in order to accurately estimate the balanced nutrient requirements of the peanut crops and recognize the relationship between the pod yields and the nutrient accumulations. This was due to the fact that there were large variabilities observed in the IE values among the field experiments.
In the present study, the data with HI lower 0.4 was removed due to the fact that the crop yields had been subjected to different constraints, such as disease, pests, and drought, with the exception of N, P, or K supply limitations [23]. The parameters of a and d were calculated using data with HI ranging from 0.4 to 0.82 in order to ensure that the peanut growth had been mainly limited by nutrient constraints rather than biotic or abiotic stress conditions. The constant a and d values of the nutrient were calculated by excluding the upper and lower 2.5th (Set I), 5th (Set II) or 7.5th (Set III) percentiles of the IEs (Table 3), which were indicated by the relationships between the pod yields and nutrient accumulations in the dry plant matter at maturity. The relationship between the pod yields and the nutrient requirements of the total dry plant matter was simulated by the QUEFTS model using three groups of a and d values under a potential yield of 10 t/ha, respectively. As shown in Figure 3, the N, P, and K uptake requirements (YU) which were simulated by the QUEFTS model were similar under three a and d sets for the different pod yield targets, with the exception being when the yield target approached a potential yield 10 t/ha. This study suggested that Set I should be adopted as the standard parameter set in the QUEFTS model in order to accurately estimate the balanced nutrient requirements of the peanut crops and recognize the relationship between the pod yields and the nutrient accumulations. This was due to the fact that there were large variabilities observed in the IE values among the field experiments.  , d) constants. In the figure, the calculations excluded the upper and lower 2.5th (Set I), 5th (Set II), and 7.5th (Set III) percentiles of all the IE data (HI range from 0.4 to 0.82); YD, YA, and YU represent the maximum dilution, maximum accumulation, and balanced uptake of N, P, and K in plant dry matter, respectively; Potential yield was set at 10 t/ha.

Estimation of Nutrient Uptake Requirements
In the present study, across the potential pod yields (ranging from 6 to 10 t/ha), the QUEFTS model simulated the balanced nutritional uptake requirements for N, P, and K with the Set I a and d values (17.0 and 35.7 kg pod/kg N; 135.6 and 364.9 kg pod/kg P; 31.3 and 154.4 kg pod/kg K) which  (Set III) percentiles of all the IE data (HI range from 0.4 to 0.82); YD, YA, and YU represent the maximum dilution, maximum accumulation, and balanced uptake of N, P, and K in plant dry matter, respectively; Potential yield was set at 10 t/ha.

Estimation of Nutrient Uptake Requirements
In the present study, across the potential pod yields (ranging from 6 to 10 t/ha), the QUEFTS model simulated the balanced nutritional uptake requirements for N, P, and K with the Set I a and d values (17.0 and 35.7 kg pod/kg N; 135.6 and 364.9 kg pod/kg P; 31.3 and 154.4 kg pod/kg K) which were derived from the current database. As shown in Figure 4, the QUEFTS model predicated linear increases in the pod yields if the nutrient were taken up to the balanced amounts until the pod yields reached approximately 60% to 70% of the potential pod yields, and the IE values of the N, P, and K were constant (Table 4). During the constant slope increasing phase, the balanced nutrient uptake requirements to produce 1000 kg of pods were 38.4 kg N, 4.3 kg P, and 14.0 kg K. In addition, the corresponding optimal IE values were 26.0 kg pod/kg N, 235.0 kg pod/kg P, and 71.6 kg pod/kg K for balanced nutrition, respectively. Therefore, the optimal balanced required N: P: K ratio in the total plants were 9.0:1.0:3.3 (Table 4), which was similar to the average of observed nutrient uptake ratio of 9.2:1.0:3.3. The average observed values of N, P, and K were higher than the values simulated Agronomy 2020, 10, 119 8 of 11 by the QUEFTS model. This was determined to be due to the fact that the N, P, and K predicted by the QUEFTS model were the optimal nutrient requirements under the conditions of the balanced absorption of the three nutrient elements (N, P, and K) [15]. As shown in Figure 4a, a large number of data points were concentrated in the locality of the lower boundary, which reflected an excess or luxuriant uptake of N nutrient in the current database. However, Fang and Zhao [30] concluded that the following concentrations were needed in order to produce 1000 kg of pods for the different peanut varieties: 37.4 to 49.1 kg N; 6.8 to 8.7 kg P; and 15.8 to 21.7 kg K. The higher values of the balanced nutrient uptake requirements predicted by the QUEFTS model may have been the results of over fertilization practices in the farmland areas in China.
The removal of the nutrient from the pods or harvested plant parts must be replenished in the soil in order to maintain soil fertility. Guidance for appropriate fertilizer recommendations has been provided in many practical algorithms [31]. The removal of nutrient from the peanut pods was simulated by the QUEFTS model in this study, as shown in Figure 4. It can be seen in the figure that linear increases in pod yields had occurred in the cases where the grain nutrient were balanced until pod yields reached approximately 60% to 70% of the potential pod yield. The balanced N, P, and K removal amounts to produce 1000 kg of pods were determined to be 29.4 kg N, 2.9 kg P, and 4.9 kg K, respectively (N:P:K = 10.2:1.0:1.7). It was found that when compared to balanced nutrient requirements of the total plants, approximately 79.5%, 67.4%, and 34.7% of N, P, and K had accumulated in the pods and were removed. Therefore, these values should be considered in order to sustain soil fertility when formulating fertilizer recommendations for peanut crops.
Agronomy 2020, 10, x FOR PEER REVIEW 8 of 12 were derived from the current database. As shown in Figure 4, the QUEFTS model predicated linear increases in the pod yields if the nutrient were taken up to the balanced amounts until the pod yields reached approximately 60% to 70% of the potential pod yields, and the IE values of the N, P, and K were constant (Table 4). During the constant slope increasing phase, the balanced nutrient uptake requirements to produce 1000 kg of pods were 38.4 kg N, 4.3 kg P, and 14.0 kg K. In addition, the corresponding optimal IE values were 26.0 kg pod/kg N, 235.0 kg pod/kg P, and 71.6 kg pod/kg K for balanced nutrition, respectively. Therefore, the optimal balanced required N: P: K ratio in the total plants were 9.0:1.0:3.3 (Table 4), which was similar to the average of observed nutrient uptake ratio of 9.2:1.0:3.3. The average observed values of N, P, and K were higher than the values simulated by the QUEFTS model. This was determined to be due to the fact that the N, P, and K predicted by the QUEFTS model were the optimal nutrient requirements under the conditions of the balanced absorption of the three nutrient elements (N, P, and K) [15]. As shown in Figure 4a, a large number of data points were concentrated in the locality of the lower boundary, which reflected an excess or luxuriant uptake of N nutrient in the current database. However, Fang and Zhao [30] concluded that the following concentrations were needed in order to produce 1000 kg of pods for the different peanut varieties: 37.4 to 49.1 kg N; 6.8 to 8.7 kg P; and 15.8 to 21.7 kg K. The higher values of the balanced nutrient uptake requirements predicted by the QUEFTS model may have been the results of over fertilization practices in the farmland areas in China.

