Spatial Coupling Between Specialized Production of Fruits and Vegetables and Front-End Cold-Chain Logistics

2.1 Location quotient

Location quotient reflects the degree of specialized production. It measures the distribution of the production factors of an industry in various regions. The quotient is often adopted to judge whether an industry is sufficiently specialized in a region. The location quotient of industry j in region i in year t can be calculated by:

$L{{Q}_{ijt}}=\frac{{{{e}_{ijt}}}/{{{e}_{it}}}\;}{{{{E}_{jt}}}/{{{E}_{t}}}\;}$     (1)

where, e_ijt is the output of industry j in region i in year t; e_it is the output of major APs in region i in year t; E_jt is the total output of industry j in Xinjiang in year t; E_t is the total output of major APs in Xinjiang in year t.

In this paper, the location quotients of vegetable industry and fruit industry are denoted as LQ_V and LQ_F, respectively.

2.2 Dynamic factor analysis

Dynamic factor analysis is a multivariate statistical method that effectively combines principal component analysis (PCA) and linear regression. During the analysis, a 3D array is constructed based on samples, variables, and time, and used to compare the states of different regions in both spatial and temporal dimensions. The specific steps of dynamic factor analysis are as follows:

Step 1. Normalize the 3D data to eliminate the impact of dimensional difference;

Step 2. Derive the mean covariance matrix S_T from the covariance matrix S(t) of each year:

${{S}_{T}}=\frac{1}{T}\sum\nolimits_{t=1}^{T}{S(t)}$    (2)

Step 3. Solve the eigenvalues and eigenvectors of the mean covariance matrix S_T, and find the variance contribution rate and cumulative variance contribution rate of each common factor.The mean covariance matrix fully displays the impacts from static structural difference and dynamic changes of data.

Step 4. Extract the common factors and establish the initial factor loading matrix.

Step 5. Construct the static score matrix of each sample:

${{C}_{ih}}=(\overline{{{Z}_{i}}}-\overline{Z})'{{a}_{h}},\text{ }i=1,2,\ldots I;t=1,2,\ldots T$     (3)

where, $\overline{Z_{i}}=\frac{1}{T} \sum_{t=1}^{T} Z_{i t}$  is the mean vector of each sample; $\bar{Z}=\frac{1}{I} \sum_{i=1}^{I} \bar{Z}_{i}$ is the mean vector of the sample set.

Step 6. Calculate the dynamic score of each sample:

${{C}_{iht}}=({{Z}_{it}}-\overline{{{Z}_{t}}})'{{a}_{h}},h=1\ldots k,t=1\ldots T$    (4)

where, $\bar{Z}_{t}=\frac{1}{I} \sum_{i=1}^{I} Z_{i t}$ is the mean of each variable in year t.

Step 7. Calculate the mean score E based on the dynamic scores of the samples in each year:

$E=\frac{1}{T}\sum{{{C}_{iht}}}$     (5)

2.3 Coupling coordination model

The coupling degree describes how much systems (subsystems) or elements interact with and influence each other. It is a mirror of the static coupling state between systems (subsystems) or elements. Coupling coordination degree reflects the degree of harmony and consistency between systems or elements during development, i.e. the dynamic changes and variation in coordinated state of a system in disorderly (low-level) and orderly (high-level) development.

Inspired by the physical model of capacity coupling coefficient, a coupling degree model was established for the level of F&V specialized production and the quality of FECC logistics:

${{C}_{t}}=\sqrt{\frac{{{U}_{xt}}\times {{U}_{yt}}}{{{({{U}_{xt}}+{{U}_{yt}})}^{2}}}}$    (6)

The coupling degree model accurately measures the interaction and influence between systems (subsystems) or elements, yet fails to reflect the coordination degree of the interactive development. For example, the coupling degree model could output a high value if F&V specialized production and FECC logistics are both immature in a region. It is impossible to know whether the two factors have synergistic effect, or how coordinated is their interactive development.where, U_xt is the level of F&V specialized production in year t; U_yt is the quality of FECC logistics in year t. The coupling degree C_t is positively correlated with the relevance between the two factors. A high C_t means F&V specialized production and FECC logistics interact with and influence each other; a low C_t means the two factors each develops in its own ways.

