The study, published in Environmental Earth Sciences, integrated annual time series of four Sentinel-2 spectral indices with topographic variables derived from the Copernicus GLO-30 digital elevation model. The researchers computed NDVI to estimate photosynthetic activity, SAVI to reduce the influence of exposed soil in degraded areas, NDMI to capture water content in vegetation and soil, and BSI to highlight unvegetated surfaces. Each index was extracted from surface reflectance imagery at 10-meter resolution within a window of plus or minus 40 days centered on January 5th each year, from 2018 through 2026, with cloud masking applied through the Scene Classification Layer. For every scar, the spatial median of each index was aggregated annually, producing 765 observations across the 85 scars and nine years.

The scars themselves were filtered from a pre-existing inventory of 335 features mapped after the 2017 event, which generated an estimated 90 to 272 millimeters of rain within just four hours. An internal buffer of 10 meters eliminated edge pixels and reduced spectral mixing with intact surroundings, while additional filters removed polygons smaller than 100 square meters and highly irregular shapes. The resulting sample of 85 scars represents roughly 25.4 percent of the original inventory and about 16.6 percent of its total mapped area of 117.56 hectares. These scars sit almost entirely over volcanic rocks of the Serra Geral Formation, underlain by extremely shallow, stony Leptosols that confine hard rock or limiting material to within 25 centimeters of the surface, a configuration that limits water storage and root penetration and creates high susceptibility to rapid saturation during extreme events.

To unify the information held in the four indices, the team ran a principal component analysis on eight regeneration variables: the annual change rates and final 2026 values of NDVI, SAVI, NDMI, and BSI, with BSI’s sign inverted so that higher values consistently meant better condition. The first two components explained 96.1 percent of the standardized variance, and k-means clustering on their scores, with the optimal number of groups determined by average silhouette width, sorted the scars into three contrasting classes. The high-regeneration group contained 42.3 percent of the scars, the medium group 31.8 percent, and the low group 25.9 percent. The high class showed the largest mean values of NDVI, SAVI, and NDMI together with the lowest BSI, while the low class occupied the opposite corner of the spectral space, with the gap between extreme groups reaching 0.201 for NDVI and 0.180 for BSI.

The temporal trajectory that emerged from nine years of monitoring was emphatically nonlinear. Mean annual NDVI climbed from 0.446 in 2018 to 0.824 in 2026, SAVI rose from 0.669 to 1.236, and NDMI more than tripled from 0.094 to 0.313, while BSI fell continuously from near zero to strongly negative values, signaling that exposed soil was progressively blanketed by regrowing vegetation. The sharpest improvements concentrated in the initial years, particularly between 2018 and 2021, after which the rates of increment leveled off. The authors attribute this stabilization partly to optical index saturation in maturing, closing canopies, a well-documented sensitivity limit of normalized indices in dense vegetation, and partly to real ecological slowing of the kind reported in post-landslide studies in Japan, where full recovery was estimated to take on the order of a decade or more.

Linear mixed models with quadratic temporal terms confirmed that curvature in the recovery trajectories was statistically real rather than a visual artifact. Likelihood ratio tests comparing models with and without the quadratic time term were highly significant for all four indices, with chi-square statistics exceeding 300 for NDVI and SAVI, 44.3 for NDMI, and 68.0 for BSI, all at p below 0.001. The models included standardized geomorphometric variables as fixed effects, interactions of linear and quadratic time with each terrain attribute, and the number of valid images per year as a quality control, while random effects for intercept and slope per scar accommodated the repeated measures structure. Residual diagnostics showed no serious violations of normality or homoscedasticity, and variance inflation factors stayed below 4, ruling out multicollinearity among predictors.

Slope left its clearest imprint on NDMI, the moisture index. Scars on gentle slopes maintained consistently higher NDMI values throughout the series, whereas scars on the steepest terrain showed the lowest values in most years. The model coefficients confirmed the pattern: the main effect of slope on NDMI was negative at beta equal to minus 0.0326 with p equal to 0.0065, and the interaction with time was also significant, indicating that although steepness reduces mean moisture levels, the leveling off of moisture recovery is less pronounced on steeper slopes. The authors note that because NDMI, like NDVI, is a normalized band-ratio index barely affected by light and shadow differences on hillsides, these contrasts reflect genuine differences in vegetation moisture rather than calculation artifacts, consistent with the greater water availability and easier seed fixation and root anchoring on less inclined surfaces reported in comparable studies.

