Experimental assessment of the impacts of fire and agriculture on the carbon dynamics of a páramo grassland in northern Ecuador

Abstract

Prescribed fires, associated to extensive cattle raising, and the expansion of crop agriculture are two of the main threats to the integrity and functionality of páramo ecosystems in the Northern Andes. In addition to affecting plant and animal communities, these practices could significantly affect soil carbon storage and hydrological regulation. In this study, we present results from a field experimental assessment of the impacts of fires and potato cultivation on carbon dynamics in a grassland ecosystem in the páramo of Northern Ecuador. Using the static chamber method and a portable gas analyzer, we characterized carbon dynamics in 27 experimental 6.25 m² plots, randomly assigned to control, vegetation burning, or potato cultivation treatments. Unlike Control plots, Burning and Potato plots predominantly lost carbon during the first half of the experiment, and then shifted to negative values of net carbon exchange (NEE) around day 150. Specifically, NEE in the potato plots decreased from an average of 0.9 g CO₂ m^-2 hr^-1 during the first 150 days of the experiment, to −0.6 g CO₂ m^-2 hr^-1 during the second half of the study. Corresponding values for burning treatment plots represent a decrease from 0.5 g CO₂ m^-2 hr^-1 during the first half of the experiment to −0.3 g CO₂ m^-2 hr^-1 in the second half. Low rates of NEE in Control plots suggest that daytime carbon uptake capacity may be reduced under the evaluated conditions, although annual carbon balance cannot be inferred from the present dataset. These results underscore the need to further evaluate the practice of anthropogenic fires in páramos and limit the expansion of agriculture, as critical strategies for the long-term maintenance of páramo ecosystem services.

1 Introduction

Current conservation of tropical mountain ecosystems sits at a critical crossroad. On the one hand, recent decades have seen a significant increase in academic interest and understanding about these ecosystems and their irreplaceable role in supporting human wellbeing (Grêt-Regamey and Weibel, 2020). On the other hand, human activities -including rapid climate change and land-use changes, are increasingly impacting mountain environments and threatening their long-term persistence (Price and Egan, 2017; Schickhoff et al., 2022). Of special concern are the impacts that mountain ecosystems transformation could bring upon ecosystem carbon dynamics, not only because of the potential loss of carbon from these characteristically soil carbon-rich ecosystems (Calispa et al., 2023), but also because such loss of soil organic matter can lead to deterioration of other ecosystem services such as the capacity to hold and regulate water (Rawls et al., 2003).

Páramo ecosystems in the northern Andes experience intense pressure from encroachment of agricultural activities (Peyre et al., 2021; Thompson et al., 2021) and fires associated with extensive cattle raising (Matson and Bart, 2013; Zomer and Ramsay, 2021; Obando-Cabrera et al., 2025), which might be exacerbated by ongoing climate change (Buytaert et al., 2002; Cuesta et al., 2012; Cresso et al., 2020). In addition to changing native plant cover and affecting water quality and aquatic ecosystems (Matson and Bart, 2013; Rey-Romero et al., 2022; Duchicela et al., 2024), these activities may alter carbon dynamics through changes in environmental conditions (e.g., soil temperature and moisture) and soil structure (Poulenard et al., 2001; Podwojewski et al., 2002). Despite the potential importance of these changes, most of our understanding about fires and agriculture is biased towards their impacts on plant community structure (Matson and Bart, 2013), or soil physical characteristics, including total carbon storage in páramo soils and vegetation (Thompson et al., 2021; Patiño-Gutiérrez et al., 2024). In contrast, little is known about the effect of these activities on the actual dynamics of soil carbon, particularly in relation to land uses that are key drivers of environmental degradation.

The impacts of land use on the carbon dynamics of mountain ecosystems have been previously analyzed in extensively used ecosystems of the European Alps. Although results are variable, net carbon ecosystem exchange (NEE) seems to be driven by land management and, specifically, by the effects of mowing, grazing, and fertilization on gross ecosystem productivity (GPP), and ecosystem respiration (R_eco) (Schmitt et al., 2010; Zeeman et al., 2010; Berninger et al., 2015). In these ecosystems, grassland management tends to increase net carbon storage because of fertilization and the resulting increase in GPP. The magnitude of these increases, however, also depend on seasonality and its influence on temperature and water availability.

Carbon dynamics of tropical grassland ecosystems have been much less and partially evaluated. In the Venezuelan páramos, Sarmiento and Bottner (2002) and Cabaneiro et al. (2008) showed that, although potato cultivation can result in decreased decomposition of organic matter, these declines in soil respiration can be overcome by the addition of labile organic matter that results from fallow periods of up to 15 years. In contrast, a warming experiment with open-top chambers in the Colombian páramos found no effect of increased temperatures on the rates of soil respiration and NEE (Lasso et al., 2021). Finally, results from an Eddy-covariance tower in a grassland páramo in Southern Ecuador showed that, over a 2-year period, the ecosystem acted as a net source of carbon to the atmosphere losing between 68 and 128 gC m^-2 year^-1 (Carrillo-Rojas et al., 2019). This last result is specially concerning as rapid loss of carbon could hamper the contribution of páramo soils to climate change mitigation and, potentially, their role in water supply and regulation (Buytaert et al., 2006; 2011; Patiño-Gutiérrez et al., 2024).

