Siting Principles of the Ancient Postal Buildings Under Environmental Constraints

Abstract

Human–environment interactions in antiquity were fundamentally shaped by environmental constraints, with spatial patterns of human construction works reflecting strategic resource optimization. This study employed Geographic Information System (GIS) and binary logistic regression (BLR) to analyze the siting principles of ancient postal buildings in Fujian, China, integrating related environmental factors of elevation, slope, relief amplitude, and distance to rivers. The results revealed significant spatial differentiation, with elevation exhibiting the strongest influence on siting preference, followed by slope, relief amplitude, and distance to rivers. Clustering patterns along coasts and rivers indicated a strategic balance between transmission efficiency and military defense needs. The applicability of the integrated GIS–BLR approach in studying the ancient postal system demonstrates its extensibility to other ancient settlement systems while offering insights for contemporary conservation practice and sustainable development.

1. Introduction

The postal system is of long standing and leaves rich heritage in the world’s communication history. Documented communication activities can be traced back to around 2000 BC in ancient Egypt [1] and approximately 1600 BC in China’s Shang Dynasty [2]. As a vital tool of imperial governance, the postal system was vigorously developed, with relay transmission becoming prevalent in the ancient world. This established an unobstructed, convenient, and efficient network across empires for long-distance communication. In the 6th century BC Persian Empire, cavalries served as messengers [1]. During the 1st century BC, the Romans established the cursus publicus to handle delivery tasks, integrating it into their military and administrative systems until the Byzantine period [3]. Meanwhile, China’s Qin Dynasty unified the postal system, serving the empire for nearly 2000 years [2].

Modern China preserves outstanding postal heritage, some recognized as World Cultural Heritage for their outstanding values, such as the Yucheng Yi (Figure 1) along the Grand Canal [4]. Most extant postal sites date from the Ming Dynasty (1368–1644). When founding the Ming Empire in 1368, Zhu Yuanzhang ordered the rectification of the postal system to consolidate the centralized authoritarian rule. The Ming postal system comprised three facility types, Yizhan (驿站, post station), Diyunsuo (递运所, delivery station that specialized in transporting goods), and Jidipu (急递铺, urgent post station) [5], which not only facilitated the dissemination of governmental decrees and interregional economic–cultural exchanges but also played a vital role in transmitting military information and supplies, making it an indispensable element of the military defense with significant historical and military values. Academic research on the Ming postal system has achieved substantial depth, primarily within historical scholarship. Studies on the historical development and institutional framework established a theoretical foundation for understanding the Ming postal system [2,6,7]. Scholars have systematically documented its structure [8], management system [9], operational mechanisms [10], contributions [11], and collapse causes [12]. Regarding individual facilities, Z. Yang’s comprehensive study of Yizhan [13], Zheng’s research on Diyun suo [14], and Lin’s investigation into Jidipu [15] addressed critical gaps on the micro-level perspectives, offering essential context for this study focusing on postal buildings. Collectively, these achievements formed a theoretical framework for the Ming postal research, laying a solid foundation for follow-up research. It is noteworthy, however, that, due to the massive scale and numerous constituent units of the postal system combined with the incomplete records, significant gaps remain in the field of spatial analysis. The only representative achievement is the Ming Dynasty Yizhan Database by Harvard University [16], yet the spatial distribution of Diyunsuo and Jidipu remain unclear, with their integrated spatial patterns and siting principles remaining largely unexplored.

Conducting spatial analysis is crucial for comprehensively understanding the ancient postal system. The spatial distribution of cultural heritage inherently links to the natural and anthropogenic environmental factors. Analyzing the relationship between cultural heritage distribution and the environment can reveal spatial patterns across different countries or regions. This analytical approach positively contributes to understanding historical human activities while deepening knowledge of human–land relationships, including cultural exchanges, social transformations, and paleoenvironmental evolution [17]. By the 1970s, discussions on the impact of spatial environment on cultural heritage had emerged, addressing both natural factors (e.g., topography, climate) and anthropogenic dimensions (e.g., productive capacities, cultural practices, technological development) [18]. Simultaneously, attention was paid to the spatial–temporal evolution dynamics, spatial–temporal distribution characteristics, and environmental evolution of the heritage [19]. Recent advances in remote sensing and GIS technologies enhanced methodological capabilities for cultural heritage studies. GIS spatial analysis tools, such as Slope, Aspect, and Kernel Density, have been widely used to extract spatial information of heritage sites [20,21]. Concurrently, statistical methods analyze, evaluate, and interpret these data, quantifying associations between heritage distributions and environmental parameters [22]. The integration of GIS and statistics exhibits strong potential within cultural heritage research, enabling applications such as functional performance assessment [23], identification of influencing factors [24], and quantitative analysis of spatial characteristics [25]. This combined approach significantly advances the study of human–environment interactions, not only greatly expanding the field of traditional spatial analysis but also conceptually revolutionizing the archaeological interpretation of space, assisting heritage conservation. Consequently, this integrated method provides a directly transferable methodological framework for investigating the spatial characteristics of the ancient postal system.

