Giant rosette plant morphology as an indicator of recent fire history in Andean páramo grasslands

Ecological Indicators 45 (2014) 37–44

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Giant rosette plant morphology as an indicator of recent fire history in Andean páramo grasslands Paul M. Ramsay ∗ School of Biological Sciences, Plymouth University, PL4 8AA, United Kingdom

a r t i c l e

i n f o

Article history: Received 5 December 2013 Received in revised form 6 February 2014 Accepted 6 March 2014 Keywords: Burning Espeletia Fire mortality Growth rates Ecuador Colombia

a b s t r a c t High-altitude Andean páramo grasslands are fire-dependent systems but reconstructing recent fire history is difficult using conventional approaches. From Venezuela to Ecuador, páramos are usually dominated by giant rosette plants of the genus Espeletia. This study assesses Espeletia’s potential as an indicator of recent fire history. Their peculiar morphology is an adaptation to the mountain environment. Fire removes dead leaves which sheathe the single stem, but they begin to reaccumulate after the fire. It is this reaccumulation of leaves, plus post-fire mortality rates, that might indicate recent fire history. Adult mortality during the first two years after the fire varied according to fire intensity, from 8% (low intensity) to 56% (very high intensity), and was low in the absence of fire (2.5%). Growth rates were much faster at 3600 m (14.8 cm y−1 ) than at 4100 m (1.6 cm y−1 ), and so was leaf turnover (94 compared with 50 leaves y−1 ). Taller plants grew faster than shorter ones. Dead leaf cover on the stems successfully predicted time since fire in four sites of known fire age. Espeletia does represent a useful indicator of fire history but requires calibration to account for local growth rates. At lower altitudes, Espeletia plants could provide information in fires during the previous 20 y, and longer periods at higher altitudes (where plants grow more slowly and live longer). It is a relatively cheap method that could be used to support a wide range of wider studies where recent fire history is influential. A protocol for calibrating the use of Espeletia as an indicator is proposed. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction High-altitude páramo grasslands form a chain along the Andes Mountains from Venezuela and Colombia, through Ecuador, to northern Perú and occupy a total of around 35,000 km2 (Hofstede et al., 2003). Páramos also exist in the mountains of Costa Rica (Kappelle and Horn, 2005). For mountain grassland, Andean páramo is relatively biodiverse, with around 5000 plant species or 10–20% of Andean floral richness (Rangel Ch, 2000). Furthermore, as much as 60% of páramo plant species are endemic to this ecosystem (Luteyn, 1999). The páramos provide water for people, agriculture and industry at lower altitudes and the soils are rich in organic material, storing large amounts of carbon (Hofstede et al., 2003). These páramos are fire-dependent systems (Horn and Kappelle, 2009). Mounting evidence suggests that extensive grasslands and fires did exist before the arrival of humans (Di Pasquale et al., 2008), but that a significant increase in fire frequency after their arrival

∗ Tel.: +44 1752 584646; fax: +44 1752 584605. E-mail address: [email protected] http://dx.doi.org/10.1016/j.ecolind.2014.03.003 1470-160X/© 2014 Elsevier Ltd. All rights reserved.

has been responsible for an increase in the extent of páramos at the expense of forest cover, especially at lower altitudes (Laegaard, 1992; White, 2013). What is absolutely clear is that local fire regimes determine the detail of biodiversity composition and ecological dynamics in modern páramo grasslands. Differential fire mortalities of growth forms, enhanced recruitment of some species after fires, and fire-induced changes to species interactions like competition for light have all been demonstrated for páramo plants within landscape-scale fire mosaics in Ecuador (Laegaard, 1992; Ramsay, 1999; Ramsay and Oxley, 1996, 1997; Sklenáˇr and Ramsay, 2001), with some evidence of consequences on populations of páramo animals (Suárez and Toral, 1996). In addition, páramo fires have been shown to affect soils and hydrology (Buytaert et al., 2006; Harden, 2006; Poulenard et al., 2001) and agricultural production (Hofstede, 1995; Ramsay and Oxley, 2001). Therefore, any plan aimed at protecting biodiversity and managing natural resources in the páramo must take account of fires. Normally, management plans for protected páramo areas prohibit fires because they are assumed, anecdotally, to be damaging to páramo soils, hydrology and biodiversity. Such dictats are often not enforced and, even where they are, can result in significant fires