Fires and potato cultivation are among the most widespread and frequent human-induced changes in páramo landscapes. Although the distribution and dynamics of these activities vary widely, large areas of páramo are devoted to potato cultivation (Thompson et al., 2021; Sherwood et al., 2023; Vergara and Llorente, 2025), which remains a critical economic activity and one of the main staple crops in the northern Andes. Moreover, between 2001 and 2013, an average of 4,300 ha of páramo burned each year in the Eastern Andes of Colombia alone (Borrelli et al., 2015; Armenteras et al., 2020). Most of these fires are linked to extensive cattle-raising practices, where fire is used to remove necromass accumulated in tussock grasses (e.g., Calamagrostis spp.), thereby facilitating cattle access to more palatable plants. Since both potato cultivation and fires can induce significant changes in the environmental conditions of páramo ecosystems (Peña-Quemba et al., 2016), it is likely that these activities might also alter carbon dynamics in páramo landscapes by affecting processes such as plant productivity, ecosystem respiration, and microbial activity. However, these potential changes have not yet been characterized. In this context, this study presents the results of a experimental assessment of the direct effects of burning and potato cultivation on the carbon dynamics of a grassland ecosystem in the páramos of Northern Ecuador. We hypothesize that burning and potato cultivation would result in increased ecosystem respiration and decreased gross primary productivity, thus reducing the net rate of carbon accumulation in the ecosystem. Ultimately, the aim of this experiment was to characterize potential changes in carbon dynamics that could result from burning and potato cultivation, two of the most common land-uses encroaching in páramo ecosystems in the Northern Andes.

2 Materials & methods

We worked in the Chakana Conservation Reserve (0°26′57.88″”S; 78°18′22.09″”W), in a tussock grassland ecosystem in the Northern Ecuadorian páramo. This area has a long history of grazing and burning dating from at least the 1880s, but cattle was removed around 2014. Since then, vegetation is slowly recovering and is dominated by dense tussock grasses (Calamagrostis intermedia), scattered shrubs (Monticalia peruviana, M. arbutifolia, Baccharis odorata) and herbs including Lachemilla orbiculata, Rockhausenia nubigena, Gentiana sedifolia, among others. The páramo grassland at our study site is representative of the secondary vegetation that develops in the Ecuadorian Andes subjected to cycles of burning and grazing. During the study, mean temperature at the site was 6.7 °C, ranging from −2.2 °C to 16.4 °C, and annual precipitation was 1,230 mm.

To assess the impacts of burning and potato cultivation on carbon dynamics, we established 27 plots (6.25 m² each) distributed in three blocks (Figure 1). The blocks were separated by 170 m–400 m and located between 3,715 and 3,740 m of elevation. Within each block, plots were randomly assigned to one of three treatments (Control, Burning, Potato), each with three replicates. For the Burning treatment, plant biomass within the plots was burned using a propane gas torch between 15 and 22 August 2022. At each plot, the torch flame was applied starting on the top of each tussock, gradually burning the tuft until all the biomass had been consumed and the flames touched the base of the tussock. The procedure was repeated on the next tussock until all the standing biomass in the plot was burned. This burning procedure was selected to avoid spread of the fire beyond the area of the plots, and to simulate the fast-burning fires that usually consume páramo grassland vegetation with restricted immediate impacts on soil temperature (Ramsay and Oxley, 1996; Carrión-Paladines et al., 2023). At the same time, at Potato plots, following common agricultural practice in this area, we first cut and removed the vegetation; then, soil was tilled to a depth of approximately 30 cm and three rows of potato were planted, attaining a density of 18 plants m^-2 (mean plant density at the end of the experiment was 10 plants m^-2). Three weeks after planting, 50 g of inorganic fertilizer (Agripac 10% N, 30% P, 10% K) were applied to the Potato plots. In this way, this treatment reflects the compounded effect of removing the vegetation, tilling the soil, and planting the potato plants. Additionally, plots were fenced with wood boards to a height of 50 cm, to reduce the potential impacts of rabbits and other small herbivores. To monitor temperature patterns, we deployed dual temperature loggers (HOBO U23 Pro v2), distributed in nine plots (one logger per treatment, per block). For each logger, the shielded sensor was located 1.5 m above the ground, while the second was buried at 3 cm depth.

