CN108564200A - A kind of soil fertility prediction technique building geographical MDS minimum data set based on yield - Google Patents

Abstract

The invention belongs to the technical field of soil detection, and in particular relates to a method for predicting soil fertility based on yield to construct a geographic minimum data set, comprising the following steps: (1) collecting soil samples; (2) detecting the contents of various fertility indicators of the samples; (3) ) carry out the analysis and calculation of Nemerow's single item fertility index and comprehensive fertility index to the index content measured in step (2); (4) utilize SPSS data processing software to carry out principal component analysis to the index content measured in step (2), select the principal component Analyze medium and high factor loading indicators, and enter the minimum data set; (5) perform standardized calculation on the data set in step (4); (6) calculate the minimum data set fertility index on the standardized minimum data set in step (5). According to the correlation analysis results of the minimum data set fertility index and yield, the fertility index can predict soil fertility; the soil fertility prediction method of the present invention can uniformly evaluate the soil fertility of camellia oleifera with large spatial differences, and has a wide range of applications.

Description

A Soil Fertility Prediction Method Based on Yield Constructing Geographically Minimal Dataset

technical field

The invention belongs to the technical field of soil detection, and in particular relates to a method for predicting soil fertility by constructing a geographic minimum data set based on yield.

Background technique

Camellia oleifera (Camellia oleifera Abel.), Camellia oleifera is an evergreen small tree, and its seed oil is rich in unsaturated fatty acids such as oleic acid and linoleic acid. It is an important woody edible oil tree species in the south. Zhejiang Province currently has 200,000 hectares of camellia oleifera forests, ranking fourth in the country, of which 70% are distributed in Lishui and Quzhou areas, which is an important resource advantage and potential for farmers in mountainous areas to increase their income and become rich. At present, more than 60% of Camellia oleifera forests in Zhejiang Province are old Camellia oleifera forests with extensive management, low yield and low efficiency. Such forests have bad phenomena such as slow tree growth, decreased fruit size, and obvious annual growth, which seriously restrict the production and quality of Camellia oleifera. improve. The "National Camellia Camellia Industry Development Plan (2009-2020)" clarifies that the transformation of low-yield forests is an effective way to increase the yield of Camellia oleifera in the short term, and proposes a plan for the transformation of low-yield forests in Zhejiang Province. In recent years, the reconstruction of low-yield Camellia oleifera forests in Zhejiang Province has been gradually promoted at an annual rate of nearly 10,000 hectares. As an important measure for low-yielding forest transformation, precise water and fertilizer management has attracted more and more attention. However, there are some problems in fertilization of camellia oleifera at the present stage, such as ignoring the differences in the fertility requirements of camellia oleifera site conditions, ignoring the role of trace elements, and immature evaluation of camellia oleifera soil fertility. It is urgent to establish a soil fertility evaluation system suitable for camellia oleifera.

The establishment of the minimum soil fertility data set is the basis and an important part of soil fertility evaluation and determination of key fertilization factors. Common methods for constructing the minimum data set include principal component analysis and correlation analysis, and many scholars use the norm value to extract indicators. The present invention selects principal component analysis as the main research method to establish the minimum data set of camellia oleifera soil fertility in Zhejiang Province, evaluate the soil fertility of the main camellia oleifera production area, and use the Nemerow index method in conjunction with the output of the sample plot to verify the minimum data set, The research results have important guiding significance for improving the development of camellia oleifera planting industry in Zhejiang Province, improving blind fertilization, and protecting the environment.

Soil fertility is the ability of soil to supply the nutrients necessary for plants, as well as various soil properties and states related to nutrient supply ability. Soil fertility evaluation is to select the index that can best reflect the soil quality and production capacity and judge the soil quality. Therefore, the construction of the minimum data set is the key link of soil fertility evaluation. At present, there are no related reports on the construction of the minimum data set for Camellia oleifera soil fertility.

Contents of the invention

The purpose of the present invention is to overcome the deficiencies of the prior art, and to provide a soil fertility prediction method based on the yield to construct the geographical minimum data set. In the case of large differences in various nutrients in the soil, the evaluation method can be applied to various The fertility index evaluation of the nature and state of the soil, and can accurately judge the situation of Camellia oleifera soil productivity.