To solve the problem, the coupling degree model was extended into a coupling coordination degree model:

${{D}_{t}}={{(C\times T)}^{\frac{1}{2}}}$, $T=\alpha {{U}_{xt}}+\beta {{U}_{yt}}$    (7)

The coupling coordination degree D_t is positively correlated with the coordination between the two factors during coupling development. A high D_t means F&V specialized production and FECC logistics are well coordinated; a low D_t means F&V specialized production and FECC logistics are poorly coordinated.where, T is the coordination index about how much the overall development of F&V specialized production and FECC logistics contributes to the coupling coordination degree; α and β are two undetermined parameters, respectively reflecting the importance of the level of F&V specialized production and the quality of FECC logistics. Both factors are equally important to the coordinated development process. Thus, the values of α and β were both set to 0.5.

2.4 Evaluation index system

Evaluation index system is an organic collection of indies that reflect the various features and their correlations of the evaluation object(s). The system must be reasonable and operable, and describe the object(s) in a systematic and objective manner. Considering our research objects and purposes, this paper sets up an evaluation index system for the quality of FECC logistics (FECCLS) in F&V producing regions based on expert survey and relevant literature.

Once harvested, F&V must be always kept at a low temperature, making FECC facilities a must have. The quality of F&V FECC logistics is affected by all the relevant elements in the whole logistics process. Therefore, our evaluation index system should consider relevant industries, operating environments, as well as suppliers and consumers, and have both similarities with and differences from the evaluation index system for regional logistics development.

Table 1. Our evaluation index system

Note: 1. “F&V import and export” is measured by the local standards of importers and exporters (Source: Urumqi Customs District); 2. “Agricultural machinery transport volume” refers to the total volume (unit: ton·km) of field transport, highway transport, and water transport by various agricultural machinery in the current year (Source: Statistical Reporting System of Xinjiang Agricultural Machinery Administration); 3. “Number of cold-storage equipment” refers to the number of cold-storage equipment used to store the harvested APs (Source: Statistical Reporting System of Xinjiang Agricultural Machinery Administration); 4. “Total area of facilities” refers to the total area of polytunnels, solar greenhouses, and multi-span greenhouses in the current year (Source: Statistical Reporting System of Xinjiang Agricultural Machinery Administration); 5. All the other tertiary indices were collected from the Xinjiang Statistical Yearbooks.

4.1 Coupling coordination analysis

According to relevant literature, the coupling coordination degree was evenly divided into five levels: strongly uncoordinated (0.0-0.2 point), slightly uncoordinated (0.2-0.4 point), slightly coordinated (0.4-0.6 point), moderately coordinated (0.6-0.8 point), and strongly coordinated (0.8-1.0 point). The coupling coordination between the level of F&V specialized production and the quality of FECC logistics is illustrated in Figure 3.

As shown in Figure 3, F&V specialized production and FECC logistics became more and more coordinated with the elapse of time. Overall, the level of coupling coordination gradually shifted from strongly uncoordinated to strongly coordinated.

4.1.1 Coupling coordination of vegetable industry

The coupling coordination degrees between the specialized production level of vegetables and the quality of FECC logistics in different regions of Xinjiang were calculated by formulas (6) and (7), and sorted in descending order (Table 5). The coupling coordination degree of each region in each year was evaluated against the scale in Figure 3.

As shown in Table 5, Urumqi is the only moderately coordinated region in terms of the coupling coordination of specialized production level of vegetables and the quality of FECC logistics; Changji and Bayangol are slightly coordinated; Kashgar, Karamay, Ili, Tacheng, Aksu, Hami, and Turpan are slightly uncoordinated; Bortala, Hotan, Kizilsu, and Altay are strongly uncoordinated. The different levels of coupling coordination reflect the regional difference in the development of specialized production of vegetables and FECC logistics.