FIGURE 1

Starting on September 2022, CO₂ fluxes were measured in each plot. Sampling frequency was determined by weather conditions, but plots within each block were always sampled within the same day. CO₂ dynamic was characterized in each plot using a static chamber method implemented through a Picarro GasScouter™ G4301, a fixed PVC ring base (40 cm in diameter) enclosing the main type of plant cover in each treatment, and a clear polycarbonate chamber, equipped with a fan. For each measurement, we placed the chamber on the ring base and sealed the resulting seam with a rubber belt. We then placed the lid on the chamber and recorded net flux of CO₂ through the gas scouter for 2 minutes. This first estimate represents net ecosystem exchange (NEE), with negative values representing net CO₂ uptake by the ecosystem, and positive values representing net CO₂ loss to the atmosphere. Following NEE estimation, we removed the lid to refresh the air within the chamber for at least 1 min, then replaced the lid, covered the chamber with a light-proof cloth hood, and repeated the CO₂ flux. As the cloth hood blocks sun light, stopping photosynthesis, CO₂ fluxes become positive within the chamber and represent ecosystem respiration (R_eco). Gross primary productivity (GPP) was calculated as the difference between NEE and R_eco. Each CO₂ measurement was complemented with three consecutive measurements of photosynthetically active radiation (PAR), carried out with a Santacary XAR-P10 Quantum PAR Light Meter. In total, we performed six to eight CO₂ flux measurements in each plot, over a period of 295 days.

Because our field methods did not allow for modeling of annual carbon balance, we calculated the GPP:R_eco ratio, which has been used as a conservative proxy to infer if an ecosystem is likely to be accumulating carbon over time (Amiro et al., 2010; Planas-Clarke et al., 2020). Because our GPP:R_eco ratios didn´t account for nighttime respiration, we used a value of GPP:R_eco = 2 as rough approximation to a 24-h neutral carbon balance (Law et al., 2002; Planas-Clarke et al., 2020). In this approach, if the GPP:R_eco ratio is higher than 2, it is assumed that the ecosystem must be accumulating carbon, whereas lower values would suggest the opposite (Law et al., 2002; Heskel et al., 2013).

In addition to characterizing carbon dynamics, we also estimated the amount of plant biomass carbon lost from the experimental plots. For the Burning treatment, at the onset of the experiment we weighed the aboveground plant biomass that was extracted from the Potato plots before tilling. As the vegetation within each block was very homogeneous, we assumed that the aboveground biomass extracted from the Potato plots, is equivalent to the biomass burned in the Burning plots. For the Potato treatment, at the end of the experiment we cut and weighed all the potato plants and the potatoes excavated from each plot at the end of the experiment. For both treatments, final biomass values were corrected for moisture content established from subsamples of each type of plant material (dried to constant weight at 65 ^oC).

To determine how each treatment influenced the carbon balance we applied linear and generalized mixed effects models. In both cases, we accounted for our randomized block repeated measures design by including the interaction of time and block as random intercepts. To analyze how the treatments influenced NEE, GPP, and R_eco we used linear mixed effects models, with a gaussian distribution and identity link. To determine if the treatments significantly explained differences in the dependent variables, we employed a type II Wald Chi-squared test (Quinn and Keough, 2002), and assessed differences between treatments using Tukey multiple comparisons of general linear hypothesis (Hothorn et al., 2008).

To test the effects of Potato and Burning treatments on the GPP:R_eco ratio we fitted a partially Bayesian binomial mixed effects logistic regression, using GPP:R_eco ratio thresholds (>2 or <2) as a Boolean dependent variable. We used a partially Bayesian model to address singular fit problems arising from maximum marginal likelihood estimations of group level variances (Chung et al., 2013; 2015). “Time” = 0 days and “Treatment” = Control were used as the reference categories. All tests were performed in R with the lme4, blme, car and multcomp packages using RStudio ® environment (Hothorn et al., 2008; Bates et al., 2015; R Core Team, 2022; Posit team, 2024).

3 Results

Potato and Burning treatments resulted in increased daytime soil temperature. On average, mean soil day temperatures at the Potato (95% C.I: 12.0 °C–12.9 °C) and Burning plots (11.6 °C–12.5 °C) were 20% and 16% higher, respectively, than in the Control plots (10.0 °C–10.7 °C). These differences in soil temperature between treatments remained relatively constant until the end of the experiment (Figure 2).

FIGURE 2

There was a significant effect of Treatment on NEE and GPP (Type II Wald Chi-square test; NEE: GPP: ). For both variables, post hoc comparisons indicated no difference between Potato and Burning treatments, but a significant difference when comparing those treatments with the Control plots (Table 1; Figure 3). In contrast, R_eco values were not statistically different between treatments ().

FIGURE 3

Mean values of GPP on the Control plots were 40% and 61% higher than on the Burning and Potato treatments, respectively (Figure 3; Supplementary Table S1). In contrast, rates of ecosystem respiration were similar, with mean values that varied less than 13% between treatments. Resulting patterns of NEE imply that Control plots were mostly fixing carbon, with values ranging from −0.9 to −0.3 g CO₂ m^-2 h^-1 (95% confidence interval), while Potato and Burning plots were mostly losing it, at average rates of 0.4 and 0.2 g CO₂ m^-2 h^-1, respectively (Supplementary Table S1).