The technical solution adopted by the present invention to solve its technical problems is:

A soil fertility prediction method for constructing a geographic minimum data set based on yield, comprising the following steps:

(1) Sampling: Set the sampling location according to the output, and collect the Camellia oleifera soil samples;

(2) Detection: the content of each fertility index of the Camellia oleifera soil sample collected in detection step (1);

(3) Data processing and analysis of the full data set:

Evaluation of Nemerow's individual fertility index: Pi=Ci/Si

Pi: the individual fertility index of a certain index i in the soil;

Ci: measured data of a certain index i in the soil;

Si: The standard value of a certain index i in the soil.

Evaluation of Nemerow's Comprehensive Fertility Index

Ptotal: comprehensive index of soil fertility; (Ci/Si) ² min: the square of the minimum value of a single fertility index (Ci/Si) ² ave: the average square of all fertility indexes in the soil; n: the number of soil fertility indexes (4) Minimal data set data processing analysis:

S1: Use SPSS data processing software to carry out principal component analysis on the index content measured in step (2), and extract some principal components according to the principle that the characteristic value is greater than 1 and the cumulative contribution rate is greater than 70%;

S2: According to a certain standard, select the high factor loadings of several principal components obtained in step S1;

S3: According to a certain standard, select the high factor loadings of several principal components obtained in step S2 to enter the minimum data set:

(5) The minimum data set fertility index standardization:

Since the soil organic matter, total nitrogen, alkaline nitrogen, etc. and crop growth effect curves are generally S-shaped, the S-shaped membership function is used to standardize the fertility index.

Among them, x1 is the low critical value of the fertility index; x2 is the high critical value of the fertility index; x is the measured value of the fertility index.

(6) The establishment of the minimum data set fertility index

The principal component analysis method was used to obtain the variance of the common factor, based on which the weight coefficient was determined, and the soil fertility index NFI was calculated according to the following formula.

Wi: weight coefficient of fertility index; Fi: standard value of fertility index

(7) The correlation between the fertility index and yield of the smallest data set

The fertility index of the smallest data set and the comprehensive index and yield of the full data set were analyzed for correlation data.

As a preference, in the step (1), the sampling method is as follows: the sample plot is divided into upper, middle and lower slopes, and the S-shaped sampling method is used to collect 0-40 cm layers of soil, wherein 15-20 samples are randomly selected for each slope. point, the weight of each sample point shall not be less than 100g.

As a preference, after the samples of each slope are fully mixed, 1 kg is retained by quartering, put into a plastic bag, numbered, and the sample information is registered, and taken back to the laboratory for air drying.

Preferably, in the step (2), the fertility index is pH, total nitrogen, total phosphorus, organic matter, available nitrogen, available phosphorus, available potassium, soil texture, available trace elements iron, copper, zinc, manganese, exchangeable calcium , Exchangeable magnesium.

Preferably, in the step (4), the principle for selecting high factor loads in the principal components is: the absolute value of the high factor loads is greater than 90% of the maximum factor loads in the principal components.

As preferably, in described step (4), the standard that the high factor load of selecting principal component enters minimum data set is:

(a) When there is only one high factor loading indicator of a principal component, the indicator enters the minimum data set;

(b) When there are two high factor loading indicators of a principal component, do a correlation analysis on the two indicators. If the absolute value of the correlation coefficient r<0.4, both indicators will enter the minimum data set; if the absolute value of the correlation coefficient If r≥0.4, the indicators with high factor loads will enter the data set;

(c) When there are three or more high factor loading indicators of a principal component, the correlation analysis is performed in pairs. (1) When the absolute value of the pairwise correlation coefficient of all indicators is r<0.4, all indicators enter the minimum Data set; (2) When the absolute value of pairwise correlation coefficient r ≥ 0.4 exists for all indicators: if the number of indicators with absolute value of pairwise correlation coefficient r ≥ 0.4 accounts for more than 70% of the total number of indicators, the correlation coefficient with other indicators The high factor loading indicator with the largest sum of absolute values enters the dataset; otherwise all indicators enter the dataset.