38

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

when they do occur (Keating, 2007). They also ignore the role fire might play in maintaining the ecological context that many such plans aim to protect. Understanding the nature of the local fire regime is key to understanding the impact of burning on biodiversity, ecosystem services and landscape. Unfortunately, the complexity of fire regimes makes them difficult to understand. Páramo fire frequency and intensity change according to fuel availability, wind direction, slope, vegetation composition and structure, climatic conditions and human intervention—resulting in complex landscape mosaics (Horn and Kappelle, 2009; Keating, 2007; Ramsay, 1999). The unpredictability of wildfire events means that most studies of the effects of fire are carried out retrospectively, once a fire has occurred, and comparisons with pre-fire conditions are impossible. Only a few manipulative studies using experimental or prescribed fire in the páramo have been carried out (Keating, 1998; Ramsay and Oxley, 1996), too few to understand the effects of the diverse combination of fire events that characterize the páramos. The time needed for recovery of páramo vegetation, soils and hydrology after fire is an essential component of any effective management plan. One way to determine recovery time is to monitor plots through time after a fire until recovery has been achieved, but such studies take time and involve long-term commitment of resources to accomplish—none have been published to date for páramo ecosystems. Alternatively, ad hoc post-fire investigations can provide valuable information about recovery after fire, but often fail because detailed information about fire history in the páramo zone is lacking (Horn and Kappelle, 2009). Remoteness and rugged topography mean observational records are limited. In other fire-dominated systems, satellite imagery and/or air photography can provide useful historical insights into fire regimes. Unfortunately, significant cloud cover makes this difficult for páramo studies. Some means to determine time since fire at a landscape scale would be extremely valuable in promoting the understanding of fire dynamics and ecosystem response in the páramo (Horn and Kappelle, 2009). Ecological indicators offer one possible source of information. The absence of some fire-sensitive species may offer crude indicators of fire prevalence (e.g., Bromley, 1971), but do not provide enough information for detailed investigations. In Costa Rican páramos, the annual rings of some shrubs allow past fires to be dated (Williamson et al., 1986), but this approach does not transfer to more equatorial páramos where shrubs do not have annual growth rings. In Ecuador, patches of same-age Puya giant rosette plants have been proposed to estimate time since fire: pulses of recruitment occur soon after fires and the steady growth rate of the rosettes allow the time of recruitment of the cohort to be estimated up to the 28–30 y lifespan of these plants (García-Meneses and Ramsay, 2014). However, one drawback of this method is that the recruitment pulse may take place within a window of several years, resulting in a similar margin of error of that length. Another giant rosette offers the prospect of estimating time since fire with more precision. Espeletia plants are characteristic of páramos from Venezuela to the north of Ecuador (Luteyn, 1999). They have a peculiar morphology, shared with several other tropical mountain plants around the world where the growth form has evolved independently (Smith, 1994). The typical morphology of an Espeletia plant is depicted in Fig. 1: a rosette of leaves sits on a tall stem which retains its dead leaves (known as marcescence). The morphology is a response to the particular demands of the highaltitude tropical climate (Carlquist, 1994; Goldstein and Meinzer, 1983). The important point here is that a central reservoir in the tall stem, insulated by the dead leaves, provides a supply of unfrozen water to the leaves in the early morning when the soil water is potentially still too cold to be of use. The marcescent leaves are retained throughout the life of the plant, with the aid of phenolic

Fig. 1. Morphological components of a burned Espeletia plant. Sometimes, marcescent leaf bases and exposed stem show more complex patterns than the simple (but common) pattern shown here. The plant represented here would measure around 2 m in height.

compounds that resist decay, and some old plants may still retain leaf remains more than 100 y old (Acosta-Solís, 1984). Fire removes the leaf blades of the marcescent leaves, leaving behind the dense leaf bases which sheathe the stem (these may also be removed by repeated fires). Assuming the plant survives the fire, the living rosette on top of the plant regrows and the new leaves eventually become marcescent, building up a new section of stem clothed in dead leaves. Thus, the stem cover of marcescent dead leaves might be an indicator of time since fire, if the plants have a reliable growth rate regardless of fire. Interestingly, the relationship has been used in reverse, estimating Espeletia growth rate by means of a known fire date (Verweij, 1995). This study presents growth rates for one species, Espeletia pycnophylla subsp. angelensis, in burned and unburned situations in northern Ecuador, and considers mortality rates in a range of fire intensities. Using the measured growth rates and the morphology of burned Espeletia plants in places with known fire history, the reliability of time-since-fire estimates using this method are assessed. Finally, the applicability of this indicator to widespread use in the northern páramos is discussed. 2. Methods 2.1. Study sites The study was carried out within one páramo grassland area in Carchi Province, Ecuador, incorporating Volcán Chiles (at altitudes 3600–4100 m) and El Ángel (in the buffer zone for the Reserva Ecológica El Ángel at altitudes 3600–3700 m).