Elevation, by contrast, primarily modulated the bare-soil trajectory. Scars at higher altitudes exhibited faster reductions in BSI over time, with the linear interaction between time and elevation negative and significant at beta equal to minus 0.0027 and p equal to 0.045, meaning the annual rate of soil-cover gain accelerated with altitude. The high-elevation class recorded a total BSI reduction of 0.310 between 2018 and 2026, against 0.228 for the low class. One plausible mechanism is the solar radiation regime imposed by the rugged Serra Gaucho relief: slopes in lower valley positions may remain shaded by opposing higher slopes for part of the day, reducing the energy available for photosynthesis, while scars higher on the slopes receive more abundant direct insolation. The researchers are careful to frame topographic shading as a plausible mechanism rather than a demonstrated driver, since radiation was not measured directly, and unmeasured altitudinal gradients in temperature, precipitation, soils, or land use may also contribute.

Not every terrain attribute mattered. Scar shape, quantified by the size-independent circularity index, exerted no significant effect on any spectral index, and total terrain curvature, the second derivative of elevation, showed no significant primary effects or interactions for any index, with p-values above 0.05 throughout. Aspect fared no better as an inferential factor: the circular distribution of scar orientations showed low angular concentration, and while the high-regeneration class contained a slightly higher proportion of north-facing scars at 27.8 percent compared with 9.1 percent in the low class, the differences were too subtle to assert that orientation determines recovery under the region’s humid subtropical climate with relatively distributed rainfall. The terrain roughness index did not affect mean regeneration levels, though a significant quadratic interaction for NDMI suggested that moisture recovery levels off more sharply on surfaces with greater microrelief irregularity, possibly because micro-depressions act as temporary water reservoirs.

The practical upshot is a mapable framework for prioritizing restoration. Because 42.3 percent of scars achieved relatively greater spectral recovery while steep, low-moisture sites lag behind, the approach can flag where erosion and landslide reactivation risk remain elevated and where targeted intervention would pay off most. The authors stress that satellite metrics capture surface spectral signals rather than floristic composition or the structural reestablishment of native forest, and that conclusions about scar geometry apply strictly to the morphological range retained after filtering. They also acknowledge that the 30-meter DEM grid smooths fine-scale micro-topography, which may explain the limited significance of local curvature, and that the findings rest on a single basin. Still, within the Rolante watershed, where the scars were historically dominated by native forest cover and early regeneration is driven by fast-growing pioneers such as Trema micrantha and Solanum mauritianium, the message is clear: after the rain has done its damage, it is the shape of the land that decides how fast the forest returns.

Subject of Research: Vegetation recovery in landslide scars is controlled by topography, with slope and elevation shaping nonlinear spectral recovery trajectories in southern Brazil

Article Title: Topographic controls on vegetation recovery in mass-movement scars after an extreme rainfall event in Southern Brazil

Article References: Cargnin, B. D. R., de Oliveira Andrades-Filho, C., Ferreira, M. P., Herrmann, P. B., & de Oliveira, G. G. (2026). Topographic controls on vegetation recovery in mass-movement scars after an extreme rainfall event in Southern Brazil. Environmental Earth Sciences, 85(17), Article 438. https://doi.org/10.1007/s12665-026-13180-0

Image Credits: AI Generated

DOI: 10.1007/s12665-026-13180-0

Keywords: landslides, vegetation recovery, remote sensing, Sentinel-2, topography, Brazil, geomorphometry, ecological restoration, Topographic, controls, vegetation, recovery

The research, published in the journal Natural Hazards and Earth System Sciences, was led by Jeng-Hann Chong of the University of New Mexico and California State University Northridge, together with colleagues at several institutions. The team combined three field campaigns with high-resolution satellite imagery and numerical modeling to track how the likelihood of debris flows changed over four years in five steep, severely burned basins in the Angeles National Forest. Their central finding is striking: unsaturated soil hydraulic conductivity, a measure of how quickly water can soak into the ground, was roughly an order of magnitude higher during the dry season than during the wet season, sharply reducing the modeled probability of debris flows in summer even years after the fire.

