Due to the\u00a0structure of\u00a0the\u00a0public procurement and the\u00a0required schedule, there was no scope within\u00a0the\u00a0project for conducting hydrological measurements in\u00a0the\u00a0field. The\u00a0approach was therefore based on a\u00a0synthesis of\u00a0available expert knowledge, analysis of\u00a0spatial data, and categorisation of\u00a0roads according to their morphology, drainage type, interaction with the\u00a0watercourse network, and position relative to the\u00a0landscape. Based on the\u00a0literature review, key types of\u00a0interactions between the\u00a0road network and runoff processes were identified, and the\u00a0design of\u00a0the\u00a0road classification framework was developed at two levels: the\u00a0basic level relied solely on data analysis, while the\u00a0detailed typology refined the\u00a0basic level using insights from field mapping, but not from measurements of\u00a0the\u00a0actual hydrological interaction between the\u00a0road and its surroundings. These methodological limitations and the\u00a0chosen framework are further elaborated in\u00a0the\u00a0following sections.<\/p>\n
THEORETICAL PRINCIPLES AND FOUNDATIONS<\/h2>\n
The natural hydrological regime of mountain areas arises from the interaction of topography, soil properties, vegetation cover, and atmospheric inputs. In its natural state, precipitation water partly infiltrates into the soil and flows subsurface, while the remainder runs off on the surface, especially during high-intensity rainfall and saturated soil profile. This system, however, can be significantly\u00a0disrupted by linear structures, among which roads play a\u00a0particularly prominent role. Roads affect both surface and subsurface runoff, with the\u00a0consequences of\u00a0these impacts varying according to the\u00a0road\u2019s\u00a0position in\u00a0the\u00a0landscape, its morphology, surface modifications, and drainage measures.<\/p>\n
Intervention in\u00a0the\u00a0natural terrain\u00a0morphology disrupts the\u00a0continuity of\u00a0surface water flow. Roads often follow the\u00a0fall line of\u00a0a\u00a0slope or are situated in\u00a0slope depressions, thereby creating preferential runoff pathways. Water running off from the\u00a0surrounding terrain\u00a0accumulates on the\u00a0surface of\u00a0the\u00a0compacted or paved road and is conveyed along the\u00a0road alignment, or alternatively in\u00a0ditches or wheel ruts. Such concentrated runoff is then directed either into the\u00a0nearest watercourse or towards the\u00a0slope edge, where it may trigger erosion processes and destabilise the\u00a0soil profile. This phenomenon is often referred to as \u201cthe\u00a0function of\u00a0a\u00a0road as a\u00a0surface water collector.\u201d In\u00a0other situations, however, roads may act as distributors \u2013 that is, water from the\u00a0road or ditch is dispersed into the\u00a0surrounding environment, for example by seeping into the\u00a0slope or through transverse drainage features. At points where ditches, culverts, or erosion gullies connect directly to the\u00a0hydrographic network, roads function as inflow points that link surface runoff directly to recipients, thereby significantly accelerating the\u00a0catchment response.<\/p>\n
Roads also significantly disrupt subsurface water flow. Due to their construction and use, the\u00a0subgrade beneath higher-category roads is often heavily compacted, which reduces the\u00a0soil infiltration capacity and redirects water into the\u00a0surface system. In\u00a0addition, natural conductive horizons are interrupted, which would normally allow lateral (slope) water flow in\u00a0shallow soil layers.<\/p>\n
Particularly problematic are situations where a\u00a0road is cut into a\u00a0slope and runs across it, almost or entirely along the\u00a0contour lines. In\u00a0such cases, the\u00a0slope toe or the\u00a0side of\u00a0the\u00a0road cut disrupts the\u00a0natural shallow drainage layers through which subsurface flow occurs. Water then emerges at the\u00a0surface from the\u00a0disturbed slope, resulting in\u00a0the\u00a0conversion of\u00a0subsurface runoff into surface runoff. The\u00a0outcome is not only the\u00a0loss of\u00a0the\u00a0slope\u2019s\u00a0infiltration function but also an increased risk of\u00a0erosion and accelerated drainage. This mechanism may lead to the\u00a0formation of\u00a0secondary spring outflows or even to the\u00a0development of\u00a0small watercourses along road embankments, although under natural conditions no surface runoff would occur at all. In\u00a0some cases, these effects combine \u2013 for example, when water accumulates upslope of\u00a0a\u00a0road due to a\u00a0barrier, increasing profile saturation and subsequently emerging at the\u00a0surface as a\u00a0secondary spring outflow, thereby increasing the\u00a0amount of\u00a0surface runoff.<\/p>\n
The\u00a0mechanisms described are supported by numerous studies demonstrating changes in\u00a0the\u00a0hydrological regime caused by linear structures, a\u00a0brief selection of\u00a0which is presented in\u00a0the\u00a0introduction to this article. On the\u00a0basis of\u00a0these hydrological concepts, a\u00a0classification framework was designed that takes into account the\u00a0mode of\u00a0interaction between roads and both surface and subsurface runoff. This framework serves as the\u00a0foundation for the\u00a0road network typology, which is described in\u00a0detail in\u00a0the\u00a0following sections.<\/p>\n
ROAD NETWORK TYPOLOGY<\/h2>\n