Given the resource limitations in ancient China, postal buildings were strategically sited at locations with optimal accessibility and natural advantages to maximize transmission efficiency, demonstrating strong environmental dependency. Thus, systematically analyzing the siting principles of the postal buildings and exploring their spatial distribution patterns form the pivotal objectives for this study. Building on the above studies, we adopt the integrated approach of GIS and statistical analysis to investigate the relationship between the ancient postal system and its environment. The methodological framework (Figure 2) comprises four stages. First, spatial datasets including all documented postal buildings (containing names, administrative jurisdictions, geospatial position, inter-station distances) and related environmental parameters (terrain and river network) are integrated in the GIS platform for visualization. Secondly, critical environmental data (including terrain elevation, slope, aspect, relief amplitude, and distance to rivers) are extracted for the correlation test, identifying significant factors associated with postal system distributions. Subsequently, taking the validated environmental factors as the independent variables, with postal buildings and random points as dependent variables, binary logistic regression modeling is implemented to detect the siting preference. Finally, the spatial distribution patterns and underlying reasons of the postal system are analyzed within the historical and geographical contexts, providing theoretical support for formulating sustainable conservation strategies for postal heritage.

2. Materials and Methods

2.1. Study Area

This study focuses on Fujian during the Ming Dynasty, as archaeological surveys indicate that most extant postal heritage in China originate from this period. Administratively, Fujian comprised 9 prefectures and 57 counties [26], covering approximately 12.4 km² (Figure 3).

Fujian was situated along the southeastern maritime frontier of the Ming territory. Its landscape featured rugged mountains, continuous gullies, and intertwined rivers, forming complex geographical conditions [27]. Economically, Fujian served as an essential logistics hub throughout the Ming territory [28]. Militarily, it faced persistent Japanese pirate incursions since the dynasty’s founding, making coastal security precarious. Therefore, the Ming government built a dense of Wei forts (卫城, primary defensive fortresses) and Suo forts (所城, secondary defensive fortresses) along the coast to garrison soldiers and resist invaders [29] (Figure 3b). The dual demands of commercial exchanges and military communication led to a well-constructed postal system in Fujian and left well-documented records that enable systematic research.

Such complexity of the natural environment, density of postal relay system construction, and typicality of functional demands, coupled with relatively sufficient documentary support, make Fujian a representative region for studying ancient postal siting principles and environmental relationships.

2.2. Data Sources

Data of Fujian’s postal system are primarily drawn from three Ming Dynasty official compilations: the official historical book “Ming Hui Dian” [30], the geographic record “Huanyu Tongqu” [31], and the local chronicle “Ba Min Tongzhi” [32]. These documents have detailed records, including names, districts, bearings, routes, and interval mileage for all the postal buildings. Through systematic analysis, 755 postal buildings in Fujian of the Ming Dynasty were identified (see complete dataset in Table S1).

The geospatial data originates from the ASTER GDEM 30M digital elevation data (resolution: 27.08 m × 27.08 m), provided by the Geospatial Data Cloud platform, Computer Network Information Center, Chinese Academy of Sciences [33]. The river network data sources are from the Harvard University Historical Database [34]. Notably, dual verification ensured data reliability: (1) 85.4% of Ming-era heritage sites in Fujian (35/41) retain original surface positions, confirming stable terrain elevation; (2) river network data shows consistent trajectories with Ming waterways in historical atlases [35].

2.3. Spatial Localization

Based on collected textual sources, the spatial localization of postal buildings was conducted for further analysis. The methodology systematically combines the following:

• Field investigation for precise geospatial positioning

Although far from now, there are still a few postal buildings that can be traced by their relics, thus the precise geographical coordinates are obtained through field investigation. Accordingly, there are 2 Yizhan built in the Ming Dynasty with remains, of which Shenqing Yi is a remnant of the old post station gate, and Rouyuan Yi was built near the original site by later generations.