Currently, only a few regions, such as Urumqi, Changji, and Bayangol, are in coordinated states. Urumqi is moderately coordinated, while the latter two regions are slightly coordinated. Meanwhile, most regions, namely, Kashgar, Karamay, and Bortala, are in uncoordinated states.

About 64% of all uncoordinated regions belong to the slightly uncoordinated level. Among them, Kashgar, Karamay and Ili are gradually approaching the slightly coordinated state. There are fewer strongly uncoordinated regions concerning the specialized production of vegetables than those concerning the specialized production of fruits. Among the strongly uncoordinated regions, Bortala and Hotan are improving in coupling coordination level.

Over the time, the coupling coordination of specialized production level of vegetables and the quality of FECC logistics is increasing in all regions of Xinjiang, except Altay.

Table 5. Coupling coordination degrees between the specialized production level of vegetables and the quality of FECC logistics in different regions of Xinjiang

3.png

Figure 3. Coupling coordination between the level of F&V specialized production and the quality of FECC logistics

Table 6. Coupling coordination degrees between the specialized production level of fruits and the quality of FECC logistics in different regions of Xinjiang

4.1.2 Coupling coordination of fruit industry

The coupling coordination degrees between the specialized production level of fruits and the quality of FECC logistics in different regions of Xinjiang were calculated, and sorted in descending order (Table 6). The coupling coordination degree of each region in each year was also evaluated against the scale in Figure 3.

As shown in Table 6, Aksu, Kashgar, and Turpan are slightly coordinated in terms of the coupling coordination of specialized production level of vegetables and the quality of FECC logistics; Ili, Changji, Bayangol, Hami, Urumqi, and Hotan are slightly uncoordinated; Tacheng, Bortala, Altay, Kizilsu, and Karamay are strongly uncoordinated. Overall, 21% of all regions in Xinjiang are slightly coordinated, 43% are slightly uncoordinated, and 36% are strongly uncoordinated.

The slightly coordinated regions have always achieved the highest coupling coordination degrees. This means the specialized production of vegetables has entered the state of stable coordinated development with FECC logistics, with prominent interactions between the two factors.

The slightly uncoordinated regions have relatively low coupling coordination degrees. Hotan is the only slightly uncoordinated region, whose coupling coordination degree rises steadily. The coupling coordination degrees of the other slightly uncoordinated regions are still fluctuating. With the elapse of time, the coupling coordination levels are rather unstable. The changes in Bayangol and Hami are relatively small, indicating that the specialized production of fruits and FECC logistics in the two regions are slightly affected by uncertain factors.

The strongly uncoordinated regions lag far behind the slightly coordinated regions in coupling coordination degree, owing to their weaknesses in resources, geopolitics, social economy, and infrastructure that suit specialized production of fruits and FECC logistics. It is difficult for the two factors to realize coordinated development in a short time, despite the measures taken by the government.

4.2 Causes of coupling differences between vegetable industry and fruit industry

The above analysis shows that the dynamic coupling coordination between the level of specialized production of fruits and FECC logistics is not as high as that between the level of specialized production of vegetables and FECC logistics. The main reasons are as follows:

First, there are differences between fruit industry and vegetable industry in the spatial pattern of specialized production levels. For specialized production of fruits, the superior regions concentrate in less-developed counties in the south and east; these countries have vast stretches of open ground. For specialized production of vegetables, the superior regions center on Urumqi, Changji and other important cities. Hence, vegetable industry has stronger logistics supports than fruit industry.

Second, there are differences between fruit industry and vegetable industry in planting conditions, planting techniques, and the completeness of agricultural facilities. In vegetable industry, agricultural facilities like polytunnels, solar greenhouses, and multi-span greenhouses are widely adopted, such that vegetables can be planted even if the geographical environment is unfavorable. By contrast, such facilities are not widely applied in the fruit industry, which greatly limits the plan table varieties.

Third, there are differences between fruit industry and vegetable industry in industry development and consumer demand. As basic food sources for humans, vegetables provide necessary vitamins, dietary fiber, and minerals. The traditional diet in China has a high demand for fresh foods like vegetables. That is, the stable supply of vegetables is a matter of people’s livelihood. By contrast, the distribution channels of fruits are more market-oriented and diverse. High-quality fruits are often sold as high-end products for a small group of consumers, creating a certain amount of added value. The difference in product value results in disparity between vegetables and fruits in the elasticity of supply and demand.