As opposed to the Control plots, in which mean NEE, GPP, and R_eco were relatively constant throughout the experiment, these parameters exhibited clear temporal trends in the Potato and Burning plots (Figure 4). In these treatments, plots were predominantly losing carbon (positive NEE values) during the first half of the experiment and then shifted to negative NEE values around day 150 (Figure 4) corresponding to the increase of biomass of growing potato plants and recovering vegetation after fire. Specifically, NEE at the Potato plots decreased from an average of 0.9 g CO₂ m^-2 hr^-1 (95% CI: 1.4 to 0.4 g CO₂ m^-2 hr^-1) during the first 150 days of the experiment, to −0.6 (−0.05 to −1.2 g CO₂ m^-2 hr^-1) during the second half of the study. Corresponding values for the Burning treatment plots represent a decrease from 0.5 (0.9–0.06 g CO₂ m^-2 hr^-1) during the first half of the experiment to −0.4 (0.05 to −0.8 g CO₂ m^-2 hr^-1) in the second half. In contrast to NEE, R_eco exhibited no discernible temporal pattern in the Potato and Burning plots (Figure 4).

FIGURE 4

The GPP:R_eco ratio was significantly different between treatments. Specifically, while this ratio was greater than 2 in 25% of the measurements in the Control plots, corresponding values for the Potato and Burning plots ranged between 6% and 7.5% (Supplementary Table S2). Similarly, at the beginning of the experiment, the odds that a control plot was fixing carbon were 1,212.9 and 147.5 times greater than the corresponding odds for a potato or burning plot respectively (Supplementary Table S2; Supplementary Figure S1). When compared to each other, the odds of carbon fixation in burning plots were 8.22 times higher than in potato plots. Interestingly, the directionality of the effect of these two treatments (potato and burning) changed when their interaction with time was considered (Supplementary Table S2). In other words, at the beginning of the experiment, plots with potato or burning as a treatment were very unlikely to fix carbon but, as time passed, the likelihood of carbon fixation in both treatments gradually increased (Supplementary Figure S1).

In addition to losses due to net carbon dynamics, potato and burning plots experienced additional carbon losses due to carbon exports in the form of the potato cropping and burning of aboveground biomass. Based on field measurements, we estimate that these losses accounted for 830 (380–1,280) gC m^-2 for the potato crop, and 1828 (1,505–2,151) gC m^-2 from the burning of the tussock grasses.

4 Discussion

In this 295-day, plot-scale experiment, burning and potato cultivation in a páramo grassland were associated with substantially higher net carbon losses than in control plots, with mean daytime NEE values 170% and 180% higher, respectively, than in control plots. (Figure 3; Supplementary Table S1). These NEE changes could reflect increased ecosystem respiration driven by changes in environmental conditions, and/or reduced GPP.

Regarding the first possibility, our data shows that, even though mean soil temperatures were significantly higher in the potato and burning treatments (Figure 2), this did not alter ecosystem respiration, unlike reports from previous experiments (Curiel Yuste et al., 2017). In fact, ecosystem respiration remained comparable among treatments throughout the experiment (Figure 4), so the large NEE declines in land-use treatments are more consistent with reduced GPP than with increased R_eco, at least under the soil temperature and moisture conditions encountered during our study. Our results differ from previous studies in the Venezuelan páramo which report declines in soil respiration in potato agricultural plots (Sarmiento and Bottner, 2002; Cabaneiro et al., 2008).

Regarding C outflux resulting from reduced GPP, NEE showed a temporal pattern closely matching aboveground biomass recovery in land-use plots, as reflected by the patterns in GPP (Figure 4). In the burning and potato plots, NEE returned to predominantly negative values after roughly 150 days from treatment onset, even while soil temperatures remained higher than in the control plots. This suggests that the recovery of NEE to values roughly similar to those in the control plots was mostly driven by the growth of potato plants and the partial recovery of vegetation and the resulting increase in GPP.

Our data and previous studies show that tussock grass cover can mitigate the extremes of soil temperature (Scott, 1962; Monteiro et al., 2011). Removing this protection produced a persistent temperature increase that could have boosted microbial activity in the cold páramo environment of our site. Yet, treatments did not measurably affect R_eco. Although we have no specific data to test this possibility, it is possible that, during our experiment, microbial activity might have been limited by low soil moisture availability across all treatments (Wood et al., 2013; Song et al., 2015). This possibility remains largely speculative and needs further analysis, as we do not have information regarding other factors that could affect ecosystem respiration, including the quality of microbial substrates, and the partition between autotrophic and heterotrophic respiration.

Regarding the persistence of burning and cultivation impacts on carbon dynamics, our results suggest that the short-term effects of these activities were substantial, but transitory. Approximately 150 days after treatments onset, NEE in Burning and Potato plots approached levels comparable to those in Control plots, likely because NEE changes were driven mainly by reduced GPP from vegetation removal, rather than altered ecosystem respiration. Different scenarios, including the trampling and grazing that frequently follow páramo fires, or alternative cropping systems, might result in different temporal patterns or more persistent impacts on carbon dynamics. This is especially true for the burning treatment because prescribed fires in the páramo are normally associated to extensive cattle raising and this land use was not simulated in our experiment. Trampling (soil compactation), browsing, and fertilization associated with grazing might exacerbate or prolong the impacts of this activity on carbon dynamics (Clay and Worrall, 2013; Oliver et al., 2017).