2

Extracting spatial data from historical records

For vanished facilities, historical texts provided precise locational clues. For example, in Ouning County, Jianning Prefecture, the “Chengxi Post” was recorded in the “Ba Min Tongzhi” as “Chengxi Post, located in the west of the prefectural city, outside the Tongji Gate”. Searching for “Tongji Gate” within the current administrative map of Jianou City, it is located in the northwest of the city and along the moat, which is consistent with the record.

3

Positioning through historical–modern toponym matching

Toponymic continuity enabled approximate positioning where direct evidence was lacking. For example, in Houguan County, Fuzhou Prefecture, the “Xiaoruo Post” was recorded in the “Ba Min Tongzhi” as “Xiaoruo Post, located in the northwest of the prefectural city, in the 34th Du”. Searching for “Xiaoruo” within the current administrative map of Fuzhou City, “Xiaoruo Village”, which is located in the northwest of Fuzhou, was found, matching the record; then, the approximate location of “Xiaoruo Post” can be identified.

4

Locations speculation based on transport routes

According to the above 3 methods, 64.5% of postal buildings’ locations were identified. For the remaining facilities, the locations can be speculated based on the transmission routes and mileage recorded in historical documents.

Integrating all postal buildings’ location information into WGS 1984 geographic coordinate system in the GIS platform, the spatial dataset of Fujian’s postal system was established (Figure 4). Although we strove to restore the real history, historical document limitations constrained positional precision to town-level resolution. Archaeological evidence indicates typical Ming town dimensions that spanned 1–3 li (0.56–1.68 km) [36]. Consequently, localization errors remained bounded within 2 km, comprising merely 0.4% of Fujian’s macroscopic spatial scale (480 km × 530 km). Critically, the substantial sample size (n = 755) ensures that such micro-scale positional deviations exert negligible impact on macro-scale spatial analysis. The reconstructed dataset preserves key spatial structure and can be used for further analysis.

2.4. GIS Spatial Analysis

Given that humanity preferred to conduct construction activities in environments with flat terrain and abundant water sources since ancient times, this study investigates environmental influences on postal system distribution through two primary dimensions:

• Terrain features: Elevation (vertical distance above sea level), slope (steepness gradient of terrain), aspect (orientation of the terrain slope), and relief amplitude (maximum elevation difference per unit area).
• Distance to rivers.

The data related to these factors were extracted using GIS spatial analysis tools, including the following:

• Slope: Employed for analyzing terrain slope gradient;
• Aspect: Employed for analyzing terrain aspect orientation;
• Block Statistics: Employed for analyzing relief amplitude;
• Euclidean Distance: Employed for calculating the distance to rivers.

Superimposing the raster layers (Figure 5) and the spatial dataset of Fujian’s postal system, the “Extract Multi Values to Points” tool was utilized to extract precise environmental variable values for each postal building (Table S2).

Notably, Euclidean Distance instead of Cost Distance was employed to calculate the distance to rivers, preventing duplicate calculations of terrain factors. The Euclidean Distance tool quantifies straight-line distance of each cell to the nearest source (cell-to-cell distance) [37], primarily serving geographical distance measurement without cost/weight parameters. In contrast, Cost Distance analysis computes shortest weighted paths that integrate terrain complexity, transportation constraints, and other friction factors. Given that terrain factors have already been incorporated as independent parameters within the analytical framework, adopting Cost Distance analysis would cause overcompensation, distorting the spatial modeling outcomes.

2.5. Correlation Test

To identify whether the various environmental factors can be used for further analysis, a correlation test is required. As the distribution of postal buildings within each environmental parameter was highly correlated with the geographic conditions of the region itself, sub-region area proportions critically influence distribution patterns. Therefore, it is necessary to take the area proportion in different sub-regions of each environmental factor into consideration; therefrom, the correlation test adopts Equation (1), as follows:

$$W_{\mathrm{i}}={\displaystyle \frac{A_{\mathrm{i}}}{E_{\mathrm{i}}}}$$

(1)

where A_i is the actual number of postal buildings distributed in sub-region i of the environment factor and E_i is the expected number of postal buildings distributed in the same region. E_i is obtained from Equation (2) as follows:

$$E_{\mathrm{i}}=A{\displaystyle \frac{a_i}{a}}$$

(2)

where A is the total number of Fujian’s postal buildings, a_i is the area of sub-region i for each environmental factor, and a is the total area of Fujian.