The 2020 Lake Fire burned approximately 125 square kilometers of Los Angeles County, making it the second-largest fire in the county that year. According to Burned Area Emergency Response assessments, about 18 percent of the burn area experienced high burn severity. The researchers focused on five northwest-facing basins underlain by Late Cretaceous quartz diorite, with mean slopes of about 34 degrees. The United States Geological Survey had assigned these basins a debris-flow likelihood of 80 to 100 percent at a peak 15-minute rainfall intensity of 24 millimeters per hour, making them ideal natural laboratories for studying how hazard evolves over time.

Field work took place at three key moments: three months after the fire in December 2020, 45 months after the fire in April 2024 during the wet season, and 50 months after the fire in September 2024 during the dry season. At basin outlets, the team collected sediment samples for grain-size analysis and used a mini disk infiltrometer to measure how fast water penetrated the soil surface without disturbing it. During the first survey, a water droplet penetration test revealed strong soil water repellency, a hallmark of freshly burned ground. Because access to the basins was difficult early on, the team assumed a conservative hydraulic conductivity of 10 millimeters per hour for the initial period, based on measurements from the nearby 2016 Fish Fire in the San Gabriel Mountains, which shares similar bedrock, slopes, and vegetation.

To track vegetation recovery, the researchers processed 61 monthly PlanetScope satellite images at 3-meter resolution, spanning from before the fire through September 2024. They computed the differenced Normalized Difference Vegetation Index, or dNDVI, which highlights where vegetation was lost or gained between two dates. Negative dNDVI values traced the burn perimeter immediately after the fire and matched the areas flagged as high severity by emergency response teams. In later years, positive values along the channels and within the basins signaled the rapid growth of grasses and shrubs, while negative values along the river suggested sediment deposition or erosion during storm events. A ground-based point-intercept survey along a riverbank measured roughly 81 percent vegetation cover, in reasonable agreement with the satellite-derived estimate of about 70 percent.

The rainfall record added crucial context. Peak 15-minute rainfall intensity reached only about 13 millimeters per hour in the first year, well below the modeled debris-flow threshold. Over the full study period, twelve events exceeded 15 millimeters per hour, four exceeded 20 millimeters per hour, and none exceeded 25. Cumulative annual rainfall varied enormously, from 127.5 millimeters in the first year to 1329 millimeters in the third, reflecting the whiplash between drought and atmospheric-river-driven wet years that characterizes Southern California’s climate. This variability matters, because soil moisture left behind by preceding storms shapes how the next storm behaves.

The modeling itself relied on the KWAVE numerical model, which routes overland flow using the kinematic wave approximation, computes infiltration with the Green-Ampt equation, and accounts for canopy interception with the Rutter model. The simulated runoff then fed into a slope-dependent dimensionless-discharge model that estimates the rainfall intensity-duration threshold needed to trigger a runoff-generated debris flow. Immediately after the fire, the simulations showed that most basins required a peak 15-minute intensity of 30 millimeters per hour to initiate debris flows. By the wet season of year four, the picture had changed in a counterintuitive way: despite substantial vegetation regrowth, all basins could still generate debris flows at 30 millimeters per hour, and one basin could be triggered at just 20 millimeters per hour, an intensity that was actually recorded twice in February 2024.

The dry season told a very different story. With hydraulic conductivity roughly ten times higher than in spring, the soils acted like a sponge, absorbing rainfall and suppressing runoff. Three of the four studied basins showed little to no surface runoff even at 30 millimeters per hour, and the simulations indicated that triggering a debris flow would require intensities of about 80 millimeters per hour, corresponding to a 100- to 200-year recurrence interval storm. The researchers attribute this seasonal shift partly to the opening of macropore fractures in dried soils, which lets water drain rapidly, and partly to grasses and shrubs drawing moisture out of the top half-meter of soil through evapotranspiration. A related study after Arizona’s Contreras Fire found a roughly fivefold seasonal increase in hydraulic conductivity; the Lake Fire measurements showed even larger swings.

