Populus nigra ita. is an important tree species for preventing rainfall-triggered shallow landslides and hydraulic bank erosion in New Zealand. However, the quantification of its spatial root distribution and reinforcement remains challenging. The objective of this study is to calibrate and validate models for the spatial upscaling of root distribution and root reinforcement. The data were collected in a 26-year-old “Tasman” poplar stand at Ballantrae Hill Country Research Station in New Zealand. We assessed root distribution at different distances from the stem of four poplar trees and from eleven soil pits along a transect located in a sparse to densely planting poplar stand. 124 laboratory tensile tests and 66 field pullout tests on roots with diameters up to 0.04 m were carried out to estimate root mechanical properties. The results show that the spatial distribution of roots can be well predicted in trenches of individual tree root systems (R² = 0.78), whereas it tends to overestimate root distribution when planting density was higher than 200 stems per hectare. The root reinforcement is underestimated within single tree root systems (R² = 0.64), but it performs better for the data along the transect. In conclusion, our study provided a unique and detailed database for quantifying root distribution and reinforcement of poplars on a hillslope. The implementation of these models for the simulation of shallow landslides and hydraulic bank erosion is crucial for identifying hazardous zones and for the prioritization of bio-engineering measures in New Zealand catchments. Results from this study are useful in formulating a general guideline for the planning of bio-engineering measures considering the temporal dynamics of poplar’s growth and their effectiveness in sediment and erosion control.

Poplar (Populus sp.) is an important species for preventing shallow, rainfall-triggered landslides and hydraulic bank erosion in New Zealand. However, quantifying the spatial root distribution pattern and reinforcement remains challenging. This study aimed to find the Root Bundle Model with the Weibull survival function (RBMw), a root distribution model (RDM), and a root reinforcement model for the implementation in models such as BankforMAP and SlideforMAP. Our study was conducted within a 26-year-old “Tasman” poplar stand at Ballantrae Hill Country Research Station in the North Island of NZ. We measured root distribution at distances of 1.5, 2.5, 3.5, and 4.5 m from the stem of four poplar trees whose diameters ranged from 0.41 to 0.56 m and from eleven soil profiles along a transect located in a sparse to a densely planted poplar stand. This created a unique database of root distribution. 124 laboratory tensile tests and 66 field pullout tests on roots with diameters up to 0.04 m were carried out. The root distribution model well predicted spatial root partition in trenches of single tree root systems with R2 = 0.78 and in the transect with R2 = 0.85. The model tends to overestimate root distribution when planting density was higher than 200 stems per hectare. The maximum lateral root reinforcement model tends to underestimate forces in single tree root systems with R2 = 0.64, but it well performs along the transect within the stand with different planting densities. The basal root reinforcement model performed well in predicting its vertical distribution as a function of soil depth. In conclusion, our study provided a detailed dataset for the quantification of root distribution and reinforcement of poplars on a hillslope for the purpose of increasing slope stability and mitigating hydraulic bank erosion. The implementation of these data in models for the simulation of shallow landslides and hydraulic bank erosion is fundamental for the identification of hazardous zones and the prioritization of bio-engineering measures in NZ catchments. Moreover, the results are used to formulate a general guideline for the planning of bio-engineering measures considering the temporal dynamics of poplar’s growth and their effectiveness in sediment and erosion control.

Surface runoff in alpine catchments is sensitive to rainfall intensity, resulting in substantial variations in runoff coefficients across rainfall scenarios. Although this dependency is well documented in the literature, it is often neglected in practical applications by assuming a fixed runoff coefficient for all events. Moreover, quantitative assessment of infiltration processes is time-consuming, and operational hazard-assessment methods commonly rely on qualitative infiltration assumptions rather than in situ measurements. In this study, we test a new framework that links simple field infiltration tests with the estimation of event-dependent surface runoff coefficients. To this end, we compared single-ring infiltrometer measurements (SR) with 1 ${\mathrm{m}}^2$ rainfall-simulation experiments (RSEs) across four land-use types: meadow, afforestation, broadleaf forest, and spruce forest. This study’s novelty lies in the cross-scale integration and explicit evaluation of established methods for operational hazard assessment, conducted in an alpine catchment near Brienz, Switzerland. The results show that directly upscaling surface runoff coefficients based on field-measured saturated hydraulic conductivity values provides a solid basis for estimating their order of magnitude, although discrepancies can be large depending on the approach used and the rainfall intensity considered. At $i=140$ mm ${\mathrm{h}}^{-1}$, RSE-derived mean apparent infiltration ranged from $i=94.7$ mm ${\mathrm{h}}^{-1}$ in spruce forest to $i=123.3$ mm ${\mathrm{h}}^{-1}$ in mixed broadleaf forest; the lowest individual plot value was $i=56.8$ mm ${\mathrm{h}}^{-1}$ for spruce forest. The Stone model fitted the trench-level RSE data with $R^2\ge 0.94$, whereas the two-parameter Bonetti inversion reached a parameter bound in 9 of 12 trenches and did not yield physically interpretable $K_{\mathrm{fs}}$ distributions. Relative to RSE-Stone at $i=140$ mm ${\mathrm{h}}^{-1}$, SR-Bonetti predicted approximately 2–7 times higher surface runoffs for the three non-spruce land uses, while agreement was obtained only for the sampled spruce site. We discuss several potential reasons for the discrepancy between the RSE and SR results, including disturbance of the organic layer during SR measurements, seasonal variability in soil conditions and differences in the number of repetitions. Although the proposed framework provides a feasible approach to improving the quantitative assessment of flash-flood hazards in alpine catchments smaller than 10 ${\mathrm{km}}^2$, further research is needed to validate its results and enhance its practical applicability.

Protection forests play a critical role in mitigating surface landslides and controlling hydrological processes, yet their identification and assessment remain a challenge in forest and land management. This study, conducted as part of the PRIN-PNRR MILETO project, introduces in Italy a novel procedure for identifying protection forests using a deterministic statistical approach tailored to surface landslides in Italy. The SlideForMAP software forms the core of this methodology, integrating key inputs on soil and vegetation characteristics to assess landslide susceptibility. By explicitly incorporating the role of vegetation, the software offers a refined analysis of areas prone to landslides. Computationally efficient, the method supports the evaluation of extensive regions, facilitating applications at a regional scale. In southern Italy the MILETO project has implemented this methodology to map and evaluate protection forests though case studies. These areas, often characterized by steep terrain and varying climatic conditions, are particularly prone to hydrogeological hazards like landslides. The project focuses on linking hydrological and soil stability models with vegetation dynamics, a key determinant in mitigating landslide risk.These outputs provide actionable insights for forest and land managers. The hazard maps enable planners to pinpoint locations where protection forests mitigate landslide risks most effectively, while heat maps highlight areas for intervention to enhance forest functionality. This systematic approach bridges the gap between theoretical modeling and practical forest management, supporting sustainable landscape practices and disaster risk reduction approaches. By focusing on direct protection forest detection, this case study in southern Italy contributes to integrating environmental modelling and geospatial data to create a robust framework for safeguarding vulnerable regions.