4.3 Causes of spatial coupling differences

The mean LQ_F, mean LQ_V, and mean quality of FECC logistics in each region of Xinjiang were prepared into two scatterplots (Figure 4). Each scatterplot was split into four quadrants by the LQ axis and the critical line of LQ=1.

Then, the causes of spatial coupling differences were analyzed by comparing the left subgraph with Table 5, and the right subgraph with Table 6. The coupling coordination levels of regions in the four quadrants are summarized in Table 7 below.

It can be seen that the F&V specialized production and FECC logistics can achieve coupling and coordinated development (slightly, moderately or strongly coordinated), only if both factors are on high levels; the development will be uncoordinated, if the F&V specialized production is near the critical line or the FECC logistics is of poor quality.

In the first quadrant (high specialization and good logistics), the mean score of FECC logistics is positive, and the LQ_V or FQ_V is greater than 1. The regions in this quadrant have achieved a certain scale in the specialized production of vegetables or fruits. The FECC logistics in these regions is of high quality, under the positive spillover effects of social economy, political geography, infrastructure, and population density. The advanced FECC logistics manages to radiate the surroundings and link up adjacent regions.

In the second quadrant (low specialization and good logistics), the mean score of FECC logistics is positive, and the LQ_V or FQ_V is smaller than 1. The regions in this quadrant are not good at specialized production of vegetables or fruits, under constraints of environment, agricultural location, and planting conditions. The lack of specialization does not match the advanced FECC logistics, resulting in uncoordinated development.

In the third quadrant (low specialization and poor logistics), the mean score of FECC logistics is negative, and the LQ_V or FQ_V is smaller than 1. The regions in this quadrant are neither good at specialized production of vegetables or fruits, nor advanced in FECC logistics. Hence, the two factors cannot develop coordinately, leading to strong incoordination. It is possible for slightly uncoordinated regions to gradually improve their coupling coordination level. Under unchangeable constraints like resource endowment, the strongly uncoordinated regions (e.g. Altay) are not suitable produce vegetables or fruits, and should develop other special APs. The geography, social economy, or population in such regions cannot effectively drive the growth of FECC logistics.

In the third quadrant (high specialization and poor logistics), the mean score of FECC logistics is negative, and the LQ_V or FQ_V is greater than 1. Many fruit producing regions fall in this quadrant, and differ in the level of dynamic coupling coordination. Some of them are slightly coordinated, some are slightly uncoordinated, and some are strongly uncoordinated. This means poor FECC logistics does not necessarily bring incoordination, but regions with poor FECC logistics could achieve slight coordination at the most. In this quadrant, the multiple superior regions of specialized production of fruits have relatively low coupling coordination, owing to the insufficient services of the FECC logistics. There are immense difficulties in building FECC logistics infrastructure, because these regions cluster in remote villages of southern Xinjiang. The logistics capacity is far smaller than the actual demand for fruits. For instance, Turpan and Kizilsu, both have advantages in specialized production of fruits, are faced with strong incoordination arising from poor FECC logistics.

In summary, the spatial coupling coordination between F&V specialized production and FECC logistics is relatively low in Xinjiang. In many regions, the two factors restrict the development of each other, failing to achieve mutual progress or coordinated development. To improve the dynamic coupling coordination, different strategies should be adopted for vegetable industry and fruit industry. For vegetable industry, there is ample room for growth in specialization level; more regions should be encouraged to develop agricultural facilities, paving the way for specialized, scale production. For fruit industry, superior producing regions are highly concentrated. Each region has its unique strength in specialized production, yet lacks supports from FECC logistics. The causes of incoordination or slight coordination should be identified based on the quadrant of each region, before putting forward effective countermeasures.

41.png

42.png

Figure 4. Scatterplots of LQ_F, LQ_V and quality of FECC logistics in each region of Xinjiang