Perhaps the most surprising result came from the sensitivity analysis, in which the team varied hydraulic conductivity, grain size, and vegetation cover one at a time. Vegetation cover turned out to have remarkably little influence on debris-flow likelihood: a basin with full vegetation cover behaved almost identically to a bare one when other soil properties were held constant, likely because canopy storage saturates quickly during intense rain. Grain size mattered more, with halving the median grain size substantially increasing the number of slopes that could trigger flows in some basins. But hydraulic conductivity dominated, and because it can change by an order of magnitude within a single season, the authors argue that it exerts stronger control on hazard than any slowly evolving property of the landscape. This contrasts with findings from the 2016 Fish Fire, where vegetation recovery played a larger role in raising triggering thresholds.

The implications reach well beyond one burn scar. Hazard assessments that treat post-fire debris-flow risk as a smooth function of time since burning may badly misjudge the danger in any given month, particularly in Mediterranean climates with strong wet-dry contrasts. The study also suggests that recovery is not a one-way street: hydraulic conductivity may follow a non-monotonic, seasonally oscillating path rather than steadily returning to pre-fire levels. The authors caution that their measurements were limited to basin outlets and single visits per season, and that future work should sample within basin interiors, integrate lidar to quantify erosion, and characterize vegetation type and density more precisely. But the core message is clear: to know when a burned mountainside will let go, forecasters must ask not only how long ago the fire burned, but what season it is, and how wet the soil has become.

Subject of Research: Seasonal controls on post-fire debris flow likelihood following the 2020 Lake Fire in Southern California

Article Title: Seasonal influence on post-fire debris flow likelihood after the 2020 Lake Fire

Article References: Chong, J.-H., Ghosh, A., Page, B. T., Jesmok, G., Berg, D. V., Lopez, M., Upadhyay, D., Stone, D. J., Hauswirth, S. C., Lindsey, E. O., & Scuderi, L. A. (2026). Seasonal influence on post-fire debris flow likelihood after the 2020 Lake Fire. Natural Hazards and Earth System Sciences, 26(10), 4723-4740. https://doi.org/10.5194/nhess-26-4723-2026

Image Credits: AI Generated

DOI: 10.5194/nhess-26-4723-2026

Keywords: post-fire debris flows, wildfire, hydraulic conductivity, Lake Fire, soil hydrology, vegetation recovery, rainfall thresholds, remote sensing, NDVI, Angeles National Forest, debris flow modeling, seasonality

The challenge that motivated the study is one that anyone who works with satellite archives knows intimately: the record is finite. Satellites do not observe the world continuously or forever. Cloud cover, orbital gaps, sensor outages and the simple fact that any archive has a beginning and an end mean that scientists rarely witness the full life cycle of a disturbance. For a burned pixel, the true moment of ignition and the true moment of recovery may both fall outside the window of observation. Statisticians call this censoring, and ignoring it can seriously bias estimates of how long recovery actually takes. If you only count the scars whose entire story fits inside your observation window, you systematically exclude the slowest-recovering landscapes, the ones that matter most for understanding ecological resilience.

Nora Khalil of Capital University in Egypt tackled this problem by borrowing a conceptual toolkit from survival analysis, the branch of statistics originally developed to model human lifetimes and the reliability of mechanical components. In her formulation, each burned pixel has a persistence lifetime: it begins when the vegetation’s spectral signal degrades past a defined threshold and ends when that signal recovers to near its pre-fire baseline. Some of these lifetimes are fully observed, some are right-censored because the pixel was still degraded when the archive closed, and a few span the entire analysis window, providing only the information that persistence exceeded the window’s length. Crucially, although some episodes appear doubly censored in calendar time, the statistical inference is conducted entirely on the persistence-time scale, where all incomplete episodes enter the likelihood as right-censored observations.