The\u00a0road network typology was defined in\u00a0terms of\u00a0its potential impact on the\u00a0hydrological regime, considering both surface and subsurface runoff. It was designed and tested on the\u00a0road network in\u00a0the\u00a0mountainous environment of\u00a0KRNAP but formulated in\u00a0a\u00a0general way so that it could be applied more or less anywhere in\u00a0the\u00a0Czech Republic, primarily in\u00a0protected areas. The\u00a0typology was developed based on a\u00a0combination of\u00a0digital spatial analyses, field observations, and the\u00a0hydrological principles described in\u00a0the\u00a0previous section.<\/p>\n
The\u00a0types of\u00a0roads evaluated included all roads with the\u00a0potential to influence the\u00a0direction and volume of\u00a0precipitation runoff: public roads as defined by the\u00a0Road Act, including the\u00a0network of\u00a0local and purpose-built roads; roads of\u00a0categories 1L\u20134L according to \u010cSN 73 6108 \u2013 Forest Road Network\u00a0[9]; as well as significant hiking trails and other unregistered but mapped paved roads leading to buildings or intersecting water conveyance structures.<\/p>\n
The\u00a0proposed typology is applied not to roads as continuous entities, but to their homogeneous sections, for example those with a\u00a0single type of\u00a0construction or pavement, or with specific slope characteristics. The\u00a0primary basis for classification is the\u00a0road\u2019s\u00a0function in\u00a0terms of\u00a0its ability to interrupt shallow subsurface and surface runoff, followed by its accumulative or conveyance function, that is, its capacity to retain\u00a0runoff or, conversely, to discharge it rapidly into the\u00a0hydrographic network. Point features for conveying surface water, such as small bridges and culverts, are included in\u00a0the\u00a0typology only to a\u00a0limited extent due to their low level of\u00a0mapping and documentation. The\u00a0detailed procedure for dividing the\u00a0road network into homogeneous sections is presented later in\u00a0this text, in\u00a0the\u00a0Road Network Map section.<\/p>\n
Factors for typology classification<\/h3>\n
The\u00a0ability of\u00a0a\u00a0road section to influence runoff conditions depends on a\u00a0range of\u00a0factors, which carry different weights in\u00a0various combinations.<\/p>\n
The\u00a0proposed classification uses the\u00a0following main\u00a0factors:<\/p>\n
\n
- \n
- Affected component of\u00a0runoff,
\n
\n<\/li>\n - Potential runoff volume (catchment area),
\n
\n<\/li>\n - Potential to influence runoff velocity.<\/li>\n<\/ol>\n
The\u00a0first factor enters the\u00a0classification directly as a\u00a0categorical variable with two classes. To represent potential runoff volume, a\u00a0universal runoff characteristic in\u00a0the\u00a0form of\u00a0catchment area was chosen, in\u00a0an effort to avoid the\u00a0considerable uncertainties associated with methods for quantifying runoff from mountain\u00a0and forested areas in\u00a0unmonitored catchments. The\u00a0final main\u00a0factor, the\u00a0potential to influence runoff velocity, in\u00a0practice depends on a\u00a0range of\u00a0detailed characteristics; for the\u00a0proposed classification, the\u00a0following were selected:<\/p>\n
- \n
- presence of\u00a0longitudinal drainage features,<\/li>\n
- design of\u00a0road drainage and the\u00a0occurrence and technical design of\u00a0cross\u00a0drains,<\/li>\n
- construction of\u00a0the\u00a0road embankment with regard to permeability,<\/li>\n
- construction (pavement) of\u00a0the\u00a0road surface with regard to permeability,<\/li>\n
- orientation of\u00a0the\u00a0road relative to the\u00a0slope,<\/li>\n
- arrangement of\u00a0the\u00a0road\u2019s\u00a0cross-section relative to the\u00a0landscape,<\/li>\n
- longitudinal and transverse slope,<\/li>\n
- tructures on the\u00a0road and their water conveyance methods,<\/li>\n
- alignment or crossing with a\u00a0watercourse.<\/li>\n<\/ul>\n
Typology levels<\/h3>\n
The\u00a0road typology was designed in\u00a0a\u00a0structured form with two levels of\u00a0detail: basic and detailed. The\u00a0basic level allows for the\u00a0classification of\u00a0road network sections solely based on analyses of\u00a0commonly available data, while the\u00a0detailed level provides further refinement of\u00a0the\u00a0basic typology through field surveys and sophisticated data analyses. This two-tiered structure is necessary given the\u00a0extent of\u00a0the\u00a0road network in\u00a0the\u00a0study area, which, under normal time and personnel constraints, does not permit complete physical mapping.<\/p>\n
Basic typology level<\/h4>\n
This level of\u00a0road network classification was designed for application solely on the\u00a0basis of\u00a0easily accessible data within\u00a0a\u00a0GIS environment. It was tested at the\u00a0spatial scale of\u00a0the\u00a0entire KRNAP, with a\u00a0view to its potential use in\u00a0any other area within\u00a0Czechia. The\u00a0main\u00a0data source is the\u00a0topological network of\u00a0linear road objects as recorded in\u00a0the\u00a0ZABAGED database (version 2021). For testing the\u00a0typology, a\u00a0specialized road network dataset maintained by the\u00a0KRNAP GIS department was used. To determine the\u00a0average slope characteristics of\u00a0road sections, the\u00a0DMR4G elevation model was applied, and for identifying extremes (peaks and local minima) in\u00a0longitudinal profiles, the\u00a0DMR5G model was used.<\/p>\n
In\u00a0the\u00a0first phase of\u00a0the\u00a0study, general criteria were formulated for classifying road sections into classes of\u00a0the\u00a0basic typology level, separately for surface runoff (SR) and subsurface runoff (SSR). These are summarised in\u00a0Tab.\u00a01<\/em>.<\/p>\n
Tab.\u00a01. Classes of\u00a0the\u00a0basic level of\u00a0road typology and general combination of\u00a0key characteristics for classification of\u00a0individual segments<\/h5>\n