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

The climate depends significantly on altitude, but published climate data do not exist for these mountain areas and the following details are based on unpublished data from the author’s long-term datalogger studies. Soil temperatures at 10 cm depth peak around 11 ◦ C at 3600 m and 7–9 ◦ C at 4000–4200 m (falling by about 1–2 ◦ C at night at all altitudes). Soil surface and air temperatures vary much more, according to vegetation and cloud cover. At 3600 m daytime air temperatures can reach 15–20 ◦ C, and soil surface temperatures in full sun at midday have been observed >40 ◦ C. At altitudes above 4000 m night-time frosts occur often, but are relatively rare at lower altitudes. Night-time temperatures would be expected to fall below zero at 3600 m only on a few nights of the year, occurring unpredictably. The páramo in this region is considered humid, but annual precipitation also varies altitudinally in relation to the cloud base. The underlying geology is volcanic: fumaroles and thermal springs are present around the summit of Volcán Chiles. An upper layer of highly organic soil (often reaching several metres in depth) sits above the rocks, up to approximately 4200 m altitude; above this vegetation is scattered and soil is poorly developed. The reserve experiences regular burning (Olivera and Cleef, 2009), despite protection in law, with some fires started by visitors. The slopes of Volcán Chiles and the reserve’s buffer zone have been regularly burned by farmers and hunters in the past. Thus, the whole páramo zone consists of a mosaic of patches in different stages of recovery after fires (Valdospinos Navas, 2008). The vegetation across this altitudinal range is dominated by large tussocks (mostly of Calamagrostis, Cortaderia, Rhyncospora and Carex), and giant rosette plants of Espeletia pynconphylla and Puya hamata (Ramsay, 2001). On 6th December 1996, a fire burned over a very large area (>1000 ha) of páramo on Volcán Chiles. This provided an opportunity to study the response of Espeletia plants at 4100 m. A similar opportunity, at 3600 m, was provided by a similarly sized fire on 3rd August 2009. This large fire burned for three days in the buffer zone for the Reserva Ecológica El Ángel, during which time attempts were made to control the fire by firemen, reserve staff, and our team of ecologists. As the fire burned, putative fire intensities were assigned to areas used later in the study by means of qualitative observations of fire temperature (energy release), fire passage, glowing combustion and post-frontal smouldering. Five fire “treatments” were designated, each approximately 1 ha in area: a recently unburned control (where the fire would have burned without intervention), and burned areas of low, medium, high and very high fire intensities. All five areas were located within a distance of 600 m of each other, and were known to have burned last in August 2002 (Bustos Insuasti, 2008). 2.2. Mortality after fire To examine Espeletia mortality after fire, plants were studied at 3600–3700 m in El Ángel, in July 2011, about two years after the fire. For the unburned control, and each of the four fire intensity areas, four replicate samples of 30 Espeletia plants were observed within 1 m either side of a randomly located transect. Each plant was noted as alive or dead, and differences in mortality between sites were assessed using a one-way General Linear Model ANOVA and Student–Newman–Keuls (SNK) post hoc test. All analyses in this study were conducted with Statistica 10 (StatSoft, Tulsa). Overall plant height was also measured (see Fig. 1). In many cases, however, dead plant stems were damaged and could not be measured. Nevertheless, the survey was sufficiently close to the time of the fire that all plants with a developed stem could be found, even the damaged, dead ones. It was not possible to assess the mortality of very young, herbaceous Espeletia plants because there were no longer visible traces of the dead plants.

39

2.3. Growth rates Leaf production and stem height of recently burned and unburned plants, selected at random, were measured after the fires at 4100 m and 3600 m. Soon after each fire, the tip of an emerging leaf in the leaf bud, approximately 3 cm long, was marked with paint. Marked leaves grew normally and eventually died, but the leaf tip remained permanently marked until that part of the leaf broke off. Two measures of growth rate were recorded: • Increase in stem height (to the nearest 5 mm) from the point of the marked leaf attachment on the stem to the base of a leaf bud in the rosette of similar size to the one originally marked. • Leaf production leaves produced since the marked leaf, up to and including an emerging leaf of similar size to the one originally marked. The time intervals over which growth was observed were 238 and 613 days at 4100 m and 3600 m, respectively. Differences between the growth rates of burned and unburned plants were tested in each case with General Linear Model Analysis of Variance. The relationship between plant height and burning was tested with a General Linear Model Analysis of Covariance. 2.4. Consistency of marcescent leaf cover If the marcescent leaf cover of Espeletia might be used as an indicator of time since fire, there must be consistency of leaf cover across individuals within an area of the same fire history. To determine whether such consistency exists and whether it is affected by plant height, Espeletia populations at four altitudes on Volcán Chiles were surveyed: • • • •

3650 m, approximately 1 y after a fire (18 N 169940 89800) 3750 m, approximately 1 y after a fire (18 N 169850 89420) 3850 m, approximately 8 y after a fire (18 N 170150 88950) 4100 m, approximately 8 y after a fire (18 N 172350 89100)

At each altitude, the marcescent leaf cover (see Fig. 1) of 30–50 plants were measured from each of four height categories: 0.50–0.99 m; 1.00–1.49 m; 1.50–1.99 m and >2.00 m. Plants shorter than 50 cm were considered “immature” and not measured. 2.5. Validating the predictions Based on measured growth rates at 3600 m, it should be possible to determine the time interval since the last fire by measuring the marcescent leaf cover of Espeletia plants at this altitude. For four sites of known time since fire, 50 Espeletia plants of 1.0–1.3 m height were randomly selected and their marcescent leaf cover measured. For each place, the predicted time since fire was calculated and validated against the known time. An adjustment, based on the time needed to regrow a complete, average-sized leaf rosette, was also applied to the estimates. 3. Results 3.1. Mortality after fire There were clear differences in mortality rates of Espeletia depending on fire intensity, indicated by increasing mortality from the unburned site to very high fire intensity (Fig. 2; ANOVA, F4,15 = 122.34, p < 0.001). Mortality was somewhat selective: in all sites, plants which died were taller than those that survived (Fig. 3).

40

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

3.3. Consistency of marcescent leaf cover

d

60

The marcescent leaf cover in Espeletia populations was highly skewed by a few plants with much higher than usual cover (Table 2). Also, the smallest plant size category tended to have smaller minimum values than the other size categories. These outliers increased the variance and resulted in higher mean leaf covers, especially in plants taller than 2 m. Removing the outliers at both extremes and using only measurements from the central 80% of a sample substantially reduced variance for taller plants.

Mortality (%)

50

40

c

30

b

20

a 10

3.4. Validating the predictions

a

0 No fire

Low

Medium

High

Very High

Fire Severity Fig. 2. Mortality of Espeletia pycnophylla in five sites representing different fire intensities in the páramo of El Ángel, Ecuador. Standard error bars are shown and means sharing a letter were not significantly different using an SNK test.

