Current state of knowledge

The páramos biome

The páramos form a tropical alpine grassland biome located mostly in the Northern Andes (between 11°N and 8°S) and parts of Central America (Figure 1), above the tropical mountain forest biomes (⁓3000 m a.s.l.) and below the cryosphere (⁓5000 m a.s.l.) (Josse et al., 2009; Luteyn, 1999). Their delineation is not always clear because of deforestation and increasing encroachment of the lower páramos for agricultural purposes (López Sandoval and Valdez, 2015; Tovar et al., 2013a) and ongoing influence of global-warming that induced glacier retreat (Morueta-Holme et al., 2015). R. G. M. Hofstede et al., (2003) estimated that the páramos occupy an area of around 35000 km2 of the tropical Andes. Geographically, the largest extent of páramos occurs in Colombia and Ecuador and smaller, disconnected patches exist in the Andes of Venezuela (Páramos de Mérida) and Costa Rica (Cerro Chirripó). The southern limit of the páramos is known as the Jalca grasslands (Sánchez-Vega and Dillon, 2006; Tovar et al., 2012), a transition vegetation to the Puna biome of the central and south Peruvian Andes (Cuesta et al., 2017; Ochoa-Tocachi et al., 2016a). However, ecosystems with very similar biogeographical and hydrometeorological characteristics occur as far south as Bolivia, where they are known as the Páramo Yungeño (Jørgensen et al., 2014). The páramos are highly fragmented because of the complex topography of the Northern Andes and underwent in the past changes in isolation and connectivity (Flantua and Hooghiemstra, 2018). Glacial and interglacial periods raised and lowered respectively the upper forest line leading to changes in connectivity that have directly impacted the flora of the region (Flantua and Hooghiemstra, 2018; van der Hammen, 1982; Hooghiemstra and Van der Hammen, 2004). These changes are one of the main drivers of the high levels of endemism and species diversification making the páramos the fastest evolved biome among biodiversity hotspots (Madriñán et al., 2013). The vegetation of the páramos (Figure 2) is dominated by tussock grasses (Calamagrostis, Stipa and Festuca sp.) with scarce forest patches (e.g., Polylepis sp.) with transitions to acaulescent rosettes (e.g., Werneria nubigena, Hypochaeris sessiliflora) and cushions plants (e.g., Azorella sp., Plantago rigida) at higher elevations (Luteyn, 1999; Ramsay and Oxley, 1997). The total number of plant species recorded in the páramos is 3595 distributed among 540 genera of which 14 are endemic (Sklenar and Balslev, 2005). The páramos have the highest number of plant species among tropical alpine flora (Sklenář et al., 2014) and are fundamental part of the habitats of emblematic species such as the Andean condor (Vultur gryphus) and the spectacled bear (Tremarctus ornatus).