The data came from Landsat surface-reflectance time series at Monitoring Trends in Burn Severity sampling locations within a large fire scar in interior Alaska, centered near 65.5 degrees north latitude. Khalil assembled NDVI histories, a standard greenness index derived from the difference between near-infrared and red reflectance, for roughly one hundred burned pixels, drawing on observations from Landsat 5, Landsat 7 and Landsat 8 between 2000 and 2020. Cloudy, hazy and snow-covered observations were discarded. For each pixel, a healthy-vegetation baseline was defined as the 80th percentile of its own NDVI record, an approach that adapts to local conditions rather than imposing a single regional standard. A burn episode was declared when smoothed NDVI fell below 60 percent of that baseline, and recovery was declared when it climbed back to at least 85 percent.

To mimic the incomplete monitoring conditions typical of real satellite archives, the analysis was restricted to a window spanning 1 January 2010 to 31 December 2018. This deliberate choice produced a balanced mix of observation types: of the 90 persistence episodes in the final dataset, 49, or 54 percent, were fully observed, 35, or 39 percent, were right-censored, and 6, or 7 percent, remained degraded throughout the entire window. Nearly half the episodes were therefore incomplete, a striking illustration of why naive approaches that use only raw observed durations would discard a substantial portion of the available information and skew the picture of recovery.

The heart of the methodological contribution lies in how the persistence times were modeled. Khalil assumed they follow a Weibull distribution, a flexible two-parameter family whose shape parameter reveals whether the recovery hazard, the probability of recovering in the next interval given that recovery has not yet occurred, increases or decreases with time, and whose scale parameter sets a characteristic duration. Standard maximum likelihood estimation of these parameters is efficient when the model is correct, but it is notoriously sensitive to outliers. Satellite data are full of them: residual cloud contamination, geolocation errors, mixed pixels and misclassification can all produce anomalous readings that distort the estimated hazard and bias ecological interpretation.

To guard against such contamination, the study employed a weighted-likelihood estimator with Huber-type influence weights. The idea is elegant: each observation is compared against the currently fitted model, and observations that disagree strongly with it receive progressively reduced weight, down to a floor set by a tuning constant. Nothing is discarded outright; even the most discordant episodes retain some influence, but atypically long-lived scars no longer dominate the fit. The tuning constant itself was chosen data-driven, by minimizing the squared discrepancy between the fitted Weibull distribution and a nonparametric Kaplan-Meier or Turnbull estimate of the survival curve over its central range. Uncertainty was quantified through a parametric bootstrap with 1000 replications, preserving the observed censoring pattern in each synthetic dataset.

The results paint a coherent picture of post-fire recovery in the subarctic. The fitted Weibull model indicates a median persistence of approximately 2.7 to 2.9 years, with a 90th-percentile persistence of roughly five to six years, meaning one in ten burned areas remains spectrally scarred for at least that long. The shape parameter consistently exceeded one, indicating an increasing recovery hazard: the longer a scar has persisted, the more likely it is to recover in subsequent months. This pattern aligns with ecological expectations, since vegetation regrowth tends to accelerate once seedlings and resprouting vegetation establish themselves. Bootstrap diagnostics showed that the robust estimator narrowed the dispersion of parameter estimates compared with maximum likelihood, driven by downweighting a small number of very long-lived scars, while leaving the central tendency essentially unchanged. Sensitivity analyses across tuning constants from 1.0 to 1.5, and across alternative NDVI thresholds, baselines and smoothing choices, confirmed that the substantive conclusions are not artifacts of any single arbitrary specification.

An honest comparison with alternative distributions revealed some interesting tensions. The lognormal model actually attained the smallest AIC and BIC information criteria, with the gamma distribution close behind, yet the estimated medians and upper-tail quantiles were broadly comparable across all three families. Khalil retained the Weibull as the primary model for two reasons: it offers a direct monotone hazard interpretation that maps naturally onto ecological regrowth, and it integrates seamlessly with the established robust censored-likelihood machinery. A Weibull probability plot showed approximate linearity over the central quantile range, with departures confined to the sparse tails where few uncensored observations exist. The fitted survival curve closely tracked the Kaplan-Meier and Turnbull empirical estimators over the central duration range of roughly 500 to 1500 days.