The beneficial effects that the present invention obtains are:

(a) The present invention selects the principal component analysis research method, determines the minimum data set of Camellia oleifera soil fertility, evaluates the soil fertility of Camellia oleifera, and utilizes the Nemerow index method in conjunction with the output of the sample plot to verify the minimum data set; the present invention is based on the geographic minimum The soil fertility prediction method constructed by the data set can well reflect the relationship between Camellia oleifera yield and soil fertility. The research results have important guiding significance for improving the development of Camellia oleifera planting, improving blind fertilization, and protecting the environment;

(b) The method for predicting soil fertility of the present invention based on the production-based minimum geographic data set can comprehensively evaluate soil fertility indicators, including the detection and evaluation of trace elements, and can reflect the yield of camellia oleifera and various soil fertility indicators to a greater extent and the soil fertility of camellia oleifera with large spatial differences can be evaluated uniformly, with a wide range of applications.

Description of drawings

Fig. 1 Nemerow comprehensive index, minimum data set fertility index and yield map of camellia oleifera soil in Zhejiang Province.

In the figure: a: minimum data set fertility index; b: Nemerow composite index; c: yield.

Detailed ways

The present invention and its specific implementation will be described in further detail below in conjunction with the examples, but it is not meant to limit the scope of the present invention.

Example 1:

Take Camellia oleifera soil samples in Zhejiang Province, and according to the average yield of the sample plot and the county where the sample plot is located, three oil yields in the sample plot are greater than 150kg·ha ^-1 , 75-150kg·ha ^-1 , and less than 75kg·ha ^-1 Level, to divide the soil measurement results.

When sampling, the sample plot is divided into upper, middle and lower slopes. The S-shaped sampling method is used to collect 0-40cm layer of soil, among which 15-20 sample points are randomly selected for each slope, and the sample weight of each sample point is not less than 100g. After fully mixing the samples of each slope, take 1kg by quartering method, put them into plastic bags, number them, register the sample information, take them back to the laboratory to air-dry, and test the contents of various fertility indicators. The specific division results of the test and the content of soil fertility indicators in each county are shown in Table 1, Table 2 and Table 3.

Table 1 Soil measurement results of yield greater than 150kg·ha ^-1

Table 2 Soil measurement results of yield 75-150kg·ha ^-1

Table 3 Soil measurement results of yield less than 75kg·ha ^-1

In Table 2, the average value of the soil fertility test results of the counties whose oil production is in the range of 75-150kg·ha ^-1 is taken as the standard value Si of the fertility index (see Table 4). Compared with the Si ¹ index of cultivated land soil fertility in the southern region specified in NY/T 1749-2009, the relative standard deviations are less than 10% for available iron, available manganese, total nitrogen, rapid nitrogen, and rapid potassium. The large deviations are available phosphorus, total phosphorus, available copper and organic matter, and the rest are medium deviations, which is consistent with the characteristics of the acidic red soil phosphorus and organic matter of camellia oleifera, and it also shows that these indicators may be the limiting factors of camellia oleifera production.

Si standard value of each index in table 4

Evaluate _the soil fertility indicators in Table 1, Table 2 and Table 3, and calculate the Nemerow individual fertility index P _i and the Nemerow comprehensive fertility index P in each area. The calculation formulas are as (I) and (II) Shown:

Nemerow individual fertility index evaluation: P _i ＝C _i /S _i (I)

In the formula (I), P _i : the individual fertility index of a certain index i in the soil; C _i : the measured data of a certain index i in the soil; S _i : the standard value of a certain index i in the soil;

Comprehensive fertility evaluation index:

In formula (II), (C _i /S _i ) ² min: the square of the minimum value of a single fertility index; (C _i /S _i ) ² ave: the average square of all fertility indices in the soil; n: the number of soil fertility indices ;

The _{comprehensive} calculation results of Nemerow's individual fertility index P _i and Nemerow's comprehensive fertility index P in each region are shown in Table 5;

Table 5 Soil evaluation results of each county

For the Ci/Si index of PH in each region, when PH≤5, Ci/Si=1; when the pH range is 5.0-5.5, Ci/Si=1.5;

In Table 5, the correlation between the Nemerow composite index and yield was analyzed, and the results showed that the correlation coefficient between the two was 0.348, and the correlation between the two was poor. The reason may be that some fertility indicators have no direct correlation with yield, so the internal The Merrow composite index cannot evaluate the fertility index of soil oil production, and the construction of the minimum data set fertility index that can reflect the output level of camellia oleifera is of great significance to the promotion of camellia oleifera industry.