![\"\"](\"data:image\/svg+xml;base64,PHN2ZyB3aWR0aD0iMSIgaGVpZ2h0PSIxIiB4bWxucz0iaHR0cDovL3d3dy53My5vcmcvMjAwMC9zdmciPjwvc3ZnPg==\")
<\/a>\nIn\u00a0the\u00a0first proposal of\u00a0the\u00a0basic typology level, the\u00a0existence of\u00a0longitudinal drainage features and the\u00a0terrain\u00a0configuration (the\u00a0position of\u00a0the\u00a0road relative to the\u00a0surrounding landscape) were also considered as key characteristics. Although these characteristics cannot be derived from commonly available data, within\u00a0the\u00a0framework of\u00a0the\u00a0project, procedures for analysing detailed DTMs obtained using airborne LiDAR techniques were tested, which have become increasingly accessible in\u00a0recent years. For the\u00a0analysis of\u00a0pilot sites selected within\u00a0the\u00a0project, a\u00a0DTM from 2012 with a\u00a01\u202fm resolution, provided by KRNAP, and a\u00a0DTM with a\u00a050\u202fcm resolution acquired using UAVs were available. In\u00a0both cases, the\u00a0products were derived from laser scanning. Approximately 20 cross-sections were analysed at two sites; an example from the\u00a0site below \u0160pindlerovka, including the\u00a0locations of\u00a0the\u00a0analysed profiles, is shown in\u00a0Fig.\u00a01<\/span><\/em>.<\/span><\/p>\n
<\/a>\nFig. 1. Orthophoto (left) and detailed DTM (right) derived from UAV data; cross-sections in\u00a0red are used for testing the\u00a0identification of\u00a0roadside drainage features<\/h6>\n
Analyses of\u00a0the\u00a0obtained cross-sections proved to be conditionally useful for obtaining information on the\u00a0terrain\u00a0configuration of\u00a0a\u00a0road section, or on longitudinal drainage structures. Fig.\u00a02<\/span><\/em> shows arguably the\u00a0clearest of\u00a0the\u00a0analysed cross-sections at the\u00a0pilot site below Rennerovky. On the\u00a01\u202fm resolution digital model, road ditches are mostly indistinguishable, whereas at higher resolution they are generally identifiable. However, it is often difficult to identify them amidst the\u00a0noise caused by remnants of\u00a0elevation data processing. The\u00a0terrain\u00a0configuration of\u00a0the\u00a0road was generally satisfactorily visible even on the\u00a0less detailed 1\u202fm model, whereas the\u00a0standard 2\u202fm DMR5G model published by the\u00a0Czech Office for Surveying, Mapping and Cadastre (\u010c\u00daZK) was insufficient for this purpose.<\/span><\/p>\n
<\/a>\nFig. 2. Cross-section of\u00a0a\u00a0road (right) derived from DTMs based on LiDAR scanning with 1\u00a0m resolution (red, KRNAP Administration) and 50\u00a0cm resolution (blue, CTU) for\u00a0testing the\u00a0identification of\u00a0road terrain alignment\u00a0and roadside drainage features; photograph of\u00a0the\u00a0actual condition from a\u00a0field survey (left)<\/h6>\n
In general, a significant problem was encountered with georeferencing aerial data in forested areas and with filtering the captured point clouds. The analysis was further complicated by the positional accuracy of the road network lines, which was already insufficient for this purpose, even in the case of the corrected map dataset from KRNAP, which generally exhibited higher accuracy than the ZABAGED positional data. Automating the process of generating cross-sections and identifying longitudinal drainage features from detailed DTMs proved to be unrealistic, and manual analyses were inefficient compared with a simple field survey. For these reasons, the characteristics of terrain configuration and\u00a0<\/span>the\u00a0presence of\u00a0longitudinal drainage features were not used for the\u00a0basic typology level of\u00a0the\u00a0road network and were instead assigned to the\u00a0detailed level.<\/span><\/p>\n
The\u00a0following six characteristics were used for the\u00a0final classification of\u00a0road sections according to the\u00a0basic typology level:<\/span><\/p>\n
Z\u00a0\u2013 Paved\/unpaved road (1\/0),<\/span><\/p>\n
M \u2013 Connection to a\u00a0local minimum or crossing with a\u00a0watercourse
\nYES\/NO\u00a0(1\/0),<\/span><\/p>\nS\u00a0\u2013 Average longitudinal slope of\u00a0the\u00a0section (%),<\/span><\/p>\n
D \u2013 Deviation of\u00a0the\u00a0road from the\u00a0fall line (0\u00b0\u201390\u00b0),<\/span><\/p>\n
T \u2013 Transverse slope of\u00a0the\u00a0terrain\u00a0around the\u00a0section (%),<\/span><\/p>\n
W \u2013 Contributing sub-catchment area (ha).<\/span><\/p>\n
For the\u00a0application of\u00a0the\u00a0basic typology level in\u00a0the\u00a0creation of\u00a0the\u00a0KRNAP Road Network Map, the\u00a0classification key from Tab.\u00a02<\/span><\/em> was used. The\u00a0threshold values of\u00a0the\u00a0criteria and their combinations for classifying road sections were based on a\u00a0frequency analysis of\u00a0the\u00a0occurrence of\u00a0individual parameter values across KRNAP and may therefore be influenced by the\u00a0specific characteristics of\u00a0this area. When applying the\u00a0typology in\u00a0other regions, it is advisable to carry out a\u00a0similar frequency analysis and, if necessary, adjust the\u00a0threshold values for class boundaries. However, any modification of\u00a0the\u00a0combinations of\u00a0criteria used should be based on a\u00a0significant objective reason, such as the\u00a0absence of\u00a0certain\u00a0data (for example, information on road surface characteristics may be unavailable). The\u00a0basic typology level of\u00a0the\u00a0road network is designed as an explicit combination of\u00a0the\u00a0potential to influence both components of\u00a0direct runoff, for instance B\/C. By incorporating factors identified through detailed field surveys, it can be extended to the\u00a0detailed level. The\u00a0application of\u00a0the\u00a0typology and the\u00a0presentation of\u00a0results are addressed in\u00a0the\u00a0Road Network Map section.<\/span><\/p>\n