Perhaps the most exciting aspect of the work is its generality. The same framework, persistence as a censored lifetime, estimated with adaptive robust weighting, applies naturally to other environmental processes observed under incomplete monitoring: snow-cover longevity, floodwater retention, orbital-debris decay, drought impacts, land-use transitions and vegetation browning. In the fire context, the author notes that the method could be applied to events such as the 2019-2020 Australian bushfires or large Amazonian drought fires by adapting only the NDVI extraction and baseline definitions. Future extensions could incorporate spatial pooling, climate covariates, fire severity effects and hierarchical structures combining multiple fire events. The study is not without limitations, the sample of 90 episodes with only six window-spanning cases constrains precision in the extreme upper tail, but as climate change intensifies fire regimes worldwide, having a statistically principled, robust and interpretable tool to measure how long landscapes carry their wounds has never been more timely.

Subject of Research: Statistical modeling of post-fire vegetation recovery duration using censored Weibull lifetime analysis of satellite NDVI time series

Article Title: Robust doubly censored Weibull modelling of NDVI-based burn-scar persistence in satellite time series

Article References: Khalil, N. (2026). Robust doubly censored Weibull modelling of NDVI-based burn-scar persistence in satellite time series. Advances in Statistical Climatology, Meteorology and Oceanography, 12(1), 111-121. https://doi.org/10.5194/ascmo-12-111-2026

Image Credits: AI Generated

DOI: 10.5194/ascmo-12-111-2026

Keywords: wildfire, burn scar, NDVI, Landsat, survival analysis, Weibull distribution, censored data, robust estimation, remote sensing, Alaska, vegetation recovery, MTBS

A team of researchers led by Francisco Gómez-Prieto of the Universidad Autónoma de Madrid set up an experimental network of corridors in a forest clearing at roughly 1,300 meters elevation, surrounded by pine stands of Pinus nigra, Pinus pinaster and Pinus sylvestris. The site was deliberately chosen because it resembles terrain already used by sporting events in the park. On flat ground, the team laid out four corridors, two assigned to running and two to cycling, each subdivided into transects receiving either zero, 500 or 1,000 passes. The 500-pass treatment follows the standard trampling protocol of Cole and Bayfield, while 1,000 passes reflects the participation numbers typical of real events held in the park. On two sloped areas of 12.5 and 10 degrees, corridors received 400 passes of running descents, cycling descents and, in one case, cycling ascents.

The treatments were applied over two consecutive days at the end of February 2024 by three participants with an average body mass of 75 kilograms, using standard athletic shoes for running and two mountain bikes, including an electric model on the slopes. Sampling took place before the impact and then in March, June, October and February 2025, tracking soil penetration resistance, vegetation cover and composition, bacterial counts and microarthropod abundance. This design allowed the team to separate the immediate shock of a single mass event from the slower process of ecological recovery across a full Mediterranean year.

The soil results were unambiguous. After 1,000 passes, penetration resistance in the center of the flat corridors reached 4.74 kilograms per square centimeter for running and 5.37 for cycling, both significantly higher than the less-trampled lateral zones. Mountain biking compacted the soil more than running, a pattern the authors attribute to the continuous passage of knobby tires along the same narrow tracks, in contrast to the intermittent and spatially dispersed pressure of footfalls, which become even more scattered at a runner’s longer stride. Strikingly, even a full year after the disturbance, compaction in the impacted centers remained significantly above the levels of adjacent ground, indicating that soil recovery operates on a much slower clock than vegetation.

Vegetation told a different and more hopeful story. Running at 1,000 passes stripped away half of the plant cover in the flat corridors, while cycling removed about 15 percent, and 500 passes of cycling produced no significant loss at all. Yet by June, just three months later, cover in the trampled plots was statistically indistinguishable from pre-impact conditions. The explanation lies in the flora itself: the community was dominated by trampling-tolerant perennial herbs such as Poa bulbosa and Rumex acetosella, low-growing species with narrow leaves and flexible structures. The single shrub recorded, Santolina rosmarinifolia, was rare and escaped visible damage, apparently because runners instinctively avoided its growth form. This resilience contrasts sharply with studies in Australian mountains and semi-arid American grasslands, where recovery often exceeds a year, and it underscores how strongly local plant traits govern vulnerability.