It should be noted that 50–90% of dead Espeletia plants could not be measured because their stems had broken and, from personal observation, dead taller plants were more likely to break. Thus, the positive relationship between height and mortality is likely to be even stronger than that presented here. 3.2. Growth rates There was no difference in growth rates between burned and unburned plants, either at 3600 m or 4100 m (Table 1). However, altitude affected growth rates. At 4100 m, leaf production was almost half (49.9 leaves y−1 ) that at 3600 m (93.7 leaves y−1 ), and stem height grew an order of magnitude less at the higher altitude (1.6 cm y−1 ) than at the lower altitude (14.8 cm y−1 ). At 4100 m, taller plants grew faster than shorter ones, but this relationship was unaffected by whether a plant was burned or unburned (Fig. 4; ANCOVA, F1,52 = 0.025, p = 0.876).

Mean Espeletia height (m)

3.0

Live Dead

2.5

2.0

1.5

1.0

0.5

0.0

117

1

No fire

110

1

Low

100 2

Medium

80

12

High

53

34

Very High

Fire Intensity Fig. 3. Comparison of mean live and dead plant heights of Espeletia pycnophylla in five sites representing different fire intensities in the páramo of El Ángel, Ecuador. Standard error of the means (error bars) are only shown where n > 10 (indicated in labels near the x-axis). The stems of some dead plants were broken and could not be measured.

At 3600 m, it was estimated that an average rosette of 79 leaves would need approximately 306 days to recover after fire (based on the observed leaf production rate of about 94 leaves y−1 ). Furthermore, the annual growth rate in stem height at this altitude was 14.4 cm y−1 . Using these figures, crude and adjusted estimates of the time since fire were made in four locations (Table 3). Although there is some variability, times since fire at all four locations were estimated correctly to the nearest whole year by the adjusted calculation. 4. Discussion In the absence of recent burning, only 2.5% of adult Espeletia plants died within a two-year period, similar to the 1.7% reported for E. grandiflora in Colombia (Fagua and Gonzalez, 2007). Fires killed Espeletia adults, with mortality varying from 8 to 56% as fire intensity increased. Two different kinds of mortality were observed, and both would be expected to increase with fire intensity. Fires fuelled by dead leaves of tussock grasses can reach temperatures in excess of 500 ◦ C in the zone where Espeletia rosettes are found, but often these temperatures are short-lived and are easily blocked by insulating material (Ramsay and Oxley, 1996). Fires which pass rapidly (generally low intensity) are unlikely to provoke lethal temperatures at the meristem because it is protected within a dense, insulating leaf bud (Ramsay and Oxley, 1997), but the sustained high temperatures of a high intensity fire could kill the single apical meristem. With this kind of mortality, there was no regrowth of leaves after the fire. However, adult mortality also occurred in another way, some time after the fire following several months of apparent recovery (since leaves were produced after the fire). This kind of delayed mortality was responsible for the deaths of approximately 20% of the adult population in the very high intensity fire area, though infrequently observed in lower fire intensities. The mechanism for this kind of mortality probably results from the loss of insulating leaves on the stem and susceptibility to freezing mortality at a later date (Laegaard, 1992). Fire removes marcescent leaf cover which otherwise insulates stem water on cold nights. A high, earlymorning demand for water from a large leaf rosette can be lethal if the stem water is frozen (Goldstein and Meinzer, 1983). Since freezing nights do not occur frequently at the altitudes where the mortality observations were made, it is likely that mortality could be delayed until a particularly cold night occurred once rosette recovery had generated high water demand from the leaves. In fact, the indirect effects of fire on stem insulation are cumulative. The first fire usually removes the loosely packed leaf blades, but the dense leaf bases remain because their dense packing prevents oxygen from sustaining the fire. However, repeated fires or high intensity fires remove a little of the leaf layer each time, until eventually the stem itself is exposed (Laegaard, 1992). It is this cumulative fire impact that explains the higher mortality of taller plants after fires: they are the most likely to have exposed stems

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

41

Table 1 Estimated rates of leaf production and stem height growth for Espeletia pycnophylla at 3600 m and 4100 m in the páramos of Volcán Chiles and El Ángel, Ecuador. Growth measure

Altitude (m)

Burned páramo (means ± SE)

Unburned páramo (means ± SE)

df

F

P

Stem height (cm y−1 )

3600 4100

15.4 ± 0.9 1.7 ± 0.12

14.4 ± 1.2 1.6 ± 0.13

17 53

0.392 0.180

0.539 0.673

Leaf production (leaves y−1 )