Theoretical interpretation of the W_i index follows three principles:

• W_i ≈ 1: Indicates that the postal buildings show a homogeneous distribution in terms of the environmental factors analyzed, without clustering, which means that there is no obvious correlation between the site selection of the postal buildings and this environmental factor.
• W_i > 1: Indicates that the number of postal buildings distributed in this sub-region is higher than the expected value, showing positive clustering. The larger the value, the stronger the clustering, and vice versa.
• W_i < 1: Indicates that the distribution of postal buildings is inversely related to this sub-region.

With this approach, the quantitative relationship between the locations of postal buildings and related environmental factors can be accurately represented.

2.6. Binary Logistic Regression

Logistic function is commonly employed for multi-classification problems involving multiple eigenvectors. The eigenvector for two is binary logistic regression (BLR). For BLR, the dependent variables are only “1” and “0”, where “1” represents “Yes” and “0” represents “No” [38]. BLR computes the probability of belonging to the impact factors based on dependent variable discrimination and can provide an interpretation of the discriminative results in terms of probability.

This study adopted BLR to model the siting preference of Fujian’s postal buildings mathematically, with the sites of postal buildings defined as “1”, and other randomly generated points that are not postal buildings defined as “0”. The areas with high probability are the areas with high preference for siting, from which the siting rules can be recognized. In particular, BLR requires a large number of samples to ensure that the results produce small errors [39]. Whereas the number of Yizhan and Diyunsuo is rare and not suitable for separate analysis, all the postal buildings were taken as a whole.

P is set as the probability of the postal buildings’ siting preference, with a value in the range [1, 0], and 1 − P is the probability of the non-existence of postal buildings. Taking the natural logarithm of ratio ${\displaystyle \frac{P}{1-P}}$ to obtain $\mathit{ln}{\displaystyle \frac{P}{1-P}}$, namely logit transformation of P, and noted as logit P, then the value range of logit P is (−∞, +∞). With logit P as the dependent variable, the linear regression equation is established as follows:

$$LogitP=\alpha +{\beta }_1{\chi }_1+{\beta }_2{\chi }_2+...+{\beta }_{\mathrm{n}}{\chi }_{\mathrm{n}}$$

(3)

A further conversion of Equation (3) gives the p-value, as follows:

$$P={\displaystyle \frac{\exp \left(\alpha +{\beta }_1{\chi }_1+{\beta }_2{\chi }_2+...+{\beta }_n{\chi }_n\right)}{1+\exp \left(\alpha +{\beta }_1{\chi }_1+{\beta }_2{\chi }_2+...+{\beta }_n{\chi }_n\right)}}$$

(4)

Equation (4) is the BLR model, where α is a constant term, $x_n$ is the independent variable, and ${\beta }_n$ is the regression coefficient of BLR [33].

3. Results

According to the analysis steps, the correlation between postal buildings and environmental factors was first verified according to Equation (1), and the results are shown in

Table 1. The correlation test results statistics.

Environmental Factor	Sub-Region	Area (km2)	Ai	Ei	Wi
Elevation (m)	<0	863.80	0	5	0
	0–200	23,793.62	400	147	2.72
	200–500	44,560.31	298	275	1.08
	500–1000	46,664.24	60	288	0.21
	>1000	6857.70	0	42	0
Slope (°)	<2	8337.97	134	51	2.60
	2–6	16,400.16	182	101	1.80
	6–15	42,021.79	248	260	0.96
	15–25	38,323.05	145	237	0.61
	>25	17,656.70	49	109	0.45
Aspect	Flat	1459.84	10	9	1.11
	North	14,266.79	79	88	0.90
	Northeast	14,570.94	76	90	0.84
	East	15,602.82	74	96	0.77
	Southeast	16,687.37	106	103	1.03
	South	15,187.42	103	94	1.10
	Southwest	14,685.20	108	91	1.19
	West	14,910.62	102	92	1.11
	Northwest	15,368.67	100	95	1.05
Relief amplitude (m)	0–5	4789.68	47	29	1.59
	5–15	8430.11	161	52	3.10
	15–45	34,186.23	263	210	1.25
	45–140	72,456.59	281	446	0.63
	>140	3267.88	6	20	0.30
Distance to rivers (km)	0–5	40,239.15	429	245	1.75
	5–15	55,307.78	162	337	0.48
	15–30	24,887.42	122	152	0.80
	30–50	3392.40	43	21	2.08
	>50	466.50	2	3	0.69

. From the statistics, it can be seen that the value W_i of each sub-region within the factor aspect tends to be close to 1, while that of the factors elevation, slope, relief amplitude, and distance to rivers shows a large fluctuation, which means that the siting of the postal buildings was basically uncorrelated with aspect but was related to elevation, slope, relief amplitude, and distance to rivers.