To the extent of our knowledge, no comparable studies have been carried out in other tropical mountains. However, our experiment in a páramo grassland in northern Ecuador stands in contrast to patterns reported from the European Alps. In general, studies in temperate mountains show that heavily managed grasslands tend to increase rates of NEE, GPP, and R_eco, especially when soil fertility is increased by application of manure or other forms of fertilizer (Schmitt et al., 2010; Zeeman et al., 2010). In contrast, the land-use treatments at our sites resulted in declines in carbon assimilation. This pattern suggests that the observed differences may stem from the specific land use activities that we tested in the páramo, which involve a complete but temporary removal of vegetation cover (reduced GPP), whereas pasture fertilization in the Alps typically enhances GPP, leading to greater carbon storage.

At the plot scale and over 295 days, the magnitude of the NEE response to burning and potato cultivation was similar in our experiment, but we did not evaluate cumulative effects of repeated cropping, repeated burning, or their interactions at the landscape scale. Potato cultivation may affect relatively small, lower-elevation areas, whereas individual fires can impact larger areas and broader elevation ranges. Moreover, potato cultivation can become a persistent land use, while burning at a site may occur at longer intervals. Also, while in our study area potato cultivation is mostly artisanal, in other páramo regions (i.e., the Nariño-Carchi on the Colombia-Ecuador border or the Cundinamarca-Boyacá altiplano in Colombia) potato cultivation is heavily mechanized and affecting larger areas. From this point of view, understanding landscape-scale impacts will require additional modeling of the extent and distribution of these practices, and integration with remote sensing, long-term flux measurements, and soil carbon inventories under contrasting environmental conditions.

The GPP:R_eco ratio has been used as a rough indicator of the capacity of an ecosystem to act as a net sink of carbon, with values greater than 2 suggesting that more carbon is being fixed than released, over time. This approach is imperfect as is does not incorporate direct measurements of nighttime respiration and has not been evaluated in páramo ecosystems. However, the magnitude of the differences in this ratio between our treatments is informative. Not surprisingly, we found that the vast majority (>90%) of measurements at Potato and Burning plots had GPP:R_eco values lower than 2, suggesting that plots under these land-use activities act as a net source of carbon most of the time (Supplementary Figure S1). Interestingly, although the percentage of GPP:R_eco values greater than 2 at the Control plots (25%) was five times higher than at the Potato and Burning treatments, this value is still relatively low. This pattern suggests that, under the conditions sampled, our site functioned as a net CO₂ source during much of the daytime measurement period. Combined with Eddy-covariance evidence from a southern Ecuadorian páramo that also indicates net carbon loss over 2 years (Carrillo-Rojas et al., 2019), this raises the possibility that some grazed-and-burnt páramo grasslands may currently be transitioning towards net carbon loss, with implications for climate mitigation and hydrological services. The drivers of these changes might be manifold and need to be further analyzed. On the one hand, increased CO₂ outflux could result from augmented microbial activity associated with higher soil temperatures. However, this explanation seems unlikely for our dataset as the Potato and Burning treatments, in which soil temperatures were significantly higher than on the Control plots, did not experience higher rates of R_eco. On the other hand, higher mean temperatures in this region could be leading to increased evapotranspiration, thus resulting in reduced NPP and a net loss of carbon from the ecosystem. However, further research is needed to ascertain these patterns and the mechanisms that could explain them, under a wider range of environmental conditions and especially regarding the role of soil moisture in controlling ecosystem respiration and GPP.

Regarding páramo fires, it has been suggested that páramo might be a fire-dependent ecosystem (Obando-Cabrera et al., 2025). Our experiment suggests that a single low-severity burn can temporarily increase net carbon losses by ∼180% relative to unburned plots at the plot scale. If such short-term flux responses occur repeatedly at intervals of only a few years, they could, in principle, hinder long-term soil carbon accumulation. Together with evidence that fire increases water repellency and runoff in páramo soils (Poulenard et al., 2001), these findings suggest that frequent anthropogenic burning might be incompatible with the long-term accumulation and maintenance of large soil carbon stocks and hydrological functions that characterize many páramo ecosystems. Testing this hypothesis will require integrating plot-scale flux measurements, fire frequency data, and soil carbon stock measurements across representative landscapes.

5 Conclusion

Our study highlights the significant impacts of burning and potato cultivation on carbon dynamics in páramo grassland ecosystems of Northern Ecuador. Both land use activities resulted in transient but substantial carbon losses, with carbon outfluxes from Potato and Burning plots being 170% and 180% higher, respectively, compared to Control plots. These changes were attributed to reduced GPP rather than increased R_eco, as soil temperature increases did not result on elevated R_eco rates. Instead, the recovery of aboveground biomass in land use treatment plots corresponded with a return to negative NEE values, highlighting the importance of vegetation cover for carbon assimilation. While our study provides insights into the localized impacts of burning and agriculture, further research is needed to understand their landscape-scale effects. Additionally, ongoing investigations into the mechanisms driving carbon dynamics in páramo ecosystems, particularly the role of soil moisture in regulating GPP and R_eco, are crucial for effective management and conservation strategies. Overall, our results suggest that anthropogenic fires in páramo grasslands cause substantial short-term carbon losses at the plot scale. At the same time, fire has been shown to contribute to soil drying and water repellency (Poulenard et al., 2001). Given these effects, and the relatively limited production benefits of fire-based extensive grazing systems in many páramo regions, reducing the frequency and spatial extent of anthropogenic burning should be explored as a priority for conserving soil carbon and hydrological services.