Tab.\u00a02. Combinations and values of\u00a0characteristics for classifying road segments according to the\u00a0basic level of\u00a0typology<\/h5>\n
<\/a>\nDetailed typology level<\/span><\/h4>\n
Building on the\u00a0basic typology level of\u00a0the\u00a0road network, which reflects the\u00a0potential to influence runoff characteristics in\u00a0the\u00a0affected area, the\u00a0detailed level provides a\u00a0more in-depth analysis of\u00a0the\u00a0impact on runoff conditions in\u00a0individual cases. It supplements the\u00a0basic typology level with road network characteristics that, given the\u00a0current state of\u00a0available data, can only be determined through field surveys.<\/span><\/p>\n
For the\u00a0detailed survey in\u00a0KRNAP, five pilot sites were designated. The\u00a0survey was conducted from July to November 2022 and was slightly supplemented in\u00a02023. Data were collected using the\u00a0open mobile application QField based on QGIS technology, which allows simultaneous data collection by multiple personnel and subsequent synchronization. After two initial adjustments, the\u00a0data model for collection was finalised as three separate point layers according to Tab.\u00a03<\/span><\/em>, with photographic annotations. The\u00a0categories of\u00a0individual characteristics are not provided here; following the\u00a0formal completion of\u00a0the\u00a0above-mentioned project (no. TITSMZP945), they will be accessible in\u00a0full in\u00a0the\u00a0results report V1 \u2013 Road Network Typology.<\/span><\/p>\n
Tab.\u00a03. Point data layers for field data collection and recorded characteristics<\/h5>\n
<\/a>\nDuring the\u00a0development of\u00a0the\u00a0basic typology level, or its application to the\u00a0map, road network sections and nodal points were defined. For the\u00a0purposes of\u00a0applying the\u00a0detailed typology, this division is supplemented by significant specific points of\u00a0type C (Tab.\u00a03<\/span><\/em>) identified during the\u00a0field survey (e.g., the\u00a0end of\u00a0a\u00a0longitudinal drainage feature, a\u00a0change in\u00a0surface type, etc.). The\u00a0detailed characteristics recorded at the\u00a0corresponding type B points are then assigned to the\u00a0resulting road sections. From the\u00a0recorded characteristics, four were selected and adopted as the\u00a0criteria for the\u00a0detailed typology level according to Tab.\u00a04<\/span><\/em>.<\/span><\/p>\n
Tab 4. Criteria of\u00a0the\u00a0detailed level of\u00a0road typology<\/h5>\n
<\/a>\nIt is evident that the\u00a0individual criteria are interrelated to varying degrees, and therefore the\u00a0final assessment of\u00a0their impact on surface or subsurface runoff must necessarily be based on all criteria simultaneously, which requires a\u00a0certain\u00a0level of\u00a0expertise in\u00a0runoff processes and hydrology in\u00a0general. The\u00a0effects on individual runoff components, as well as guidance for evaluating roads according to these criteria, are provided in\u00a0the\u00a0forthcoming project results prepared for publication: V1 \u2013 Road Network Typology and V3 \u2013 Methodology for Recommendations on the\u00a0Construction of\u00a0New and Modification of\u00a0Existing Roads with Regard to Minimising Surface Runoff<\/span> (hereinafter referred to as the\u00a0Methodology). These criteria were applied and graphically represented in\u00a0result V2 \u2013 KRNAP Road Network Map, the\u00a0derivation of\u00a0which is described in\u00a0the\u00a0following section of\u00a0this text.<\/span><\/p>\n
ROAD NETWORK MAP<\/h2>\n
As the\u00a0second required output of\u00a0the\u00a0project mentioned in\u00a0the\u00a0introduction, the\u00a0KRNAP Road Network Map (hereinafter referred to as the\u00a0Map) was created, applying the\u00a0road network typology to assess its influence on the\u00a0hydrological regime of\u00a0the\u00a0area. It was published in\u00a0the\u00a0form of\u00a0three cartographic atlases. The\u00a0first covers the\u00a0entire KRNAP area, applying the\u00a0basic typology level of\u00a0the\u00a0road network, while the\u00a0remaining two expand the\u00a0basic level with detailed typology criteria for the\u00a0five pilot sites, allowing a\u00a0more detailed assessment of\u00a0the\u00a0potential to influence subsurface and surface runoff. The\u00a0Map is intended primarily as one of\u00a0the\u00a0key resources for selecting locations suitable for implementing restoration measures and actions to reduce the\u00a0negative impacts of\u00a0the\u00a0road network on the\u00a0runoff regime within\u00a0the\u00a0National Park. In\u00a0addition, together with the\u00a0Map\u2019s\u00a0accompanying documentation and Methodology, it is intended to serve future authors of\u00a0similar studies in\u00a0other Czech protected areas as a\u00a0methodological guide and example for applying the\u00a0Road Network Typology and assessing road network hydrological impacts. A\u00a0brief description of\u00a0the\u00a0Map creation methodology follows; the\u00a0full version will be provided in\u00a0the\u00a0accompanying documentation, which is scheduled for publication soon together with the\u00a0Map.<\/span><\/p>\n
Input data<\/h3>\n
The\u00a0primary basis for creating the\u00a0Map was the\u00a0linear layer of\u00a0the\u00a0road network provided by KRNAP, which was preferred over ZABAGED data due to its higher positional accuracy and more extensive attribute set. Since not all protected areas in\u00a0Czechia, where the\u00a0developed methodology was intended to be applied, have access to a\u00a0similarly detailed dataset, the\u00a0map creation was also successfully tested on the\u00a0positional layers from ZABAGED, specifically by combining the\u00a0following layers:<\/span><\/p>\n
