89 7 Fire, Plant Species Richness, and Aerial Biomass Distribution in Mountain Grasslands of Northwest Argentina Roxana Aragón, Julietta Carilla, and Luciana Cristóbal INTRODUCTION Grazing and fire are the most common distur- bances in many grassland ecosystems around the world (McNaughton et al. 1993; Vogl 1974 in Oesterheld et al. 1999, De Baro et al. 1998), and they both affect biodiversity and plant com- munity dynamics. Grazing and fire influence species composition and richness, determine dominant life-forms and therefore the general structure of the community (Belsky 1992; Diaz et al. 1992; Milchunas and Lauenroth 1993; Collins et al. 1998). They can also regulate eco- system processes such as nutrient cycling (Hobbs et al. 1991) and plant productivity (McNaughton 1985; Rusch and Oesterheld 1997). Importantly, grazing and fire often occur together, and they interact deeply. Grazing and fire are both consumers of plant production. Herbivores feeding on forage can determine the fuel load. Fire, in turn, con- sumes accumulated biomass that could be used by herbivores (Oesterheld et al. 1999). Grazing can influence fire frequency and intensity, and fire determines what is left for herbivores, not only in terms of quantity but also in terms of forage quality (Hobbs et al. 1991). In addition, these disturbances provide open space for col- onization that, in turn, can modify species diversity, promote the establishment of certain species, and change the general structure of the community (Collins 1987; Pucheta et al. 1998; Valone and Kelt 1999). Grazing and fire occur naturally in many grasslands and savannas, and they also are part of many management prac- tices. In addition, burning in grasslands and savannas has an important worldwide effect because it is one of the major sources of atmo- spheric methane and CO 2 , especially in tropical areas (Crutzen et al. 1985 in Hobbs et al. 1991). Livestock raising is one of the most impor- tant land uses in many montane grasslands (Eckholm 1975). Particularly in Andean grass- lands, extensive cattle grazing is often com- bined with burning of the natural vegetation (Schmidt and Verweij 1992 in Hofstede et al. 1995, Grau and Brown 2000). Fire promotes resprouting and is believed to encourage the development of more palatable life-forms (Grau and Brown 2000). However, grazing and fire can also increase soil susceptibility to erosion, reduce species or functional richness (Lloret and Vila 2003), and modify community com- position (Pucheta et al. 1998; Diaz et al. 1992). Eventually, their positive or negative effects depend on an array of factors such as grazing intensity, fire frequency, and climate. Mountain grasslands are one of the most species-rich habitats of northwest Argentina. They are important in regulating the hydric regime and in providing economic resources (e.g. cattle ranching and scenic values). In spite of their ecological and economical importance, 3523_book.fm Page 89 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC 90 Land Use Change and Mountain Biodiversity mountain grasslands are scarcely represented in the protected areas of Argentina, and little is known about their functioning. The study site of this work, the valley of Los Toldos, is located in the upper Bermejo River basin and is con- sidered an area of high conservation priority at a national level (Brown et al. 2001). The dom- inant land use is for grazing by cattle, and this is combined with periodic fires. As was observed in other neotropical mountains, recent works suggest a decrease in land use intensity in this area (Grau et al. submitted) . The reduc- tion in the density of animals may produce changes in fire frequency and intensity that can, in turn, affect plant communities in different ways. In this chapter, we describe a study on how fires affect vegetation structure in the mountain grasslands of northwest Argentina that are used for grazing. More specifically, this study intends to investigate the effect that the time since the last fire event may have on plant species richness, vegetation structure, and bio- mass dynamics. METHODS S TUDY A REA The study was performed at the valley of Los Toldos (22 ° 30 S, 64 ° 50 W), Santa Victoria, Salta, Argentina. The study area consists of a mosaic of mountain grasslands and Alnus acuminata forest patches at an altitude of about 1700 masl. This area lies in the upper altitudinal level of the phytogeographic province of the Argentinean Yungas (subtropical montane for- est) (Cabrera 1976). The original vegetation seems to have been dominated by forest patches, but a long history of grazing in the valley probably shaped the current vegetation physiognomy (Malizia 2003). The mean annual temperature is 15 ° C, and the average precipita- tion is 1300 mm (Ramadori 1995). The precip- itation is highly seasonal, with most of the rain falling during the summer months (Bianchi 1981). The main disturbances at this altitudinal range are grazing, fire, and landslides (Grau, 2005), and livestock raising is the most com- mon land use. Cattle grazing is extensive with no fences limiting individual properties. There are no data on grazing intensity in Los Toldos, but information provided by national agricultural censuses for Santa Victoria shows a decrease in the population of domestic ani- mals during the 20th century (Grau et al. sub- mitted). These data, together with