References

• 1

  AmiroB. D.BarrA. G.BarrJ. G.BlackT. A.BrachoR.BrownM.et al (2010). Ecosystem carbon dioxide fluxes after disturbance in forests of North America. J. Geophys. Res. Biogeosciences115. 10.1029/2010JG001390

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  ArmenterasD.GonzálezT. M.Vargas RiosO.Meza ElizaldeM. C.OliverasI. (2020). Incendios en ecosistemas del norte de Suramérica: avances en la ecología del fuego tropical en Colombia, Ecuador y Perú. Caldasia42, 1–16. 10.15446/caldasia.v42n1.77353

  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  BatesD.MächlerM.BolkerB.WalkerS. (2015). Fitting linear mixed-effects models using lme4. J. Stat. Softw.67, 1–48. 10.18637/jss.v067.i01

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BerningerF.SusiluotoS.GianelleD.BahnM.WohlfahrtG.SuttonM.et al (2015). Management and site effects on carbon balances of European mountain meadows and rangelands. Boreal Environ. Res.20, 748–760.

  • Google Scholar
  • View reference in article

• 5

  BorrelliP.ArmenterasD.PanagosP.ModugnoS.SchüttB. (2015). The implications of fire management in the andean Paramo: a preliminary assessment using satellite remote sensing. Remote Sens.7, 11061–11082. 10.3390/rs70911061

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BuytaertW.DeckersJ.DerconG.De BieÁvreB.PoesenJ.GoversG. (2002). Impact of land use changes on the hydrological properties of volcanic ash soils in South Ecuador. Soil Use Manag.18, 94–100. 10.1111/j.1475-2743.2002.tb00226.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BuytaertW.CelleriR.de BievreB.CisnerosF.WyseureG.DeckersJ.et al (2006). Human impact on the hydrology of the Andean paramos. Earth-Science Rev.79, 53–72. 10.1016/j.earscirev.2006.06.002

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BuytaertW.Cuesta-CamachoF.TobonC. (2011). Potential impacts of climate change on the environmental services of humid tropical alpine regions. Glob. Ecol. Biogeogr.20, 19–33. 10.1111/j.1466-8238.2010.00585.x

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  CabaneiroA.FernandezI.Parez-venturaL.CarballasT. (2008). Soil CO2 emissions from northern Andean paramo ecosystems: effects of fallow agriculture. Environ. Sci. and Technol.42, 1408–1415. 10.1021/es071392d

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  CalispaM.VásconezF. J.SantamaríaS.SamaniegoP. (2023). “Los suelos de los páramos del Ecuador,” in Los Páramos En El Ecuador: Pasado, Presente Y Futuro (Quito, Ecuador: USFQ Press), 40–74.

  • Google Scholar
  • View reference in article

• 11

  Carrillo-RojasG.SilvaB.RollenbeckaR.CélleriR.BendixaJ. (2019). The breathing of the andean highlands: net ecosystem exchange and evapotranspiration over the páramo of southern Ecuador. Agric. For. Meteorology265, 30–47. 10.1016/j.agrformet.2018.11.006

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  Carrión-PaladinesV.FriesA.HinojosaM. B.OñaA.ÁlvarezL. J.BenítezÁ.et al (2023). Effects of low-severity fire on soil physico-chemical properties in an Andean páramo in Southern Ecuador. Fire6, 447. 10.3390/fire6120447

  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  ChungY.Rabe-HeskethS.DorieV.GelmanA.LiuJ. (2013). A nondegenerate penalized likelihood estimator for variance parameters in multilevel models. Psychometrika78, 685–709. 10.1007/s11336-013-9328-2

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  ChungY.GelmanA.Rabe-HeskethS.LiuJ.DorieV. (2015). Weakly informative prior for point estimation of covariance matrices in hierarchical models. J. Educ. Behav. Statistics40, 136–157. 10.3102/1076998615570945

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  ClayG. D.WorrallF. (2013). The response of CO2 fluxes from a peat soil to variation in simulated sheep trampling. Geoderma197–198, 59–66. 10.1016/j.geoderma.2012.12.008

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  CressoM.ClericiN.SanchezA.JaramilloF. (2020). Future climate change renders unsuitable conditions for paramo ecosystems in Colombia. Sustainability12, 8373. 10.3390/su12208373

  • CrossRef
  • Google Scholar
  • View reference in article

• 17

  CuestaF.BáezS.RamírezJ.TovarC.DevenishC.BuytaertW.et al (2012). “Síntesis de los impactos y estado del conocimiento de los efectos del cambio climático en la biodiversidad de los Andes Tropicales,” in Panorama Andino De Cambio Climático: Vulnerabilidad Y Adaptación En Los Andes Tropicales. Editors CuestaF.BustamanteM.BecerraM. T.PostigoJ.PeralvoM. (Lima-Quito: CONDESAN, SGCAN), 109–146.