- \n
- Roads, motorways,<\/li>\n
- Unregistered roads,<\/li>\n
- Streets,<\/li>\n
- Paths.<\/li>\n<\/ul>\n
The\u00a0last-mentioned category, Paths<\/span>, should ideally be distinguished in\u00a0the\u00a0final linear layer according to available detailed attributes. Since 2024, the\u00a0original classification of\u00a0Paths<\/span> as paved\/unpaved<\/span> has been replaced by a\u00a0new division into Maintained\/unmaintained paths<\/span>. Within\u00a0the\u00a0project, the\u00a0nature and impacts of\u00a0this change could not be analysed or assessed in\u00a0detail. The\u00a0above-mentioned linear layers should ideally be supplemented with data on the\u00a0classification of\u00a0the\u00a0forest road network, which should be available from its administrator.<\/span><\/p>\n
Digital terrain\u00a0models DMR4G and DMR5G, as well as the\u00a0Watercourse<\/span> layer from the\u00a0ZABAGED positional data, were also used in\u00a0creating the\u00a0map. For the\u00a0application of\u00a0the\u00a0detailed typology level, the\u00a0point layers of\u00a0characteristic and specific points from the\u00a0field survey described in\u00a0the\u00a0previous section were employed.<\/span><\/p>\n
Methodology for creating the map at the\u00a0basic typology level<\/h3>\n
Before assigning characteristic values for classifying roads according to the\u00a0proposed typology, it was first necessary to divide the\u00a0linear road elements into segments that were homogeneous in\u00a0terms of\u00a0alignment and elevation, road surface, and approximate length. This segmentation was carried out in\u00a0several steps, which are briefly summarised here; the\u00a0full procedure is provided in\u00a0the\u00a0accompanying Map documentation.<\/span><\/p>\n
The\u00a0first essential step was to correct the\u00a0topology of\u00a0the\u00a0lines so that they were broken at the\u00a0road network nodes. Overlaps and incomplete lines are undesirable. Pseudo nodes \u2013 i.e., the\u00a0junction of\u00a0two linear elements \u2013 are allowed only at locations where an attribute key to the\u00a0classification of\u00a0the\u00a0segment according to the\u00a0typology changes (in\u00a0this case only a\u00a0change in\u00a0road surface). In\u00a0the\u00a0base layer provided by KRNAP, these rules were broken in\u00a0several hundred instances and had to be semi-automatically removed. When using ZABAGED positional data, the\u00a0topology of\u00a0the\u00a0resulting network must be cleaned according to these rules after merging the\u00a0specified linear layers.<\/span><\/p>\n
Division at locations of\u00a0alignment breaks<\/span><\/h4>\n
Sharp changes in\u00a0alignment and curves are frequent locations for changes in\u00a0several road characteristics \u2013 longitudinal slope, configuration relative to the\u00a0terrain, the\u00a0presence of\u00a0longitudinal drainage features, and others. Since standard GIS tools are unable to identify these points on linear features, custom analytical scripts were developed in\u00a0the\u00a0R project environment for this purpose. In\u00a0the\u00a0first step, sharp alignment breaks were identified where adjacent vertices formed an angle of\u00a0less than 100\u00b0; in\u00a0Fig.\u00a0<\/em>3<\/span>, such a\u00a0break is marked with a\u00a0red triangle. In\u00a0the\u00a0second step, sharp curves \u2013 represented in\u00a0the\u00a0linear data by a\u00a0series of\u00a0very short segments \u2013 were identified. For this purpose, each line was approximated using points at a\u00a0constant spacing of\u00a015\u202fm, and an empirical threshold of\u00a0120\u00b0 was applied to adjacent points; in\u00a0Fig.\u00a03<\/span>, the\u00a0sharp curves are indicated by circles, with colours reflecting their significance. For the\u00a0final division of\u00a0the\u00a0lines, points that were too close to each other or located near road network nodes were filtered out, as these could otherwise have caused undesirable fragmentation of\u00a0the\u00a0road network.<\/span><\/p>\n
<\/a>\nFig. 3. Sharp break point (red triangle) and three prominent curves (orange and yellow circles) as split points of\u00a0the\u00a0access road to Tet\u0159ev\u00ed boudy<\/h6>\n
Crossing watercourses<\/span><\/h4>\n
It is advisable to split road lines at the\u00a0point where they cross a\u00a0watercourse, as such locations often (though not always) involve a\u00a0change in\u00a0elevation. The\u00a0connection of\u00a0a\u00a0segment to a\u00a0watercourse, or to a\u00a0local elevation minimum, is also one of\u00a0the\u00a0characteristics used for the\u00a0classification of\u00a0segments according to the\u00a0proposed typology. However, junction nodes of\u00a0the\u00a0road network (intersections) are often located close to crossings, and in\u00a0such places it is undesirable to split segments, as this would lead to excessive fragmentation. In\u00a0the\u00a0wider area of\u00a0KRNAP, more than 2,700 intersections with surface sections of\u00a0watercourses were identified but, after filtering out unsuitable points, fewer than half of\u00a0them were used.<\/span><\/p>\n
Identification of\u00a0elevation extremes<\/span><\/h4>\n
To ensure correct calculation of\u00a0slope parameters, it is necessary to divide road lines at local elevation extremes. Since standard GIS tools do not allow for the\u00a0identification of\u00a0these locations, custom analytical procedures were again\u00a0developed and tested in\u00a0the\u00a0R project environment.<\/span><\/p>\n