Example 2:

The minimum data set fertility index construction:

The index content measured in Table 1, Table 2 and Table 3 of Example 1 utilizes SPSS software to carry out principal component analysis, and the results are as shown in Table 6;

Table 6 Eigenvalues and contribution rates of principal components of each factor

According to Table 6, according to the principle that the characteristic value is greater than 1 and the cumulative contribution rate is greater than 70%, 5 principal components are extracted; according to the calculation formula of the principal components, the results of principal component analysis are shown in Table 7;

Table 7 Rotation component matrix

It can be seen from Table 7 that the main factors determining the size of the main component 1 are exchangeable calcium, magnesium, potassium and pH, which can be classified as exchangeable nutrients. The main components that determine the size of principal component 2 are total nitrogen, organic matter and total phosphorus, which can be classified as total nutrients. The main components that determine the size of the main component 3 are available copper, zinc, and manganese, which can be classified as trace elements. The main components that determine the main component 4 are fast nitrogen and fast phosphorus, which can be classified as fast-acting nutrients. What determines the main component 5 is mainly available iron. It can be seen that the 5 principal components extracted by principal component analysis have agronomic significance and can explain most of the variation of the whole quantity.

The high factor loads of the above five principal components are selected based on the principle that the absolute value of the high factor loads is greater than 90% of the maximum factor loads in the principal components.

It can be seen from Table 7 that the maximum factor loading in principal component 1 is exchangeable calcium 0.879, and the index whose absolute value is greater than 90% of the maximum factor loading is exchangeable calcium, magnesium, and pH. Considering that quick potassium is an important indicator of soil nutrients, and its factor load is slightly less than 90% of the maximum factor load, in order to ensure the comprehensiveness and accuracy of the evaluation, quick potassium is also included in the high factor load, that is, the high factor load of principal component 1 Exchangeable calcium, magnesium, pH and tachypotassium. The high factor loading of main component 2 is total nitrogen and organic matter, the high factor loading of main component 3 is effective copper, the high factor loading of main component 4 is fast phosphorus, and the high factor loading of main component 5 is effective iron. Because principal component 1 and principal component 2 have multiple high factor loads, the correlation of high factor loads in principal component 1 and principal component 2 was analyzed separately, and the results are shown in Table 8.

Table 8 High factor loading correlation

In Table 8, ** indicates a significant correlation at the 0.01 level (two-sided) * indicates a significant correlation at the 0.05 level (two-sided).

The criteria for selecting high factor loadings of the principal components into the minimal dataset are:

(a) When there is only one high factor loading indicator of a principal component, the indicator enters the minimum data set;

(b) When there are two high factor loading indicators of a principal component, do a correlation analysis on the two indicators. If the absolute value of the correlation coefficient r<0.4, both indicators will enter the minimum data set; if the absolute value of the correlation coefficient If r≥0.4, the indicators with high factor loads will enter the data set;

(c) When there are three or more high factor loading indicators of a principal component, the correlation analysis is performed in pairs. (1) When the absolute value of the pairwise correlation coefficient of all indicators is r<0.4, all indicators enter the minimum Data set; (2) When the absolute value of pairwise correlation coefficient r ≥ 0.4 exists for all indicators: if the number of indicators with absolute value of pairwise correlation coefficient r ≥ 0.4 accounts for more than 70% of the total number of indicators, the correlation coefficient with other indicators The high factor loading indicator with the largest sum of absolute values enters the dataset; otherwise all indicators enter the dataset.

The correlation analysis in Table 8 shows that the absolute values of the correlation coefficients between exchangeable calcium and exchangeable magnesium and other indicators in principal component 1 are greater than 0.4, and the sum of the absolute value of the correlation coefficients between exchangeable magnesium and other indicators is 2.228, which is the largest value. Enter the minimum data set. However, the sum of the absolute values of the correlation coefficients between exchangeable calcium and other indicators is 2.178, the difference between the two is very small, and because exchangeable calcium and magnesium are relatively key fertility indicators for Zhejiang Camellia oleifera soil, exchangeable calcium is also selected as the minimum data set. The high factor loadings in principal component 2 include total nitrogen and organic matter; the correlation analysis in Table 8 shows that the correlation coefficient between the two is 0.926, which is extremely significant, and the factor loading of total nitrogen is greater than that of organic matter, so total nitrogen is selected as the smallest data set. Principal Component 3, Principal Component 4, and Principal Component 5 all have only one high factor loading, so they are selected as the minimum data sets for effective copper, fast phosphorus, and effective iron. Thus, the minimum data set for the evaluation of Camellia oleifera soil fertility in Zhejiang Province is 6 indexes including total nitrogen, available phosphorus, exchangeable calcium, exchangeable magnesium, available copper and available iron.