3600 4100

96.9 ± 5.3 49.8 ± 3.1

90.7 ± 5.2 50.0 ± 5.1

17 53

0.678 0.001

0.422 0.981

and suffer direct fire damage to tissues, as well as indirect water stress later. Higher mortality of taller Espeletia plants has also been observed in Espeletia hartwegiana in Colombia (Verweij and Kok, 1992). It is not certain whether the cumulative effects of repeated fires would result in lower growth rates, but it is possible that older plants might show slowed growth rates as a consequence of sub-lethal physiological stress if their stems are exposed to cold temperatures. However, the chances of this affecting the predictions of time since fire could be minimized by avoiding older (taller) plants. Plants 1–1.5 m in height when the fire occurred would be unlikely to have experienced many repeated fires during their lives. Mortality of juvenile Espeletia plants was not recorded in this study. While fires kill significant numbers of juvenile Espeletia plants (Smith, 1981; Verweij, 1995), there is often a post-fire pulse of germination and establishment when the surrounding vegetation is more open (Laegaard, 1992): reducing competition, increasing light (and thus temperatures) and liberating nutrients into the soil (Suarez and Medina, 2001). There were no differences in growth rates of Espeletia plants between burned and unburned sites. In fact, growth rates were remarkably consistent. Given that the burned plants, used for the

growth rate assessment at 3600 m, had experienced a high intensity fire, it seems reasonable to expect that lower intensity fires do not affect subsequent growth rates either. Thus, consistent rates of recovery can be expected, irrespective of fire intensity, a very useful characteristic in terms of this plant’s potential as an indicator of time since fire. The growth of the plant is driven by a single apical meristem producing the rosette of leaves. Providing the meristem is not killed and the plant has sufficient reserves to produce the first replacement leaves, then growth should be little affected by a temporary loss of the photosynthetic apparatus. It is not clear what reserves Espeletia usually holds but on the evidence of this study, growth rates seem little affected by the complete loss of its leaves. In any case, as new leaves are added, the importance of reserves declines. At 3600 m, E. pycnophylla is able to replace its entire leaf complement in around 306 days, with other species needing 204–657 days (Table 4). However, Espeletia growth does vary with altitude and plant height. Plants at 3600 m produced twice as many leaves and more than nine times the height increase of plants at 4100 m. At 4100 m, a 1.5 m-tall plant grew in height approximately 1.75 times faster

Mean Espeletia stem growth (cm y -1)

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.5

1.0

1.5

2.0

Plant height (m) Fig. 4. The relationship between plant height and growth rate for Espeletia pycnophylla plants at 4100 m. Black symbols represent growth rates of burned plants, white symbols those of unburned plants. The fitted logarithmic relationship for all plants combined (p < 0.001, R2 = 0.79) has the formula, stem growth (cm y−1 ) = 0.96 ln(plant height in m) + 2.06.

42

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

Table 2 Marcescent leaf cover on the stems of Espeletia pycnophylla plants in relation to plant height at four different locations, all burned within the previous ten years. Minimum, 10% percentile, mean, 90% percentile and maximum marcescent leaf covers are shown, along (CV%). Adjusted mean and CV% values exclude outliers beyond the 10% and 90% percentiles. Altitude (m)

Height (m)

Marcescent leaf cover (cm) Min

10%

Mean

90%

Max

CV%

Adj mean

Adj CV%

3650 m, 1 y since fire

0.50–0.99 1.00–1.49 1.50–1.99 2.00+

2 4 5 4

3 5 6 5

5.5 8.8 8.6 11.6

7 13 12 11

10 29 12 71

31 49 22 130

5.4 8.4 8.6 7.6

20 28 16 36

3750 m, 1 y since fire

0.50–0.99 1.00–1.49 1.50–1.99 2.00+

3 5 5 4

6 7 6 4

8.3 9.0 12.8 16.6

10 11 10 12

11 14 141 176

19 23 174 205

8.4 9.0 8.5 8.1

13 16 13 24

3850 m, 8 y since fire

0.50–0.99 1.00–1.49 1.50–1.99 2.00+

3 3 4 5

4 4 6 3

5.5 5.9 6.9 11.1

7 7 8 13

8 9 10 98

23 21 17 138

5.5 5.9 6.9 8.4

16 15 11 23

4100 m, 5 y since fire

0.50–0.99 1.00–1.49 1.50–1.99 2.00+

2 2 3 4

3 4 4 5

5.1 6.8 6.2 7.6

7 10 8 10

10 16 12 12

32 44 30 28

5.0 6.4 6.0 7.5

20 28 20 18

Table 3 Predicted versus actual time since fire, based on Espeletia marcescent leaf cover, for four plots at 3600–3700 m, in the páramo of El Ángel, Ecuador. Marcescent leaf cover on stem was based on measurements of 50 plants in each area. The crude estimate of time since fire assumes a stem growth rate of 14.8 cm y−1 . The adjusted estimate includes an additional 0.84 y to account for the time needed to regrow all the rosette leaves after a fire. Site

Mean marcescent leaf cover on stem ± SE (cm)

1 2 3 4

114.6 21.7 12.6 18.0

± ± ± ±

1.4 1.3 0.7 0.6

Crude estimate of time since fire (y)

Adjusted estimate of time since fire (y)

Actual time since fire (y)

8.81 1.67 0.97 1.39

9.41 2.27 1.57 1.94

8.94 1.97 1.97 1.97

Table 4 Summary of published reports of mean Espeletia growth rates in the Andes (and Senecio keniodendron from Kenya, for additional context). Species, location and source

Stem height (cm y−1 )

Leaf production (leaves y−1 )