information provided by local people, suggest that grazing intensity in Los Toldos is currently low (between 0.5 and 1 cow per 10 hectares). The pastoral system involves transhumance, a sea- sonal movement of cattle from the highlands to midaltitude and piedmont forests (Grau and Brown 2000). Cattle are driven up to the high- land grasslands at the beginning of the sum- mer period and, in March, they are brought back to lower ranges (piedmont forest). Dur- ing the summer period (from November to March), the animals feed mainly in grassland patches, but also browse in the Alnus forest understory. Summer grazing by cattle is usu- ally prepared for by burning the vegetation in spring. The extent and frequency of burning seem to depend on the proximity to settle- ments and on the weather conditions (wind, temperature, and soil humidity) when the fire is started. As a result of these management practices, the landscape consists of a mosaic of vegetation patches, differing in the time since the last burning event occurred. S AMPLING D ESIGN AND D ATA A NALYSIS In November 2000, we conducted a survey in the study site, looking for evidence of previous fire events. Based on this survey and on infor- mation provided by local inhabitants, we iden- tified three types of vegetation patches that dif- fered in the time since the last fire event occurred: 1. Areas burned during the ongoing growing season (the last fire event probably occurred during spring 2000). These areas showed evident signs of fire, such as abundant char- coal, ashes, and burned vegetation. 2. Areas burned during the previous growing season (spring 1999), with 3523_book.fm Page 90 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC Fire, Plant Species Richness and Biomass in Mountain Grasslands of NW Argentina 91 some evidence of fire (mainly the remains of charcoal). 3. Areas not burned recently. In this case, the last fire event apparently took place at least 5 years ago (spring 1995 or earlier). This information was checked with local residents. We selected three patches of each vegeta- tion type (nine in total). The time since the last fire event was regarded as “treatment.” Hereaf- ter, we will refer to the different treatments as: <1 year (areas burned during the ongoing grow- ing season); >1 year (areas burned during the previous season), and >5 years (areas not burned for at least 5 years). Unfortunately, since burning is a common practice in this area, we did not have any plots that had no fire and could have served as a control. All the vegetation patches included in our sample were no more than 3 km apart from each other, had areas of less than 500 m 2 , and were in similar topo- graphic positions. Because of the absence of fences, vegetation patches had potentially sim- ilar grazing pressure. In December 2000, we conducted plant relevés in 1 m × 1 m plots with five plots per patch. The plots were placed every 10 m in a 50-m transect. The transects were placed at ran- dom in each patch (i.e. 1 transect in each patch). Each plot was divided into four 0.5 m × 0.5 m quadrats, and all the plant species present were recorded. In addition, we collected all the aerial biomass in ten 0.2 m × 0.2 m plots in each patch. Whenever possible, the plots were placed in two 50-m transects that were separated by 10 m. If the vegetation patches were not big enough, we placed plots in shorter transects, but always used the same number of plots. We col- lected biomass in December 2000 and in Janu- ary, February, March, and August 2001 (before the next burning event). The biomass was clas- sified into live biomass, standing dead, and lit- ter. Live biomass was further classified for dif- ferent life-forms (i.e. graminoids, tussock grasses, erect species, rosettes, prostrate spe- cies, ferns, and woody species). All the material was classified, dried to constant weight at 70 ° C, and weighted. We used ANOVA tests for the comparisons between treatments (both for total biomass and proportions). In the case of species richness and total number of species, we used Kruskal–Wal- lis tests, a nonparametric technique, because the assumptions required for parametric tests were not met. Differences in the biomass collected throughout the year were tested through repeated measures ANOVA. Species frequency was computed as the number of plots per treat- ment in which each species was recorded. Equativity was computed as: E m = – Σ ( p i log p i )/log N where p i is the proportion of the species recorded in transect m that belong to life- form i , and N is the total number of different life-forms in that transect. Small values of E imply that one or a few life-forms are dominant in the community; in other words, this index is an indication of evenness (O’Neill et al. 1988). To measure compositional similarity among plots, we performed a detrended corre- spondence analysis (DCA) with downweight- ing of rare species. Only the species that were recorded in at least two plots were considered. We computed a nonparametric Kendall’s tau correlation between plot scores in the ordina- tion space and the time since the last fire event. The analyses were performed in Statistica (StatSoft 1993) and PCORD (McCune and Mefford 1997). RESULTS We recorded a total of 149 species in the study area. In Table 7.1, we have included only the species that had a frequency 0.4 in at least one of the treatments. Among these 