Slope and direction of travel emerged as powerful modifiers of impact. On the inclined corridors, both activities compacted the soil and stripped vegetation during descents, producing denudation comparable to 1,000 passes on flat terrain despite only 400 passes being applied. Ascent cycling, by contrast, caused minimal measurable change, with any small loss in cover fully recovered by June. The authors suggest that steeper gradients channel more force into the surface and that downhill wheels cut ruts which later concentrate water flow and accelerate erosion. This echoes earlier work by Pickering and Growcock showing that slope can matter more than use intensity, and it carries a direct message for race planners: downhill segments on steep ground are where a given number of participants does the most damage.

Not every indicator proved useful. Bacterial colony-forming units per gram of dry soil fluctuated mainly with seasonal moisture rather than with activity type or pass count, and microarthropod abundance, including springtails and mites, showed no clear response to trampling at all. Instead, these soil organisms exploded in October as the Mediterranean summer drought ended, consistent with the on-off influence of humidity on arthropod demography in Mediterranean ecosystems described by Touloumis and Stamou. The authors conclude that only penetration resistance and vegetation cover delivered consistent, interpretable signals, and they recommend these two measures as the backbone of any monitoring program for recreational impacts in comparable environments.

Community-level shifts added a subtle warning. Permutational multivariate analysis revealed that from June onward, plant community composition differed significantly according to impact intensity but not activity type. Denuded patches were recolonized by fast-growing species such as Rumex acetosella, which can displace slower-establishing plants and reshape the original community. Simpson’s diversity index showed post-impact homogenization, likely reflecting the relative rise of trampling-resistant species at the expense of more sensitive ones, although heterogeneity partially returned as other species re-established. Such compositional drift may be invisible in a simple cover measurement but could accumulate over repeated events into a lasting ecological debt.

For the managers of Sierra de Guadarrama and protected areas like it, the study’s practical implications are concrete. Because the impacts observed were largely reversible when the disturbance occurred in winter, before germination and active growth, well-regulated events appear compatible with conservation provided limits of acceptable change are respected and monitoring is in place. The authors recommend avoiding routes with gradients exceeding 15 percent, particularly descents, spacing events on the same trails to allow recovery, steering clear of sensitive seasons such as spring for breeding wildlife or muddy autumn trails, and concentrating use on wide, stable paths. Where sites hold particularly fragile values, outright bans remain a legitimate option, since conservation comes first in many protected areas.

The team is careful to spell out the limits of their findings. The experiment applied a single impact event in winter, whereas most real races take place in spring, when plants are in vulnerable phenological stages and damage could persist far longer. Recovery under repeated, year-round trampling may differ substantially from recovery after one simulated race, and the sloped plots, which were not fenced, suffered vandalism that cut short their monitoring. Participant numbers, moreover, influenced some indicators but were not the dominant driver of damage; route design and event timing mattered more. As e-bikes open previously inaccessible terrain to ever larger numbers of riders, the authors argue that adaptive management, grounded in robust indicators and seasonal awareness, is the key to keeping Mediterranean mountain trails both thrilling to run and alive beneath the wheels.

Subject of Research: Experimental assessment of the ecological impacts and recovery dynamics of trail running and mountain biking in a Mediterranean mountain national park

Article Title: Impacts of trail running and mountain biking: Resilience of a Mediterranean mountain ecosystem

Article References: Gómez-Prieto, F., Tejedo, P., Faucha, M., Escudero, A., Leung, Y.-F., Valladares, F., & Benayas, J. (2026). Impacts of trail running and mountain biking: Resilience of a Mediterranean mountain ecosystem. iScience, 29(10), Article 117721. https://doi.org/10.1016/j.isci.2026.117721

Image Credits: AI Generated

DOI: 10.1016/j.isci.2026.117721

Keywords: trail running, mountain biking, soil compaction, vegetation recovery, Sierra de Guadarrama, protected areas, recreation ecology, Mediterranean ecosystems, trampling, national park management, soil microbiology, slope erosion