Accordingly, four environmental factors, namely elevation (X₁), slope (X₂), relief amplitude (X₃), and distance to rivers (X₄), were screened as the independent variables of the siting preference analysis model for postal buildings. The “Create Random Points” tool of the ArcGIS 10.3.1 platform was used to generate 700 random points (equivalent to postal buildings in number) as the experimental samples, which, together with the postal buildings, formed the dependent variables for modeling. The distribution map is shown in Figure 6, and then, the environmental data for each dependent variable was derived by the ArcGIS 10.3.1 platform, as shown in Table S3.

Thirdly, BLR analysis was computed by Statistical Package for the Social Sciences (SPSS) Statistics 26 software, using the algorithm “Backward: LR”. Model performance was evaluated through three key metrics:

• The Significance (Sig.) value serves as the criterion for determining the significance of the independent variables. If the Sig. < 0.05, it means that the independent variable has a significant impact on the model.
• The Omnibus Test of Model Coefficients is a likelihood ratio test for evaluating the overall effectiveness of the model. When the p-value < 0.05, it indicates that at least one variable’s odds ratio (OR) in the model is statistically significant, confirming model validity.
• The Hosmer–Lemeshow Test assesses the goodness-of-fit of the model. A p-value > 0.05 demonstrates adequate information extraction from the dataset, signifying high goodness-of-fit.

The calculation results showed that the Sig. values of all the variables (elevation, slope, relief amplitude, and distance to rivers) were less than 0.05, indicating statistically significant impacts on the model. The Omnibus Test confirmed overall model validity (p = 0.000000 < 0.05). The Hosmer–Lemeshow Test indicated high goodness-of-fit (p = 0.240898 > 0.05). Overall, these metrics validate the model’s statistical significance for subsequent analysis.

As presented in

Table 2. The parameters of the siting preference analysis model.

Variable	β 1	S.E. 2	Wald 3	Df 4	Sig. 5	Exp(β) 6
Elevation (X1)	−0.004444	0.000317	197.105177	1	0.000000	0.995566
Slope (X2)	0.050505	0.020066	6.334937	1	0.011838	1.051802
Relief amplitude (X3)	−0.017617	0.005207	11.447286	1	0.000716	0.982537
Distance to rivers (X4)	−0.000016	0.000007	5.670057	1	0.017257	0.999984
Constant	1.987054	0.140941	198.766831	1	0.000000	7.294017

¹ “β” is the regression coefficient. ² “S.E.” is the standard error. ³ “Wald” is the value used to test whether the independent variable has an impact on the dependent variable. ⁴ “df” is the degree of freedom. ⁵ “Sig.” is the significance level. ⁶ “Exp(β)” is the odds ratio.

, the variables of the siting preference analysis model for postal buildings demonstrate hierarchically ordered significance levels: Sig.(X₁) < Sig.(X₃) < Sig.(X₂) < Sig.(X₄). This sequential pattern indicates elevation exerted the strongest constraining effect on location decisions, followed by relief amplitude and slope, with rivers exhibiting the weakest influencing power.

Substituting the β values into Equation (4), the siting preference analysis model for Fujian’s postal buildings was obtained as follows:

$$P={\displaystyle \frac{\exp \left(1.987054-0.004444X_1+0.050505X_2-0.017617X_3-0.000016X_4\right)}{1+\exp \left(1.987054-0.004444X_1+0.050505X_2-0.017617X_3-0.000016X_4\right)}}$$

(5)

Finally, the raster data layers of each independent variable were substituted into Equation (5), and the mathematical calculation of the layers was performed using the “Raster Calculator tool” in the ArcGIS 10.3.1 platform, thereby generating a visualized graphic of the siting preference for Fujian’s postal buildings (Figure 7). In this study, the preference degree was categorized into five levels: low (0–0.25), relatively low (0.25–0.5), moderate (0.5–0.6), relatively high (0.6–0.7), and high (>0.7). Areas with relatively high and high preference indicate priority siting locations for postal buildings, that is, the areas that were more suitable for construction.

In order to verify the validity of the method, the siting preference degree values for all the postal buildings were extracted, as shown in

Table 3. The numerical distribution of the siting preference degree of Fujian’s postal buildings.

Degree	Number	Proportion
Low preference	35	4.64%
Relatively low preference	118	15.63%
Moderate preference	75	9.93%
Relatively high preference	203	26.89%
High preference	324	42.91%

. About 70% of the postal buildings in Fujian were located in areas with the high and relatively high site preference degree, proving that the site preference analysis model for postal buildings generated by the BLR method had high validity, and the results could be used for further discussion.