For the\u00a0identification of\u00a0optimal tool settings, points were generated along each road segment in\u00a0four spacing variants (2, 5, 10, and 20 m) and assigned elevation values from DMR5G. At each point, its elevation was assessed in\u00a0the\u00a0context of\u00a0the\u00a0two neighbouring points, and local maxima and minima were indicated according to the\u00a0required elevation difference threshold (five variants ranging from 10 to 150 cm). The\u00a0extremes were then classified into six levels of\u00a0significance based on their height or depth, as well as their width and continuity of\u00a0slope (presence of\u00a0inflection) in\u00a0the\u00a0evaluated surroundings.<\/span><\/p>\n
To verify potential errors in\u00a0the\u00a0alignment of\u00a0roads, the\u00a0same analysis was repeated for parallel lines on both sides of\u00a0the\u00a0roads at offsets of\u00a05 and 10 m. Automatically generated longitudinal profile graphs were systematically subjected to visual inspection in\u00a0groups, one of\u00a0which is shown in\u00a0Fig.\u00a04<\/span><\/em>. This approach confirmed sufficient accuracy of\u00a0the\u00a0alignment of\u00a0road lines compared with their parallel offsets. An offset of\u00a010 m and an elevation threshold of\u00a070 cm proved optimal for constructing longitudinal profiles. Finally, points located close to road network nodes were removed, and the\u00a0occurrence of\u00a0nearby opposite extremes was corrected.<\/span><\/p>\n
<\/a>Fig. 4. Example of\u00a0longitudinal profiles of\u00a0the\u00a0road section axis (left) and two equidistant lines at an offset of\u00a05 and 10\u00a0m (center and right), along with identified local elevation extremes. The\u00a0profiles were derived from points with 10\u00a0m spacing, with a\u00a0vertical threshold of\u00a070\u00a0cm used to indicate significant extremes<\/h6>\nSegmenting by length and determining segment characteristics<\/span><\/h4>\n
After accounting for directional changes, watercourse crossings, and elevation extremes, the\u00a0resulting road segments were divided into 200 m lengths. This completed the\u00a0homogenization of\u00a0road network segments, after which the\u00a0segment characteristics were calculated for the\u00a0application of\u00a0the\u00a0typology. The\u00a0average slope of\u00a0each segment was derived from DMR5G along the\u00a0road lines. From a\u00a0smoothed DMR4G raster, a\u00a0slope raster was generated, and within\u00a0a\u00a020 m buffer zone around the\u00a0segment axes, the\u00a0average slopes of\u00a0the\u00a0surrounding terrain\u00a0were evaluated. A\u00a0more detailed DMR5G is not suitable for this purpose, as it also captures the\u00a0elevation characteristics of\u00a0the\u00a0road itself, such as cuttings or roadside ditches. For the\u00a0same reason, DMR4G was also used to derive the\u00a0contributing areas (micro-catchments) of\u00a0road segments. A\u00a0simplifying assumption of\u00a0complete interruption of\u00a0runoff by the\u00a0road was adopted, since the\u00a0actual capacity of\u00a0a\u00a0road to retain\u00a0runoff cannot be determined without a\u00a0detailed field survey. To derive the\u00a0contributing areas, a\u00a0complex procedure was developed, which involved removing the\u00a0road axes and surface watercourses from the\u00a0DMR and expanding the\u00a0raster representation of\u00a0the\u00a0roads. A\u00a0detailed description of\u00a0this procedure is beyond the\u00a0scope of\u00a0this article.<\/span><\/p>\n
Application of\u00a0the\u00a0basic typology level and map production<\/span><\/h4>\n
Using database processing tools, values of characteristics determining the significance of each road segment for influencing surface and subsurface runoff were assigned to vector road segments according to the developed Road Network Typology. After classifying each segment into combined categories (SR and SSR), the final Map for the entire KRNAP area was produced. Given the extent and level of detail of the information displayed, the Map was organised as an atlas of map sheets at a scale of 1 : 25,000. The potential influence\u00a0<\/span>on SR is indicated by variable line thickness, while the\u00a0potential influence on SSR\u00a0is shown using a\u00a0simple accompanying label. An example map sheet is shown\u00a0in\u00a0Fig.\u00a05<\/span><\/em>.<\/span><\/p>\n
<\/a><\/h6>\nFig. 5. Excerpt of\u00a0the\u00a0road network map with the\u00a0applied basic level of\u00a0typology showing the\u00a0potential impact on runoff components<\/span><\/h6>\n
Methodology for producing the\u00a0detailed typology level\u00a0map<\/span><\/h3>\n
The\u00a0detailed typology level map of\u00a0the\u00a0road network expands the\u00a0basic level map by incorporating insights from field surveys, specifically the\u00a0characteristics listed in\u00a0Tab.\u00a04<\/span><\/em>. These and other road characteristics were collected and compiled into three point layers, including photographic annotations, which can be appropriately displayed using standard GIS tools, although displaying multiple photographic attachments per point in\u00a0ArcGIS Desktop posed a\u00a0considerable challenge. The\u00a0collected points naturally did not match in\u00a0density the\u00a0derived segmentation of\u00a0the\u00a0road network created during the\u00a0production of\u00a0the\u00a0basic level map. Similarly, the\u00a0points were generally not located directly on the\u00a0line of\u00a0the\u00a0road segment, either due to inaccurate GPS positioning in\u00a0forested mountain\u00a0areas, errors in\u00a0point placement, or even the\u00a0absence of\u00a0the\u00a0road line in\u00a0the\u00a0map base. Therefore, a\u00a0series of\u00a0preparatory steps had to be undertaken before applying the\u00a0detailed typology.<\/span><\/p>\n
