Introduction

Landforms play a pivotal role in shaping ecosystems and influencing land use practices, as well as in governing flood dynamics across diverse geographical areas. Concurrently, prioritizing the study and conservation of natural systems is imperative for attaining sustainable development objectives. Therefore, the utilization of spatial data analysis and elevation models holds significant importance in assessing the characteristics of various landforms (Lin et al., 2021; Dikau, 2020). Research indicates that geological attributes profoundly influence the formation and characteristics of landforms (Nazaruddin, 2020; Kurnianto et al., 2023). This correlation stems from the intricate relationship between elevation data and geomorphological processes, making it a valuable proxy for geological formations (Dede et al., 2024). The concept of Terrain Position Index (TPI) was elucidated at the ESRI International Conference through an informative poster presentation detailing its calculation methodology (Weiss, 2001). However, delineating many landform classifications necessitates comprehensive investigations into various geological features as primary inputs (Laamrani et al., 2014; Dragut et al., 2006).



Materials and methods

In this study, the TPI method was employed to create a map depicting the landforms of the region. The Terrain Position Index (TPI) quantifies the variation between the elevation at the central point (z0) and the mean elevation (ẑ) of the surrounding area within a designated radius (R). In the formula, (n) represents the total number of neighboring points considered during the assessment.

Z ̂=1/nR 1 nR∑▒i∈R Z_i (1)

TPI yields positive values when the central point is situated at a higher elevation compared to its surroundings, while negative values indicate a lower elevation. Typically, TPI outputs fall within the range of +1 to -1. Values beyond this range might suggest anomalies within the elevation dataset.



Moreover, the Deviation (DEV) metric assesses the geological context of the central point (z0) by integrating TPI with the standard deviation (SD) of the elevation. The equations are as follows (Wilson J., Gallant, 2000):

DEV=(z_0-Z ̂)/SD (2)

SD=√(1/nR) ∑▒i=〖(Z_i-Z ̂)〗^2 (3)

TPI<-focal (x,w=f,fun=function (x,…)×[5]-mean (x[-5])) (4)

Positive TPI values signify elevated areas like peaks, while negative values denote low-lying regions such as valleys, and zero values represent flat terrains like plains (Wilson J., Gallant, 2000) (Figure 3). It's noteworthy that in this study, TPI maps and subsequent landform maps were prepared using scales of 10, 20, 30, 40, 50, and 60 meters.

discussion and Results

TPI serves as a valuable tool for categorizing landforms by revealing disparities across various scales. Elevated TPI values (positive) typically denote ridges and peaks, whereas negative values indicate low-lying areas, with values nearing zero indicating flat or gently sloping terrain. Consequently, the study area's landforms were classified into 10 categories based on TPI variations across the six different scales employed in this investigation, as illustrated in Figure 8. The generated TPI outcomes underscored the significant impact of spatial resolution on landform classification, manifesting differences in the proportion of landform areas between high and low TPI values. For instance, transitioning from the finer scale of 10 meters to the broader scale of 60 meters resulted in a reduction in the proportion of U-shaped valleys from 5.64% to 2.61%, as well as a decrease in the percentage of valleys and cuts situated on uplands and slopes from 67.55% to 66.05%. Conversely, the proportion of high ridges experienced an increase from 0.81% to 2.84%, while that of mountain peaks rose from 0.11% to 0.39% (Table 4).

Conclusion

Valleys predominantly host bare lands, constituting the highest percentage at 92%, while agricultural lands are primarily situated in mid-range drainages and plains, covering approximately 92% of the area. Pasture lands are predominantly located in mid-slope drainages, while forest lands are observed in mid-slope drainages as well as in high and sloping areas.

Urban areas exhibit the highest proportion in mid-slope drainages and slopes, whereas garden lands are most prevalent in drains, valleys, and sloping terrains. This diversity underscores the distribution of various land uses such as agriculture, natural resources, wildlife habitats, urban areas, road networks, pastures, protected areas, agroforestry, and ecotourism across distinct landform features. It's worth noting that the analysis of land use and land shape considered a neighborhood radius of 10 meters.

Regarding flood susceptibility, low flood risk classes are predominantly associated with high lands (high ridges), while medium flood risk occurs in mid-range drainages, and high flood risk is prevalent in numerous landforms including drains, valleys, and sloping terrains. Mid-range drainages are particularly susceptible to flooding. As indicated in Table 3, the highest classes of flood hazard risks are correlated with plains, while the lowest risk classes are attributed to peaks with medium and high slopes and upstream drainages. Areas characterized by high slopes and peaks generally exhibit low flood risk, while those with mid-slope drainages and sloping lands fall within the moderate risk category, and valleys pose a relatively higher flood risk. This model offers insights into flood risk assessment and facilitates sustainable management strategies for the region.