4. Discussion

4.1. Siting Principles and Underlying Reasons

Under the influence of terrain and river factors, the postal buildings’ site selection in Fujian demonstrates strong spatial regularity, exhibiting a preference for coastal zones (Figure 8a) and riverine corridors (Figure 8b), with consistent proximity to administrative centers and military fortresses. Analysis of geographical and historical contexts of the Ming Dynasty provides insights into the underlying drivers of these spatial patterns.

From the geographical perspective, the flat alluvial plains and terraces along the coast and rivers offered superior conditions for constructing postal buildings. Coastal zones and river corridors are areas under fluvial erosion and accumulation that form floodplains, proluvial fans, and alluvial fans in mountainous areas while developing delta belts in plains. These regions generally feature low elevation and flat terrain, with superior geographical conditions. On one hand, such regions facilitate construction and transportation, boosting transmission efficiency. Particularly for mountainous regions like Fujian, the flat zones along coasts and rivers became optimal locations for postal buildings. On the other hand, these areas typically offer livable environments and agricultural suitability, meeting both production and living needs. Humans favored settling here, leading to dense populations, economic prosperity, and well-developed societies that evolved into regional centers. Where towns and people exist, transportation demands arise naturally, thus prompting the establishment of postal buildings. This explains why high siting preference areas of the postal buildings coincide with the distribution of administrative centers.

From the perspective of transportation, coastal zones and river corridors offered excellent shipping conditions that enhanced interregional connectivity. During the Ming Dynasty, underdeveloped technology necessitated reliance on ships, horses, vehicles, and other tools for transmission, with cost-effectiveness and efficiency prioritized in the process. Compared to land transportation, waterway transport significantly reduced labor and material expenditures, thereby creating optimal conditions for postal building construction. Concurrently, burgeoning maritime trade facilitated commodity imports through coastal ports, with goods subsequently distributed inland via postal networks. Where rivers crossed, there were frequent commercial exchanges and busy logistics, and merchants utilized postal houses for accommodations. These combined factors drove the concentrated construction of postal buildings.

The synergistic interaction between natural environmental factors (terrain, rivers) and economic dynamics (transmission efficiency, trade demands) exhibited trans-regional universality, providing a reference for understanding the spatial patterns of postal networks in other river-intensive, commercially developed areas, such as the Yangtze River Basin and the Pearl River Delta.

Furthermore, in coastal zones with high siting preference of postal buildings, a dense concentration of military defense settlements is observed. This spatial correlation indicates that the postal system’s distribution was fundamentally rooted in state governance and military defense needs, symbolizing the spatial projection of centralized imperial authority. As the primary communication network between central and regional administrations, the postal system, often termed the “national bloodline”, held strategic importance for imperial governance. The early Ming Dynasty faced persistent security challenges: northern nomadic tribes frequently raided the central plains, while Japanese pirates repeatedly invaded the southeastern coast [40]. In response, Emperor Hongwu (Zhu Yuanzhang) implemented sweeping military reforms, including the establishment of a nationwide postal network. This system enabled rapid transmission of military intelligence, supplies, and documents between the capital and provincial commands [41]. Particularly in Fujian’s coastal zones, where river estuaries were potential invasion routes for pirates, the postal system assumed heightened strategic significance. During military emergencies, small-scale conflicts could be managed by local garrisons, while large invasions necessitated immediate communication via the postal network, with interior prefectures for long-range military support. The system’s efficiency had crucial impact on the war outcome. Given resource limitations in ancient times, the authorities prioritized postal building construction in coastal zones and administrative centers. This layout ensured timely military communication, allowing central authorities to monitor frontier situations and implement coordinated responses and providing guarantees for warfare victories. This paradigm of “military defense dictating core-zone spatial planning” also appeared in northern frontier towns (e.g., Xuanfu, Datong), though with distinct operations: northern postal networks relied predominantly on land roads, contrasting sharply with the waterway-dominated pattern in the south. The differentiation reflects how environmental factors and military demands interacted to shape regional postal patterns.