The data set does not contain fast potassium, which may be related to the fact that the effect of potassium on yield has not reached a significant level reported in the literature. In recent years, the application of nitrogen, phosphorus and other macronutrients has generally been paid attention to in the transformation of low-yielding forests. However, the application of medium and trace elements has not yet attracted enough attention. Calcium-magnesium-copper-iron is an essential nutrient element for plant growth and development. Calcium-magnesium-copper-iron elements play an important role in promoting the growth of camellia oleifera. Therefore, the minimum data set including macroelements, moderate elements and trace elements is in line with the growth characteristics of camellia oleifera.

The minimum data set constructed in this example is total nitrogen, fast phosphorus, exchangeable calcium, exchangeable magnesium, available copper, and available iron.

Example 3:

The establishment of the minimum data set fertility index:

The standardization of soil fertility indicators is an important part of soil quality evaluation. The relationship between soil total nitrogen, available phosphorus, exchangeable calcium and magnesium, available copper and available iron and crop growth is a parabolic curve, that is, the indicators have an optimal value for crop growth and development. The growth range, the corresponding curve can be converted into a broken line function by determining the critical value, as described in Chinese literature: Soil Quality Index and Evaluation (Xu Jianming, Zhang Ganlin, Xie Zhengmiao, etc., Science Press, 2010). The present invention draws up critical value according to document: the data of Zhejiang forestry soil (Ye Zhongjie, Chai Xizhou, Zhejiang Science and Technology Press, 1986) and this research data, critical value result is as shown in table 9;

Table 9 Critical value of Camellia oleifera soil membership function

According to Table 9, the minimum data set indicators collected by embodiment 2 in each region are standardized, and the normalization formula is as follows:

The minimum data set after standardization, by the method of principal component analysis in embodiment 2, obtains the common factor variance of each index, utilizes common factor variance to account for the total variance ratio to obtain each index weight value coefficient, and then calculates the fertility index of the minimum data set, Its calculation formula is as follows:

In the above formula, Wi: weight coefficient of fertility index; Fi: standard value of fertility index.

The specific values of the NFI of each county in Zhejiang Province calculated in this embodiment:

Table 10 NFI Values of Counties in Zhejiang Province

Example 4:

Calculate the fertility index NFI of the minimum data set in each region of Zhejiang Province by embodiment 3, to the NFI plot that the oil production of each region and Nemerow index evaluation index and embodiment 3 calculate in the table 5 of embodiment 1, The result is shown in Figure 1;

It can be seen from Figure 1 that the Nemerow comprehensive index is consistent with the fertility index of the smallest data set and the overall trend of yield, and the correlation analysis of the three is carried out, and the analysis results are shown in Table 10;

Table 10 Correlation coefficient of Nemerow comprehensive index, minimum data set fertility index and yield

In Table 10, **.significantly correlated at the 0.01 level (two-sided) *significantly correlated at the 0.05 level (two-sided).

As can be seen from Table 10, the minimum data set fertility index is significantly correlated with the output, and is extremely significantly positively correlated with the Nemerow composite index, indicating that the minimum data set fertility index can be used to evaluate Camellia oleifera soil fertility, and the present invention solves the complex soil fertility quality problem. Evaluate more difficult questions. Therefore, by calculating the minimum data set fertility index of the soil in each region, the Camellia oleifera production in the region can be predicted.

The above-described embodiment is only a preferred solution of the present invention, and is not intended to limit the present invention in any form. For those skilled in the art, other modifications can be easily realized without violating the claims and equivalents. The invention is not limited to the specific details, while the general concepts are defined in scope.