E. pycnophylla, Ecuador, 3600–3700 m (this study) E. pycnophylla, Ecuador, 4100 m (this study) E. hartwegiana, Colombia, 4000–4150 m (Verweij and Kok, 1992) E. barclayana, Colombia, 3540 m (Cavelier et al., 1992) E. grandiflora, Colombia, 3200 m (Fagua and Gonzalez, 2007) E. flocossa, Venezuela, 3600 m (Smith, 1981) E. schultzii, Venezuela, 3000 m (Smith, 1981) E. schultzii, Venezuela, 3600 m (Smith, 1981) E. schultzii, Venezuela, 3600 m (Smith, 1984) E. schultzii, Venezuela, 4200 m (Smith, 1981) E. lutescens, Venezuela, 4200 m (Smith, 1981) E. humbertii, Venezuela, 3500 m (Smith, 1981) E. spicata, Venezuela, 4100–4200 m (Estrada, 1983) E. spicata, Venezuela, 4200 m (Estrada and Monasterio, 1991) E. timotensis, Venezuela, 4100–4200 m (Estrada, 1983) E. timotensis, Venezuela, 4150 m (Estrada, 1995) E. timotensis, Venezuela, 4400 m (Estrada, 1995) Senecio keniodendron, Kenya, 4180 m (Beck et al., 1984; Hedberg, 1969)

14.8 1.6 8.8

93.7 49.9 118

a b c d

1.8–3.1

29.1–31.1

Mean leaf number in rosette

Estimated rosette turnover (d)

79

306

70

215

36–39

456–460

7.6 4.5 2.0 1.2 0.2 1.1 1.5 4.2

125 81–119 359–894a 72 64 10 380

71 59–75

204 237–274

78 96 13 375b

394 544 489 361

110

173b

574

1.0–2.0

c

1.0–1.1 1.1c 2.5–4.5

c

73–173 73–155c

Leaf production rates reported here are too high and probably misreported (the true figures are probably ten times lower). Calculated from other figures provided. Estimated from figures in the original publication. Based on biomass measurements, not leaf numbers.

511d 657d

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

than a plant only 0.5 m tall. Since growth rates vary between individual plants and across altitudinal gradients, the use of Espeletia as an indicator must be calibrated to local growth rates. In mountains, environmental conditions vary in complex ways at a landscape level, as topography and cloud cover interact with altitude. In addition, different species of Espeletia are very likely to have different growth rates. Espeletia growth rates are poorly known though some figures are available (Table 4). This is surprising since these giant rosette plants represent “botanical big game” (Hedberg, 1969) and measuring growth rates is relatively straightforward. A project has already begun in Ecuador to measure Espeletia growth rates across a range of altitudes, and it is hoped the project can be expanded to incorporate more studies elsewhere. The upper parts of taller Espeletia plants often escape the flames, and fires sometimes miss out individual plants. Thus, after a fire, some plants will have lost all their marcescent leaf blades, some will have been only partially affected, and others will show no effects of the fire at all. In this study, taller plants had more variable leaf cover. Very young plants, although consistent in leaf cover, may have germinated and established after the fire. Therefore, only plants that were 1–1.5 m tall when the fire occurred should be used as indicators. Combining measurements of the marcescent leaf cover with the known growth rate at this altitude, it was possible to estimate the time elapsed since fire (to the nearest whole year) at four different sites. Crude estimates, based solely on annual growth rates, underestimated time since fire. Since the plants have no photosynthetic apparatus immediately after a fire, and it would normally take more than 300 days to produce the number of leaves needed to make up an average E. pycnophylla rosette, some adjustment for a period of particularly slow growth at the start of the recovery period should be made. By adding this average rosette-building period (0.84 y), a more appropriate and accurate estimate of time since fire was obtained. Unfortunately, reliable fire dates are rare in the páramo zone and it was not possible to test the predictions in sites with more than nine years since the last fire. However, given the growth rates of the plants at 3600 m, the method would only be useful for around 20 y anyway. Beyond this time, plants will have reached more than 4 m tall and it is likely that age-related senescence might slow growth rates, and lead to over-estimates of time since fire. The rarity of such tall plants and the practical difficulties in quantifying annual height gains with plants of this size means we have no information about their growth rates. On the basis of this study, a clear protocol for calibrating the use of Espeletia plants as indicators of time since fire can be proposed:

43

3. Measure the marcescent leaf cover of around 50 Espeletia plants, each with the lower limit of marcescent leaf cover at approximately 1–1.5 m above the ground. These plants are likely to be the most consistent because their size during the fire was most likely to result in complete removal of leaves. If a plant is burned more on one side than the other, always measure on the side burned most. 4. Of the 50 plants, discard the top and bottom 10% of the measurements, leaving a core of 40 measurements. This approach will remove any unusual outliers that could lead to poor estimates. Occasionally, fires miss whole plants (their leaves remain intact) and perhaps some individuals grew at an unusual rate. 5. Estimate the time required to regrow the average number of leaves in a rosette, assuming all leaves were originally destroyed by the fire. 6. Estimate the time needed to produce the observed marcescent leaf cover. For short times since fire, a simple calculation can be made by dividing the leaf cover measurement by the annual growth rate. However, for longer periods since the fire, given that plants grow faster the taller they become, a more complex calculation is needed to reflect that relationship. 7. Estimate the time since fire by adding together the results of steps (5) and (6). 5. Conclusions Where Espeletia is common, it provides a mechanism to assess fire severity and intensity (by quantifying adult mortality 1–2 y after the fire) and an estimate of time since fire by measuring marcescent leaf cover in relation to growth rates. Thus, for páramo grasslands in the north of Ecuador and throughout Colombia, recent fire history could be determined using Espeletia as an ecological indicator—in places where it is difficult to assess fire regimes in other ways. However, simple growth studies would need to be performed in order to calibrate the method to local conditions. Then, Espeletia plants would provide a useful indicator of time since fire over a period of up to 20 y at lower altitudes and much longer at higher altitudes. Such information could be incorporated easily and cheaply into other studies where recent fire histories are influential, and where the impacts of land use management need to be monitored. The principles of this method might also have potential in the fire-prone mountain grasslands of East Africa, where convergent evolution has produced remarkably similar plants, such as Senecio keniodendron (Beck et al., 1984). Acknowledgements