Validation of the QUEFTS Model
The validations of the QUEFTS model's results were evaluated by R 2 , RMSE, and n-RMSE with respect to the estimation of the nutrient uptake requirements for peanut crops. The data obtained during this study's 2018 field experiments were analyzed. There was found to be a satisfactory agreement between the simulated values of the QUEFTS model and the observed data from the field experiments of five peanut agro-ecological regions in China. As shown in Figure 5, the scatter plots between the observed and simulated values distributed near the 1:1 line, with R 2 of 0.89, 0.64, and 0.32; RMSE of N, P, and K of 34.6, 3.9, and 18.3 kg/ha; and n-RMSE values of N, P, and K of 17.3%, 21.3%, and 29.5%, respectively. However, the restrictions other than the nutritional limitations were determined to be unavoidable in the on-farm field experiments, such as disease, pests, drought, and tillage treatments, which may have accounted for the slight underrating and overrating results obtained by the QUEFTS model. Generally speaking, it was found in this study that the QUEFTS model could be effectively used to predict the nutrient requirements for peanut crops with generic applicability and robust accuracy. The removal of the nutrient from the pods or harvested plant parts must be replenished in the soil in order to maintain soil fertility. Guidance for appropriate fertilizer recommendations has been provided in many practical algorithms [31]. The removal of nutrient from the peanut pods was simulated by the QUEFTS model in this study, as shown in Figure 4. It can be seen in the figure that linear increases in pod yields had occurred in the cases where the grain nutrient were balanced until pod yields reached approximately 60% to 70% of the potential pod yield. The balanced N, P, and K removal amounts to produce 1000 kg of pods were determined to be 29.4 kg N, 2.9 kg P, and 4.9 kg K, respectively (N:P:K = 10.2:1.0:1.7). It was found that when compared to balanced nutrient requirements of the total plants, approximately 79.5%, 67.4%, and 34.7% of N, P, and K had accumulated in the pods and were removed. Therefore, these values should be considered in order to sustain soil fertility when formulating fertilizer recommendations for peanut crops.

Validation of the QUEFTS Model
The validations of the QUEFTS model's results were evaluated by R 2 , RMSE, and n-RMSE with respect to the estimation of the nutrient uptake requirements for peanut crops. The data obtained during this study's 2018 field experiments were analyzed. There was found to be a satisfactory agreement between the simulated values of the QUEFTS model and the observed data from the field experiments of five peanut agro-ecological regions in China. As shown in Figure 5, the scatter plots between the observed and simulated values distributed near the 1:1 line, with R 2 of 0.89, 0.64, and 0.32; RMSE of N, P, and K of 34.6, 3.9, and 18.3 kg/ha; and n-RMSE values of N, P, and K of 17.3%, 21.3%, and 29.5%, respectively. However, the restrictions other than the nutritional limitations were determined to be unavoidable in the on-farm field experiments, such as disease, pests, drought, and tillage treatments, which may have accounted for the slight underrating and overrating results obtained by the QUEFTS model. Generally speaking, it was found in this study that the QUEFTS model could be effectively used to predict the nutrient requirements for peanut crops with generic applicability and robust accuracy. Figure 5. Comparisons of the simulated and observed data of the N, P, and K uptake in the total peanut dry plant matter in 2018. The observed data were obtained from the experimental field plots where a Nutrient Expert (NE) Decision Support System was applied; and the simulated nutrient uptake data were estimated using the QUEFTS model. Figure 5. Comparisons of the simulated and observed data of the N, P, and K uptake in the total peanut dry plant matter in 2018. The observed data were obtained from the experimental field plots where a Nutrient Expert (NE) Decision Support System was applied; and the simulated nutrient uptake data were estimated using the QUEFTS model.

Conclusions
In this study, a database for peanut yields and nutrient uptake requirements was preliminarily established, and the results obtained demonstrated that the QUEFTS model could be used as an effective