The hydrology and meteorology of páramos

Precipitation at the páramo is known for its remarkably spatial-temporal variability (Buytaert et al., 2006d; Celleri et al., 2007) caused by the interaction between various synoptic climate processes and the complex topography. The mean annual precipitation between 2000 and 2014 calculated from the global precipitation product CHIRPS (Climate Hazards Group InfraRed Precipitation with Station Data) varies between 150 and 4090 mm yr-1 (Figure 3A) with mean maximum values reported in Costa Rica and minimum in Perú (Table 1). They are usually referred to as “wet” ecosystems (Buytaert et al., 2006a; Padrón et al., 2015)even though rainfall can be well below 500 mm yr-1 in some of the drier páramos, such as those of Chimborazo in Ecuador (Clapperton, 1990; Saberi et al., 2018). However, rainfall regimes between 1000 and 2000 mm yr-1 are more common, with values exceeding 3000 mm yr-1 in some páramos on the Amazonian slopes of the Andes. The analysis of rainfall extremes, evidenced that the southern component of tropical airflow is important for the distribution of wet convection, leading to a high intensity of precipitation in the Andean mountains. The mountains at the other hand, dampen the airflow on a large-scale, enabling local hygrothermal gradients to control extreme precipitation anomalies (Pineda and Willems, 2018). Compared to rainfall, the temperature range of the páramos is much more homogeneous, since this is one of the distinctive characteristics of their biogeographical niche (Tovar et al., 2013a). Average temperatures range around 10ºC in the sub-páramos to close to 0º on the upper fringes of the super-páramos, which borders the cryosphere (Buytaert et al., 2006b). The low latitude of páramos limits seasonal temperature variability, which is predominantly induced by the seasonality of precipitation and related cloud cover. Contrastingly, day-night temperature variations can be extreme and are a direct result of the high altitude and Equatorial position, which gives the páramos among the world’s highest influx of shortwave radiation (Buytaert et al., 2006b). From the actual evapotranspiration (ETa), 51% of the total annual rainfall according to Carrillo-Rojas et al. (2019), only the component of interception loss can be considered important at the páramos. Despite the predominance of low-intensity rainfall and frequent fog, the vegetation can lead to exceptionally high interception rates. For instance, Ochoa‐Sánchez et al. (2018) reported interception values by tussock grass in páramos between 10% to 100% of total rainfall, with a maximum storage capacity of 2 mm. Cárdenas et al. (2017) estimated that inputs from fog and drizzle can represent between 7% and 28% (120 and 212 mm yr−1) of the rainfall in Colombian páramos. In areas covered by fog, horizontal precipitation, and cloud water, interception amounts between 10% to 35% of the total precipitation have been reported for forested catchments (Bruijnzeel, 2004; Pryet et al., 2012), although Bonnesoeur et al. (2019) indicate that values between 2% and 8% may be more common. Actual evapotranspiration (ETa) shows values of 646 mm yr−1 (Buytaert et al., 2007a) and 723 mm yr−1 (Córdova et al., 2015). However, when data are limited errors are thought to be as high as 30% (Carrillo-Rojas et al., 2016; Córdova et al., 2015). Carrillo-Rojas et al. (2019) improved evapotranspiration estimates for páramos using Eddy-covariance, reporting a value of 635±9 mm yr−1for their study site in southern Ecuador. The same authors reported water deficit during specific periods and highlighted the role of air humidity variation in the control of the hydrological system. Furthermore, Wouter Buytaert & Beven, (2011) emphasized the importance of non-stationary hydrological processes such as changing evapotranspiration, infiltration and routing due to vegetation growth. In Figure 3B to 3D, indexes derived from TRMM-3b43 (Tropical Rainfall Measuring Mission) and MODIS (Moderate Resolution Imaging Spectroradiometer) were presented for the study region (Arciniega-Esparza et al., 2020 (in prep)). For the Seasonal Precipitation Index (SPI), high values represent a longer dry season compared to a rainy season and low values the absence of precipitation seasonality. The southern páramos therefore present a prolonged dry season, with the highest indexes of evaporation (EI) and aridity (AL), 50% and 100% respectively, higher than the subsequent ones (Table 1). Mid-ranges of EI were observed for most of the region with minimums in Ecuador and Colombia. In addition, low tendencies of aridity (AI) in the central and northern páramos can be observed. The high altitudinal position of the páramos compared to the elevation of most human settlements and cities makes them convenient “natural water towers” (Messerli et al., 2004) from which water can be sourced by gravity. The hydrological response of páramos is strongly related to their soil conditions and extent of wetlands (Mosquera et al., 2015). W. Buytaert et al. (2005) revealed that the probability of water stress occurrence in wet paramo soils is reduced due to their hydraulic conductivity, which prevents soil moisture to drop below 60 vol%. Páramos’ soils are therefore important regulators of runoff production (Harden, 2006). The most organic-rich soils primarily located at the foot of the hillslopes and in the bottom of the valley are commonly covered by cushion plants and present nearly saturated conditions (Buytaert et al., 2005) while the more freely draining soils are situated on the hillslopes under a cover of tussock grass. The rainfall-runoff response is mainly controlled by the variable extent of the saturated zone in the valley bottom (Correa et al., 2017). Water from valley soils compose around 40% of the runoff in a headwater catchment in the Ecuadorian páramos. During rainier periods, the contributing area expands, thus increasing the connectivity with lateral flow from hillslopes and therefore its contribution to the channel network. In drier conditions, only the deep soil horizons from the hillslopes seems to be hydrologically connected (Correa et al., 2017) and water from these horizons drains via the riparian area into streams (Crespo et al., 2011). A particular characteristic of most of the studied high Andean catchments is the presence of underlying impermeable bedrock that minimizes deep infiltration and greatly limit the groundwater contributions (Buytaert et al., 2007a). Nevertheless, some regions also present deep permeable soils, sustain important aquifers (Buytaert et al., 2006a; Favier et al., 2008) and proof the influence of shallow groundwater on stream generation especially when soil moisture decreases (Correa et al., 2017, 2018; Favier et al., 2008). Runoff ratios (ratio between annual precipitation and annual discharge) between 0.50 and 0.70 have been reported in natural wet páramos (Buytaert et al., 2007a; Ochoa-Tocachi et al., 2016a) reaching even values of 0.9 during rainfall-runoff events (Correa et al., 2016). The water yield increases with the amount of wetland, likely because of saturation excess flow (Mosquera et al., 2015). Additionally, Buytaert & Beven (2011) highlighted the importance of threshold-triggered hydrological processes, such as disconnected water storages within the catchment microtopography. In addition, threshold-driven hydrological processes, such as disconnected water storage within the microtopography of the catchment, play a crucial role in the runoff generation and catchment response (Buytaert and Beven, 2011). Figure 3 Table 1