1. Estimate annual growth rates of Espeletia plants (at several altitudes, if appropriate). This can be achieved simply and cheaply by marking plants and returning after a fixed period of time. Ideally, at least one year of growth should be measured, and further measurements over a longer period would help to fine-tune the estimates. Both height growth and leaf turnover should be measured. A simple marker of string around the stem at the point where plant height is measured (Fig. 1) has proved to be more reliable in ongoing studies over longer time periods than the paint method described in this study. 2. Identify a representative area of the place where the time since fire is required. Since fire mosaics are common in the páramo, fire histories are complex and care must be taken to ensure the study area shares the same fire history, and does not contain several overlapping fire events of different ages. Obviously, this would lead to results that would be hard to interpret. However, separate measurements in adjacent sites with different fire histories would be a valuable contribution to understanding páramo fire regimes at the landscape scale.

Permits for the work were granted by the Ecuadorian Ministry of Environment (015-IC-DFP and 001-IC-FLO-DPAC), with support from Renato Valencia and Hugo Navarrete at the QCA herbarium of the Universidad Católica del Ecuador, Quito. Field assistance in 1997 was given by Dominic Goetz, Helen Williams, Emma Tattershall and Delyth Thomas, and in 2009–11 by Paola García Meneses, Margarita ˜ Ron, Mayra Ninazunta Anaguano, Tumezghi Kidane and Pete Mino Morley. References Acosta-Solís, M., 1984. Los Páramos Andinos del Ecuador. Publicaciones Cientificas M.A.S, Quito. Beck, E., Schlutter, I., Scheibe, R., Schulze, E.D., 1984. Growth rates and rejuventation of East African giant groundsels (Senecio keniodenron). Flora 175, 243–248. Bromley, R.J., 1971. Neurolepis as a sensitive indicator of human activity in the High Andes. Am. Antiquity 36, 475. Bustos Insuasti, A.M., 2008. Propuesta metodológica para monitorear incendios de la vegetación en áreas protegidas aplicadas a la Reserva Ecológica El Ángel. Pontificia Universidad Católica Del Ecuador, Quito.

44

P.M. Ramsay / Ecological Indicators 45 (2014) 37–44

Buytaert, W., Celleri, R., De Bievre, B., Cisneros, F., Wyseure, G., Deckers, J., Hofstede, R., 2006. Human impact on the hydrology of the Andean paramos. Earth-Sci. Rev. 79, 53–72. Carlquist, S., 1994. Anatomy of tropical alpine plants. In: Rundel, P.W., Smith, A.P., Meinzer, F.C. (Eds.), Tropical Alpine Environments: Plant Form and Function. Cambridge University Press, Cambridge, pp. 111–128. Cavelier, J., Machado, J.L., Valencia, D., Montoya, J., Laignelet, A., Hurtado, A., Varela, A., Mejía, C., 1992. Leaf demography and growth-rates of Espeletia barclayana Cuatrec (Compositae), a caulescent rosette in a Colombian paramo. Biotropica 24, 52–63. Di Pasquale, G., Marziano, M., Impagliazzo, S., Lubritto, C., De Natale, A., Bader, M.Y., 2008. The Holocene treeline in the northern Andes (Ecuador): first evidence from soil charcoal. Paleogeogr. Paleoclimatol. Paleoecol. 259, 17–34. Estrada, C.A., 1983. Dinámica del crecimiento y reproducción de Espeletia en el páramo desértico. Universidad de los Andes, Mérida, Venezuela. Estrada, C.A., 1995. Procesos productivos en poblaciones de Espeletia timotensis en el páramo desértico. Universidad de los Andes, Mérida, Venezuela. Estrada, C.A., Monasterio, M., 1991. Compartamiento reproductivo de una roseta gigante, Espeletia spicata Sch. Bip. (Compositae), del Páramo Desértico. Ecotropicos 4, 1–17. Fagua, J.C., Gonzalez, V.H., 2007. Growth rates, reproductive phenology, and pollination ecology of Espeletia grandiflora (Asteraceae), a giant Andean caulescent rosette. Plant Biol. (Stuttg.) 9, 127–135. García-Meneses, P.M., Ramsay, P.M., 2014. Puya hamata demography as an indicator of recent fire history in the páramo of El Ángel and Volcán Chiles, vol. 36. no. 1. Caldasia, Ecuador-Colombia. Goldstein, G., Meinzer, F.C., 1983. Influence of insulating dead leaves and low temperatures on water balance in an Andean giant rosette plant. Plant Cell Environ. 6, 649–656. Harden, C.P., 2006. Human impacts on headwater fluvial systems in the northern and central Andes. Geomorphology 79, 249–263. Hedberg, O., 1969. Growth rate of the east African giant senecios. Nature 222, 163–164. Hofstede, R., Segarra, P., Mena Vásconez, P., 2003. Los Páramos del Mundo. Proyecto Atlas Mundial de los Páramos. Global Peatland Initiative/NC-IUCN/EcoCiencia, Quito. Hofstede, R.G.M., 1995. Effects of livestock farming and recommendations for management and conservation of páramo grasslands (Colombia). Land Degred. Rehabil. 6, 133–147. Horn, H.S., Kappelle, M., 2009. Fire in the páramo ecosystems of Central and South America. In: Cochrane, M.A. (Ed.), Tropical Fire Ecology: Climate Change, Land Use, and Ecosystem Dynamics. Praxis, Chichester, UK, pp. 505–539. Kappelle, M., Horn, S.P., 2005. Páramos de Costa Rica. Santo Domingo de Heridia, Costa Rica. Keating, P.L., 1998. Effects of anthropogenic disturbances on Paramo vegetation in Podocarpus National Park Ecuador. Phys. Geogr. 19, 221–238. Keating, P.L., 2007. Fire ecology and conservation in the high tropical Andes: observations from northern Ecuador. J. Lat. Am. Geogr. 6, 43. Laegaard, S., 1992. Influence of fire in the grass páramo vegetation of Ecuador. In: Balslev, H., Luteyn, J.L. (Eds.), Páramo: An Andean Ecosystem under Human Influence. Academic Press, London, pp. 151–170.