  • Google Scholar
  • View reference in article

• 18

  Curiel YusteJ.HereşA.-M.OjedaG.PazA.PizanoC.García-AnguloD.et al (2017). Soil heterotrophic CO2 emissions from tropical high-elevation ecosystems (páramos) and their sensitivity to temperature and moisture fluctuations. Soil Biol. Biochem.110, 8–11. 10.1016/j.soilbio.2017.02.016

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  DuchicelaS. A.LlambíL. D.BonnesoeurV.Román-DañobeytiaF. (2024). Pastoralism in the high tropical Andes: a review of the effect of grazing intensity on plant diversity and ecosystem services. Appl. Veg. Sci.27, e12791. 10.1111/avsc.12791

  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  Grêt-RegameyA.WeibelB. (2020). Global assessment of Mountain ecosystem services using earth observation data. Ecosyst. Serv.46, 101213. 10.1016/j.ecoser.2020.101213

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  HeskelM. A.AtkinO. K.TurnbullM. H.GriffinK. L. (2013). Bringing the kok effect to light: a review on the integration of daytime respiration and net ecosystem exchange. Ecosphere4, 98. 10.1890/ES13-00120.1

  • CrossRef
  • Google Scholar
  • View reference in article

• 22

  HothornT.BretzF.WestfallP. (2008). Simultaneous inference in general parametric models. Biometrical J.50, 346–363. 10.1002/bimj.200810425

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 23

  LassoE.Matheus-ArbelaezP.GalleryR. E.Garzon-LopezC.CruzM.Leon-GarciaI. V.et al (2021). Homeostatic response to three years of experimental warming suggests high intrinsic natural resistance in the paramos to warming in the short term. Front. Ecol. Evol.9, 615006. 10.3389/fevo.2021.615006

  • CrossRef
  • Google Scholar
  • View reference in article

• 24

  LawB. E.FalgeE.GuL.BaldocchiD. D.BakwinP.BerbigierP.et al (2002). Environmental controls over carbon dioxide and water vapor exchange of terrestrial vegetation. Agric. For. Meteorology113, 97–120. 10.1016/S0168-1923(02)00104-1

  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  MatsonE.BartD. (2013). Interactions among fire legacies, grazing and topography predict shrub encroachment in post-agricultural paramo. Landsc. Ecol.28, 1829–1840. 10.1007/s10980-013-9926-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  MonteiroJ. A. F.HiltbrunnerE.KörnerC. (2011). Functional morphology and microclimate of Festuca orthophylla, the dominant tall tussock grass in the andean altiplano. Flora - Morphol. Distribution, Funct. Ecol. Plants206, 387–396. 10.1016/j.flora.2011.01.001

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  Obando-CabreraL.Díaz-TimotéJ. J.BastarrikaA.CelisN.HantsonS. (2025). The paramo Fire atlas: quantifying burned area and trends across the tropical andes. Environ. Res. Lett.20, 054019. 10.1088/1748-9326/adc8ba

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  OliverV.OliverasI.KalaJ.LeverR.TehY. A. (2017). The effects of burning and grazing on soil carbon dynamics in managed Peruvian tropical montane grasslands. Biogeosciences14, 5633–5646. 10.5194/bg-14-5633-2017

  • CrossRef
  • Google Scholar
  • View reference in article

• 29

  Patiño-GutiérrezS. E.Domínguez-RiveraI. C.Daza-TorrezM. C.Ochoa-TocachiB. F.Oviedo-OcañaE. R. (2024). Effects of rainfall seasonality and land use change on soil hydrophysical properties of high-andean dry páramo grasslands. Catena238, 107866. 10.1016/j.catena.2024.107866

  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  Peña-QuembaD.Rubiano-SanabriaY.Riveros-IreguiD. (2016). Effects of land use on soil CO2 flux in the Paramo de Guerrero, Colombia. Agron. Colomb.34, 364–373. 10.15446/agron.colomb.v34n3.58791

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  PeyreG.OsorioD.RománF.AnthelmeF. (2021). Mapping the páramo land-cover in the northern Andes. Int. J. Remote Sens.42, 7777–7797. 10.1080/01431161.2021.1964709

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  Planas-ClarkeA. M.ChimnerR. A.HribljanJ. A.LilleskovE. A.FuentealbaB. (2020). The effect of water table levels and short-term ditch restoration on Mountain peatland carbon cycling in the Cordillera Blanca, Peru. Wetl. Ecol. Manag.28, 51–69. 10.1007/s11273-019-09694-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  PodwojewskiP.PoulenardJ.ZambranoT.HofstedeR. G. M. (2002). Overgrazing effects on vegetation cover and properties of volcanic ash soil in the páramo of Llangahua and La Esperanza (Tungurahua, Ecuador). Soil Use Manag.18, 45–55. 10.1079/sum2002100

  • CrossRef
  • Google Scholar
  • View reference in article

• 34

  Posit team (2024). Rstudio: Integrated Development Environment for R. Available online at: http://www.posit.co/ (Accessed November 15, 2025).