Notably, the postal system was not entirely a response to official directives; economic activities also exerted substantial influence, driving dynamic adaptations across historical periods. In the early Ming, when tributary trade dominated, postal buildings primarily served official envoys and tribute transportation. During the Zhengde and Jiajing reigns, however, the rise in private maritime merchants spawned a complex smuggling network relying on postal roads and waterways [2]. To enhance surveillance, the imperial administration established additional Xunjiansi (coastal inspection stations) in commercial hubs and their hinterlands, which objectively led to the increase in related facilities. Furthermore, during peak periods of pirate raids, coastal postal buildings experienced destruction or forced relocation due to wars [42]. The functional transformation and micro-scale spatial reconfiguration of the postal system revealed the impact of socioeconomic development and warfare on the national infrastructure system, which constituted a unique feature distinguishing Fujian’s postal system from inland systems. It also reminds us that static analysis in a single period can hardly capture the system’s full complexity. Environmental preferences formed a critical foundation for the siting of postal buildings, while political–military transformations, economic policy realignments, and disasters might impact it to varying degrees.

In summary, the siting patterns of Fujian’s postal buildings emerged from specific environmental and historical conditions while embodying universal operating principles of ancient postal networks. Environmental factors established a constraining framework for spatial organization, within which political, economic, and social forces modulated the postal system’s distribution density and functional specialization. This analytical perspective of multi-factor interaction provides a more comprehensive approach for understanding the spatial evolution of China’s ancient transportation systems.

4.2. Suggestions for Postal Heritage Conservation

The Ming postal system has existed for over 600 years. Over time, as generations changed, postal buildings were quickly replaced by newer facilities, gradually being eliminated by the times. In the urbanization process of rapid development, most of the postal buildings have been destroyed by the massive construction works due to their small scale, low grade, and single function. Nowadays, only a few postal buildings are left, basically in poor conditions, and for quite a number of them, it is difficult to even find the ruins, which is regrettable. Through comprehensive analysis of the ancient postal system, we aim to enhance public awareness regarding the preservation of such cultural heritage while providing a theoretical reference for postal heritage conservation.

Current conservation strategies for postal heritage predominantly focus on structural restoration and static display. Conventionally, emergency interventions are implemented for damaged sites, prioritizing the restoration of historical features and structural stabilization. Meanwhile, fenced site protection physically isolates heritage from their surroundings to prevent human interference and natural erosion. However, these approaches have obvious limitations: excessive emphasis on heritage itself neglects its contextual environmental connections, thus hindering a systematic understanding of their historical–geographical character and cultural significance.

Building upon the spatial distribution patterns of postal buildings revealed by the BLR model, specifically, a significant preference for coastal zones and riverine corridors as well as a highly coordinated distribution with administrative centers and military fortresses, the following specific conservation and management suggestions for the postal heritage are proposed. We aim not only to preserve heritage structures but, more critically, to sustain and revitalize their intrinsic spatial logic and environmental connectivity.

For riparian postal heritage, conduct detailed surveying and mapping of the sites and their surrounding environment, with particular attention to their proximity relationships with rivers. Accordingly, define the sites and their adjacent landscapes that manifest their waterfront-siting attributes (e.g., ancient docks, ferry remnants) as primary protection zones, where non-conservation construction is strictly prohibited. Formulate renovation plans for rivers and their riparian environments to restore the waterborne transport networks on which the postal system relied. Renovate and reinforce ancient bridges, dams, and other auxiliary facilities along rivers while installing historical interpretation signage, thus shaping rivers into vibrant corridors that integrate ecological and cultural values.

For postal heritage proximate to administrative centers, incorporate key nodes that exhibit spatial–functional dependencies with postal buildings (including prefectural/county seats, historical roads) into the integrated conservation scope, implementing regulated development intensity to preserve spacial structure and landscape compatibility. Designate integrated conservation zones incorporating remnants such as city walls and yamen (government office in feudal China) complexes. Reconstruct historical pathways connecting prefectural/county seats to postal buildings and employ landscape markers to interpret the route of document transmission and material logistics.

Regarding postal heritage adjacent to military fortresses, integrate them within the military heritage conservation framework. Sort out spatial interdependencies between postal buildings and other military fortresses as well as garrison facilities (including Wei/Suo fortress, beacon towers, etc.), delineating integrated conservation zones. Priority reinforcement and protection measures should target remains directly linked to military operations, such as postal roads, signal towers, and defensive walls, prohibiting any development that disrupts defensive spatial pattern. Simulate military postal communication scenarios (e.g., setting up postal route markers and signal simulation devices) to manifest their strategic value.

The above conservation strategies apply to postal heritage where both the remains and associated contextual features remain well-preserved, allowing identification of historical environmental characteristics. For postal heritage where environmental transformations impede holistic conservation, efforts can focus on protecting the sites. On this foundation, utilize multiple exhibition techniques to reconstruct the spatial scene of the postal system and systematically elucidate their functional dependencies on the surrounding environment.