Luteyn, J.L., 1999. Páramos: A Checklist of Plant Diversity, Geographical Distribution and Botanical Literature. The New York Botanic Garden, New York. Olivera, M.C.M., Cleef, A.M., 2009. Vegetation composition and altitudinal distribution of Andean rain forests in El Angel and Guandera reserves, northern Ecuador. Phytocoenologia 39, 175–204. Poulenard, J., Podwojewski, P., Janeau, J.L., Collinet, J., 2001. Runoff and soil erosion under rainfall simulation of andisols from the Ecuadorian páramo: effect of tillage and burning. Catena 45, 185–207. Ramsay, P.M., 1999. Landscape mosaics in the High Andes: the role of fire in paramo communities. In: Kovár, P. (Ed.), Nature and Culture in Landscape Ecology: Experiences for the 3rd Millennium. The Karolinum Press, Prague, pp. 192–199. Ramsay, P.M., 2001. The zonal páramo vegetation of Volcán Chiles. In: Ramsay, P.M. (Ed.), The Ecology of Volcán Chiles: High-Altitude Ecosystems on the Ecuador–Colombia border. Pebble & Shell, Plymouth, pp. 27–38. Ramsay, P.M., Oxley, E.R.B., 1996. Fire temperatures and postfire plant community dynamics in Ecuadorian grass paramo. Vegetatio 124, 129–144. Ramsay, P.M., Oxley, E.R.B., 1997. The growth form composition of plant communities in the Ecuadorian paramos. Plant Ecol. 131, 173–192. Ramsay, P.M., Oxley, E.R.B., 2001. An assessment of aboveground net primary productivity in Andean grasslands of central Ecuador. Mt. Res. Dev. 21, 161–167. Rangel Ch, J.O., 2000. Colombia, Diversidad Biótica III. La región de vida paramuna. Universidad Nacional de Colombia, Bogotá. Sklenáˇr, P., Ramsay, P.M., 2001. Diversity of zonal páramo plant communities in Ecuador. Divers. Distrib. 7, 113–124. Smith, A.P., 1981. Growth and population dynamics of Espeletia (Compositae) in the Venezuelan Andes. Smithson. Contrib. Bot. 48, 1–45. Smith, A.P., 1984. Postdispersal parent-offspring conflict in plants: antecedent and hypothesis from the Andes. Am. Nat. 123, 354–370. Smith, A.P., 1994. Introduction to tropical alpine vegetation. In: Rundel, P.W., Smith, A.P., Meinzer, F.C. (Eds.), Tropical Alpine Environments: Plant Form and Function. Cambridge University Press, Cambridge, pp. 1–19. Suarez, E., Medina, G., 2001. Vegetation structure and soil properties in Ecuadorian paramo grasslands with different histories of burning and grazing. Arctic Antarct. Alpine Res. 33, 158–164. Suárez, E., Toral, C., 1996. Abundancia y biomasa de lombrices de tierra en páramos con distinto uso del suelo en el Ecuador: evaluación preliminar de los efectos de las quemas. EcoCiencia, Quito. Valdospinos Navas, C.M., 2008. Aplicación del modelo de mapeo del peligro de incendio forestal usando Sistemas de Información Geográfica para evaluar el peligro de incendios de la vegetación en el páramo. In: Estudio de caso: Reserva Ecológica El Ángel, Escuela de Geografía. Pontificia Universidad Católica del Ecuador, Quito, pp. 77. Verweij, P.A., 1995. Spatial and Temporal Modelling of Vegetation Patterns. University of Amsterdam. Verweij, P.A., Kok, K., 1992. Effects of fire and grazing on Espeletia hartwegiana populations. In: Balslev, H., Luteyn, J.L. (Eds.), Páramo: An Andean Ecosystem under Human Influence. Academic Press, London, pp. 215–229. White, S., 2013. Grass paramo as hunter-gatherer landscape. Holocene 23, 898–915. Williamson, G.B., Schatz, G.E., Avlarado, A., Redhead, C.S., Stam, A.C., Sterner, R.W., 1986. Effects of repeated fires on tropical páramo vegetation. Trop. Ecol. 27, 62–69.