Harmonization and completion of\u00a0survey data<\/span><\/h4>\n
The\u00a0quality and completeness of\u00a0the\u00a0attributes in\u00a0the\u00a0field survey points is crucial for successful application of\u00a0the\u00a0detailed typology to road segments. Due to differences in\u00a0mapping strategies among individual field workers, for the\u00a0\u201ccross drainage\u201d characteristic, just over 1,000 of\u00a0nearly 2,600 points contained either no information or only the\u00a0category \u201cother\u201d. Missing characteristics had to be completed using the\u00a0collected photo documentation or, where appropriate, validated against neighbouring points. In\u00a0addition to filling in\u00a0missing attributes, the\u00a0consistency of\u00a0characteristics assigned by individual field workers was spot-checked and corrected to remove subjective bias, for example in\u00a0the\u00a0case of\u00a0the\u00a0road\u2019s\u00a0configuration relative to the\u00a0surrounding landscape.<\/span><\/p>\n
Assignment of\u00a0points to road segments<\/span><\/h4>\n
The\u00a0positions of\u00a0points collected during the\u00a0field survey using GPS were subsequently corrected and linked to the\u00a0corresponding road segments. GPS positional deviations of\u00a0the\u00a0points relative to the\u00a0road centreline in\u00a0the\u00a0map dataset ranged from a\u00a0few metres up to several tens of\u00a0metres. Although in\u00a0many cases there was evidence of\u00a0errors in\u00a0the\u00a0map base, in\u00a0order to maintain\u00a0consistency with the\u00a0original data provided by the\u00a0KRNAP Administration, the\u00a0road lines were not corrected; instead, the\u00a0recorded points were automatically moved to the\u00a0nearest position on these lines. Following testing of\u00a0various values, a\u00a0distance of\u00a020\u202fm was chosen as the\u00a0maximum threshold for moving a\u00a0point. Points beyond this threshold had to be assigned manually or completely excluded to avoid, for example, assigning a\u00a0point that characterises a\u00a0road not recorded in\u00a0the\u00a0map base. A\u00a0thorough visual inspection and correction of\u00a0incorrectly assigned points was necessary, particularly in\u00a0the\u00a0areas of\u00a0junctions.<\/span><\/p>\n
Additional segmentation and transfer of\u00a0characteristics<\/span><\/h4>\n
Some road network characteristics included in\u00a0the\u00a0detailed typology can change abruptly, such as surface type or longitudinal drainage, and it is necessary to split the\u00a0assessed road segment at these points of\u00a0change. A\u00a0subset of\u00a0specific points from the\u00a0field survey was used for this segmentation. The\u00a0recorded vertical alignment break was checked with respect to the\u00a0proximity of\u00a0a\u00a0network node or a\u00a0local extreme identified from the\u00a0DTM. Changes in\u00a0surface type or longitudinal drainage were verified using the\u00a0photo documentation and surrounding characteristic points. In\u00a0total, around 130 road segments underwent additional segmentation.<\/span><\/p>\n
Based on spatial coincidence, the\u00a0road segments were then to be assigned attributes from the\u00a0layer of\u00a0characteristic points. Prior to this, however, it was necessary to check segments with multiple characteristic points assigned. The\u00a0number itself is not fundamentally an issue if the\u00a0points contain\u00a0identical characteristics. Thanks to its easy identification in\u00a0the\u00a0field, the\u00a0cleanest attribute was the\u00a0type of\u00a0surface. On the\u00a0other hand, the\u00a0most problematic characteristic was terrain\u00a0configuration (over 100 ambiguous assignments), since these changes in\u00a0the\u00a0field are always rather gradual and no specific-point category had been established for them. After checking the\u00a0consistency of\u00a0these data using photo documentation, the\u00a0road network was finally additionally split approximately halfway between two points with differing characteristics. Subsequently, road segments were assigned the\u00a0attributes from the\u00a0characteristic points.<\/span><\/p>\n
The\u00a0final step involved addressing segments without a\u00a0corresponding characteristic point. Their number depends on the\u00a0density of\u00a0points collected during the\u00a0field survey. A\u00a0typical example of\u00a0the\u00a0situation after projecting points onto road segments is shown in\u00a0Fig.\u00a06<\/span><\/em>. For empty segments, attributes from neighbouring segments were transferred iteratively (forward\/backward), provided that no specific point indicated a\u00a0change in\u00a0the\u00a0given attribute between them.<\/span><\/p>\n
<\/a><\/h6>\nFig. 6. Diagram of\u00a0transferring the\u00a0roadside drainage attribute to adjacent segments, taking into account specific points (purple crosses) marking changes in\u00a0the roadside drainage system; green triangles indicate characteristic points<\/h6>\n
Application of\u00a0the\u00a0detailed typology level and map creation<\/span><\/span><\/h4>\n
The\u00a0map with the\u00a0detailed typology level similarly builds on the\u00a0basic-level map and complements it with an appropriate representation of\u00a0the\u00a0detailed typology criteria. The\u00a0detailed map was designed as a\u00a0set of\u00a0two cartographic atlases: one for the\u00a0subsurface component of\u00a0runoff and the\u00a0other for the\u00a0surface component. The\u00a0graphic design consists of\u00a0a\u00a0clear combination of\u00a0line thickness (representing the\u00a0potential influence on the\u00a0runoff component from the\u00a0basic typology level), a\u00a0colour code expressing terrain\u00a0configuration as a\u00a0key factor affecting the\u00a0hydrological regime, and accompanying symbols to depict the\u00a0remaining characteristics of\u00a0the\u00a0detailed typology. The\u00a0result is shown in\u00a0Fig.\u00a07<\/span><\/em>.<\/span><\/p>\n