  • Google Scholar
  • View reference in article

• 35

  PoulenardJ.PodwojewskiP.JaneauJ. L.CollinetJ. (2001). Runoff and soil erosion under rainfall simulation of andisols from the Ecuadorian páramo: effect of tillage and burning. Catena45, 185–207. 10.1016/s0341-8162(01)00148-5

  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  PriceM.EganP. A. (2017). Mountain Ecosystem Services and Climate Change: A Global Overview of Potential Threats and Strategies for Adaptation. (Paris: UNESCO).

  • Google Scholar
  • View reference in article

• 37

  QuinnG. P.KeoughM. J. (2002). Experimental Design and Data Analysis for Biologists. Cambridge: Cambridge University Press. 10.1017/CBO9780511806384

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  R Core Team (2022). R: A Language and Environment for Statistical Computing. Available online at: https://www.R-project.org/ (Accessed November 15, 2025).

  • Google Scholar
  • View reference in article

• 39

  RamsayP. M.OxleyE. R. B. (1996). Fire temperatures and postfire plant community dynamics in Ecuadorian grass páramo. Vegetatio124, 129–144. 10.1007/bf00045489

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  RawlsW. J.PachepskyY. A.RitchieJ. C.SobeckiT. M.BloodworthH. (2003). Effect of soil organic carbon on soil water retention. Geoderma116, 61–76. 10.1016/S0016-7061(03)00094-6

  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  Rey-RomeroD. C.DomínguezI.Oviedo-OcañaE. R. (2022). Effect of agricultural activities on surface water quality from páramo ecosystems. Environ. Sci. Pollut. Res.29, 83169–83190. 10.1007/s11356-022-21709-6

  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  SarmientoL.BottnerP. (2002). Carbon and nitrogen dynamics in two soils with different fallow times in the high tropical andes: indications for fertility restoration. Appl. Soil Ecol.19, 79–89. 10.1016/s0929-1393(01)00178-0

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  SchickhoffU.BobrowskiM.MalS.SchwabN.SinghR. B. (2022). “The world’s Mountains in the anthropocene,” in Mountain Landscapes in Transition: Effects of Land Use and Climate Change. Editors SchickhoffU.SinghR. B.MalS. (Cham: Springer International Publishing), 1–144. 10.1007/978-3-030-70238-0_1

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  SchmittM.BahnM.WohlfahrtG.TappeinerU.CernuscaA. (2010). Land use affects the net ecosystem CO₂ exchange and its components in Mountain grasslands. Biogeosciences7, 2297–2309. 10.5194/bg-7-2297-2010

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  ScottD. (1962). Temperature and light micro-climate within a tall tussock community. N. Z. J. Agric. Res.5, 179–182. 10.1080/00288233.1962.10419988

  • CrossRef
  • Google Scholar
  • View reference in article

• 46

  SherwoodS.ParedesM.OyarzúnP.BorjaR. (2023). “Agricultura y páramos en el centro-norte del Ecuador: cultivo de las tierras altas, amenazas ambientales y oportunidades para el futuro,” in Los Páramos Del Ecuador: Pasado, Presente Y Futuro. Editors HofstedeR.Mena-VásconezP.SuárezE. (Quito, Ecuador: USFQ Press), 246–281.

  • Google Scholar
  • View reference in article

• 47

  SongW.ChenS.WuB.ZhuY.ZhouY.LuQ.et al (2015). Simulated rain addition modifies diurnal patterns and temperature sensitivities of autotrophic and heterotrophic soil respiration in an arid desert ecosystem. Soil Biol. Biochem.82, 143–152. 10.1016/j.soilbio.2014.12.020

  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  ThompsonJ. B.Zurita-ArthosL.MüllerF.ChimbolemaS.SuárezE. (2021). Land use change in the Ecuadorian páramo: the impact of expanding agriculture on soil carbon storage. Arct. Antarct. Alp. Res.53, 48–59. 10.1080/15230430.2021.1873055

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  VergaraP.LlorenteI. D. P. (2025). Multitemporal analysis of land cover changes in the rabanal páramo ecosystem (colombia) from 2000 to 2020 and its planning regulations. Land Degrad. Dev.36, 2247–2258. 10.1002/ldr.5493

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  WoodT. E.DettoM.SilverW. L. (2013). Sensitivity of soil respiration to variability in soil moisture and temperature in a humid tropical Forest. PLoS ONE8, e80965. 10.1371/journal.pone.0080965

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 51

  ZeemanM. J.HillerR.GilgenA. K.MichnaP.PlüssP.BuchmannN.et al (2010). Management and climate impacts on net CO2 fluxes and carbon budgets of three grasslands along an elevational gradient in Switzerland. Agric. For. Meteorology150, 519–530. 10.1016/j.agrformet.2010.01.011

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  ZomerM. A.RamsayP. M. (2021). Post-fire changes in plant growth form composition and diversity in andean paramo grassland. Appl. Veg. Sci.24, e12554. 10.1111/avsc.12554

  • CrossRef
  • Google Scholar
  • View reference in article