Claims (6)

1. a kind of soil fertility prediction technique building geographical MDS minimum data set based on yield, which is characterized in that including following step Suddenly：

(1) it samples：According to sample yield sets sampling position, acquires oil tea pedotheque；

(2) it detects：Every fertility index content of the oil tea pedotheque of detecting step (1) acquisition；

(3) to the numerical value of step (2) detection, full dose data set Data Management Analysis is carried out：

Nei Meiluo individual event fertility index assessments：Pi=Ci/Si

Pi：The individual event fertility index of certain index i in soil；

Ci：The measured data of certain index i in soil；

Si：The standard value of certain index i in soil.

Nei Meiluo comprehensive fertility index assessments：

P is comprehensive：Soil fertility composite index； (Ci/Si)²min：Individual event fertility Index Min square

(Ci/Si)²ave：The mean square of all fertility indexes in soil；n：Soil fertility index number

(4) MDS minimum data set Data Management Analysis：

S1：Principal component analysis is carried out using SPSS data processing softwares to the surveyed index content of step (2), is more than according to characteristic value 1 and contribution rate of accumulative total be more than 70% principle, extract several principal components；

S2：By certain standard, the high factor loading of several principal components is chosen；

S3：By certain standard, the high factor loading for choosing above-mentioned several principal components enters MDS minimum data set；

(5) MDS minimum data set fertility index standardizes：

MDS minimum data set obtained by step (4) carries out the standardization of fertility index using S π membership functions.

Wherein x1 is the low critical value of fertility index；X2 is the high critical value of fertility index；X is the measured value of fertility index.

(6) foundation of MDS minimum data set fertility index

MDS minimum data set after step (5) standardization, acquires communality using Principal Component Analysis, determines therefrom that weight system Number calculates soil index of fertilizer NFI according to following formula.

Wi：Fertility index weight coefficient；Fi：Fertility index standard value

(7) MDS minimum data set fertility index does correlation analysis with yield.

MDS minimum data set fertility index and full dose data set composite index and yield are done into correlation data analysis respectively, as a result most Small data set fertility index and correlation with yield are preferable, and the MDS minimum data set fertility index can predict soil fertility.

2. a kind of prediction technique of soil fertility building geographical MDS minimum data set based on yield according to claim 1, It is characterized in that, in the step (1), the method for sampling is：Sample it is divided into upper, middle and lower slope, using S-shaped sampling method, acquisition 0~ 40cm layers of soil.

3. a kind of prediction technique of soil fertility building geographical MDS minimum data set based on yield according to claim 2, It is characterized in that, each gradient takes 15~20 sampling points at random, the example weight per sampling point is not less than 100g, the sample of each gradient Product take quartering to leave and take 1kg after mixing well, and are air-dried.

4. a kind of prediction technique of soil fertility building geographical MDS minimum data set based on yield according to claim 1, It is characterized in that, in the step (2), fertility index pH, full nitrogen, full phosphorus, organic matter, fast nitrogen, available phosphorus, available potassium, soil Loamy texture, effective trace elements iron, copper, zinc, manganese, exchangeable calcium, exchangeable magnesium.

5. a kind of prediction technique of soil fertility building geographical MDS minimum data set based on yield according to claim 1, It is characterized in that, in the step (4), the selection standard of high factor loading is in principal component：The absolute value of the high factor loading More than 90% of maximum factor loading in the principal component.

6. a kind of prediction technique of soil fertility building geographical MDS minimum data set based on yield according to claim 1, It is characterized in that, in the step (4), the high factor loading for choosing principal component enters the standard of MDS minimum data set and is：

(a) when a principal component high factor loading index only there are one when, then the index enters MDS minimum data set；

(b) when the high factor loading index of a principal component is two, correlation analysis is done to two indices, if related coefficient is exhausted When to value r ＜ 0.4, then two indices all enter MDS minimum data set；If related coefficient absolute value r >=0.4, factor loading are high Index enters data set；

(c) when the high factor loading index of a principal component is three or three or more, correlation analysis is done respectively two-by-two, (1) All indexs two-by-two related coefficient absolute value r ＜ 0.4 when, then all indexs all enter MDS minimum data set；(2) all indexs are two-by-two When the case where related coefficient absolute value r presence >=0.4：If the index quantity of related coefficient absolute value r >=0.4 accounts for overall performane two-by-two 70% or more of quantity then enters data set with the maximum high factor loading index of the sum of remaining index related coefficient absolute value； Otherwise all indexs enter data set.