<\/a>\nFig. 7. Example of\u00a0applying detailed typology to the\u00a0road network in\u00a0the\u00a0Rennerovky pilot area<\/h6>\n
CONCLUSION<\/h2>\n
Linear elements in\u00a0the\u00a0landscape, and particularly the\u00a0road network, have a\u00a0significant potential to influence water runoff from an area. Under specific conditions, this influence can be positive; however, it generally tends to accelerate water runoff, which is usually considered undesirable. The\u00a0project, the\u00a0selected results of\u00a0which are presented in\u00a0this text, aimed primarily to provide the\u00a0data and tools needed to identify problematic sections of\u00a0the\u00a0road network and to minimise their undesirable effects, which lead to accelerated water runoff from areas under special nature protection. The\u00a0presented typology and methodology for its application to map outputs will assist in\u00a0identifying such problematic locations, at two possible levels of\u00a0detail \u2013 the\u00a0basic level, relying solely on available map data, and the\u00a0detailed level, utilising results of\u00a0field surveys focused on a\u00a0set of\u00a0clearly defined road characteristics. The\u00a0derived maps can serve as a\u00a0basis for selecting and prioritising road sections in\u00a0protected areas that are suitable for the\u00a0implementation of\u00a0mitigation measures, or even for the\u00a0complete removal and restoration of\u00a0a\u00a0road. The\u00a0principles of\u00a0such measures are addressed in\u00a0another output of\u00a0the\u00a0aforementioned project \u2013 the\u00a0Methodology for Recommendations on the\u00a0Construction of\u00a0New and Modification of\u00a0Existing Roads with Regard to Minimising Surface Runoff <\/span>\u2013 which is scheduled for publication by the\u00a0end of\u00a02025. Together, these outputs provide managers of\u00a0natural areas \u2013 whether under strict or general protection \u2013 with tools to better reconcile human interests in\u00a0accessing the\u00a0landscape with the\u00a0protection of\u00a0its natural runoff processes.<\/span><\/p>\n
Acknowledgements<\/h3>\n
This article was written with the\u00a0support of\u00a0Research Task V\u00da1: Research and Assessment of\u00a0the\u00a0Hydrological Regime under Current and Projected Conditions, within\u00a0the\u00a0Long-Term Development Concept of\u00a0the\u00a0TGM WRI for 2025.<\/span><\/span><\/em><\/p>\n
The\u00a0Czech version of\u00a0this article was peer-reviewed, the\u00a0English version was translated from the\u00a0Czech original by Environmental Translation Ltd.<\/span><\/p>\n","protected":false},"excerpt":{"rendered":"
The article presents the results of the project Analysis of Changes in the Water Regime of Land and Watercourses in the Krkono\u0161e National Park Caused by the Network of Roads (TA CR, no. TITSMZP945), implemented as a public procurement commissioned by the Ministry of the Environment of the Czech Republic within the BETA2 applied research programme. The main output of the project is a two\u2011level typology of the road network in terms of its impact on surface and subsurface runoff. This typology was applied to the territory of Krkono\u0161e National Park (KRNAP) in the Czech Republic and presented in the form of cartographic atlases. The article describes the principles and criteria of the proposed typology and the methodology of its application in map production, which at the basic level combines spatial analyses of road network datasets, digital terrain models and the hydrographic network, and at the detailed level incorporates the results of extensive field surveys. The original analytical procedures include, among other things, the detection of directional and elevation breaks in road segments and the delineation of micro\u2011catchments for individual sections. The resulting maps provide KRNAP Administration and other managers of protected areas with a tool for identifying road segments with the highest potential impact on the hydrological regime and serve as a basis for planning compensatory measures or restoration interventions.<\/p>\n","protected":false},"author":8,"featured_media":36439,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[86],"tags":[3934,3933,1569,3543,3932],"coauthors":[1065,235,692],"class_list":["post-36718","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-hydraulics-hydrology-and-hydrogeology","tag-field-mapping","tag-gis-runoff-analyses","tag-hydrological-regime","tag-krkonose-national-park","tag-road-network-typology"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/36718","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/comments?post=36718"}],"version-history":[{"count":6,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/36718\/revisions"}],"predecessor-version":[{"id":36724,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/posts\/36718\/revisions\/36724"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/media\/36439"}],"wp:attachment":[{"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/media?parent=36718"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/categories?post=36718"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/tags?post=36718"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/www.vtei.cz\/en\/wp-json\/wp\/v2\/coauthors?post=36718"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
![\"\"](\"https:\/\/www.vtei.cz\/wp-content\/uploads\/2025\/10\/Strouhal-tab-1-1.jpg\")
<\/a>\n