Beyond enhancing spatial connectivity conservation, a corresponding management system must be established. Institute a collaborative management system to regularly monitor and evaluate heritage environment. Implement hierarchical and classified management protocols, enforcing strict controls in core protection zones while allowing moderate cultural–tourism integration projects in secondary zones, provided such developments actively contribute to interpreting the values of postal heritage. Establish dedicated conservation funds through diversified financing, including government allocations, social crowdfunding, and revenue reinvestment, to ensure sustainable implementation of protective measures.

Collectively, this multidimensional strategy constructs a “heritage–environment–history” symbiotic system centered on postal heritage, interconnected by rivers, and supported by historical nodes, such as prefectural cities and military fortress. This initiative not only preserves the sites but, more significantly, through spatial planning and management, systematically highlights the historical and geographical characteristics of ancient postal system, thereby promoting holistic and sustainable conservation.

4.3. Applicability and Limitations

The integrated approach of GIS spatial analysis and logistic regression modeling quantitatively extracted data on post buildings and related environmental factors, constructed a multivariate statistical model, and evaluated distribution patterns of postal building within the study area. The method enabled precise identification of high-preference siting areas and their spatial attributes, objectively revealing location principles. Moreover, logistic regression analysis quantified the significant relationships between environmental factors and postal building distribution patterns. Notably, these factors were widely recognized by archaeological researchers as critical natural parameters in ancient settlement selection, validating the methodological universality for complex ancient system analysis. Theoretically, this analytical framework extends beyond the postal system to other ancient infrastructures involving complex environmental variables (e.g., transportation networks, hydraulic works, military settlements). Its quantitative analysis and multi-factor integration paradigm empower researchers to accurately pinpoint key heritage zones, clarify core distribution determinants, and formulate targeted conservation strategies and planning protocols, thus advancing methodological innovations in ancient systemic heritage conservation.

From the practical perspective, the Ming postal networks maximized operational efficacy through strategic spatial planning, effectively balancing infrastructural functionality with political imperatives under resource constraints. This historical experience informs modern heritage conservation and regional planning: it is of significance to comprehensively integrate environmental conditions with functional imperatives in contemporary planning practices, optimizing resource configuration for synergistic benefits of multiple objectives that simultaneously preserve historical continuity and advance sustainable regional development.

However, the limitation of this study still exists, that is, the analyzed influencing factors remained incomplete. In addition to the terrain and water system factors, the siting of postal buildings may also be related to anthropogenic factors, which are difficult to be qualified, so they were not counted as influenced factors in this study. Overall, the identified spatial preferences represent predominant environmental tendencies, while exceptional cases require specific analysis with dialectical thoughts and proper methods.

5. Conclusions

Taking Fujian in the Ming Dynasty as the research object, this study investigated the siting principles of postal buildings through GIS spatial analysis and BLR modeling, drawing the following conclusions:

• The BLR model showed good applicability for studying the siting principles of ancient postal buildings, having passed both the Omnibus Test (p < 0.05) and Hosmer–Lemeshow Test (p > 0.05).g
• The distribution of the Ming Dynasty postal system exhibited spatial heterogeneity, with terrain (elevation, slope, relief amplitude) and river factors exerting statistically significant influences on its siting preference. Among these independent variables, elevation demonstrated the strongest explanatory power regarding siting rules, followed in descending order by slope, relief amplitude, and distance to rivers.
• Under the comprehensive influence of elevation, slope, relief amplitude, and rivers, areas with high siting probability of Fujian’s postal buildings showed strong regularity, exhibiting a preference for coastal zones and riverine corridors, with consistent proximity to administrative centers and military fortresses.

In general, the data processing and analysis functions of GIS not only visually present the spatial distribution of ancient postal system but also provide reliable tools for acquiring spatial characteristics. Logistic regression statistically interprets GIS-derived data, enhancing the accuracy of analytical results. The integrated approach of GIS and BLR demonstrates good applicability for investigating spatial distribution patterns in ancient settlement systems while offering insights for contemporary conservation practice and sustainable development. However, limitations must be acknowledged: the analysis incorporated only quantifiable environmental variables, excluding non-quantifiable micro-scale factors. Thus, the findings reveal general siting principles shaped by terrain and river systems, with exceptional cases requiring dialectical analysis within the historical context. In future work, we will continue to explore more scientific methods to deepen the ancient postal system research.