45 species, 32 were common to all the treatments. The number of species per square meter did not differ between the treatments (Kruskal-Wallis test, KW 3.29, p = .19) (Table 7.1), and the most common species were present in all the patches independently of their fire history. The most common grasses were Elionurus muticus and Paspalum notatum; Stevia yaconensis was the most frequent woody species, and the Cuphea sp. was the dominant prostrate species. The ordination of plots in the DCA was not clearly linked to the treatments. The first axis 3523_book.fm Page 91 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC 92 Land Use Change and Mountain Biodiversity of the ordination explained approximately 30% of the overall variance in species data λ 1 = 0.259, total inertia = 0.859), and plot scores were not significantly correlated with the time since the last fire event (Kendall’s tau = 0.48 p = .07). However, there seemed to be some minor changes in species composition in response to the treatments because we found that some species were differentially recorded in certain plots. Anemone decapetala and Tes- saria fastigiata were recorded only in plots that were recently burned. Eupatorium bupleurifo- lium and Ophioglossum sp. were more abundant in the >5-year treatment, whereas Baccharis tridentata and Setaria sp. were predominantly recorded in <1-year treatment (Table 7.1). In addition, the equativity of life-forms showed a slight tendency to decrease in the patches that were not burned for 5 years (KW 5.42, p = .06) (Table 7.1). The decrease in equativity in areas that were not burned for 5 years was related to the increasing dominance of woody species in comparison to other life-forms that were less frequently recorded, such as erect species and rosettes. The total aerial biomass was significantly higher in the patches that were not recently burned (F = 69.59, p < .001). The biomass in the >1-year treatment was almost twice as high as the biomass in the <1-year treatment, and the biomass in the >5-year treatment was more than 3 times the biomass in <1-year treatment (431.48 ± 27.95, 738.09 ± 63.64, and 1303.25 ± 58.79 g m − 2 for <1-year, >1-year, and >5-year treatments, respectively, Figure 7.1). There was no difference between <1-year and >1-year treatments with respect to total live biomass, but live biomass was highest in the >5-year plots (F = 22.02, p < .01). There was no differ- ence in total standing dead material (F = 3.33, p = .10), but the total amount of litter differed between the treatments (F = 71.17, p <.001). Patches that were burned in the ongoing grow- ing season (<1-year) had considerably less litter than both >1-year and >5-year patches (Figure 7.1). The relative contribution of the different biomass categories differed between the treat- ments. Live biomass had a high contribution to the total biomass in the patches that were burned during the ongoing growing season (<1- yr) (60 ± 0.5%), whereas the proportion of litter was minimum in this treatment (15, 27, and 41% in <1-year, >1-year, and > 5-year treat- ments, respectively) (F = 5.14, p = .04 for ANOVA on live biomass and F = 73.33, p < 0.001 for ANOVA on litter) (Figure 7.2). The contribution of standing dead material was reduced in the patches that were not burned for 5 years (22, 29, and 10%, respectively) (F = 9.77, p < .01). The proportion of live biomass differed between <1-year and >1-year treatments, but there was no difference between these two treat- ments and the >5-year treatment. But impor- tantly, although the overall proportion of live biomass was similar between the <1-year and >5-year treatments, the relative contribution of the different life-forms to the total of live bio- mass was quite distinct. Patches that were recently burned (<1 year) had a high proportion of erect species and ferns compared to the other treatments (Table 7.2). The proportion of tus- sock grasses plus graminoids did not differ between <1-year and >1-year treatments but their contribution was significantly smaller in the patches that were not burned for 5 years (28 and 33% in <1-year and >1-year and 15% in >5-year treatments) (Table 7.2). This differ- ence was mainly due to tussock grasses that were reduced in >5-year patches (23, 25, and 9%, respectively, in the <1-year, >1-year, and >5-year treatments). The biomass of grami- noids was similar in all three treatments. Woody species accounted for 72 ± 9% of the live bio- mass in the >5-year treatment (Table 7.2). The seasonal dynamics of the total live bio- mass, standing dead, and litter showed some similarities between the treatments. Biomass assigned to the standing dead compartment showed a peak in August in all three treatments (Table 7.3). Similarly, litter had its maximum value in August in the <1-year and >5-year treatments, but we did not detect any seasonal trend in the >1-year patches. Live biomass showed a significant decrease in August in the >1-year patches, and a small peak in March and December; however, no similar trend was detected in the other two treatments (Table 7.3). Interestingly, the relative contribution of live biomass throughout the year strongly differed between the treatments. Patches that were 3523_book.fm Page 92 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC Fire, Plant Species Richness and Biomass in Mountain Grasslands of NW Argentina 93 TABLE 7.1 Species with frequencies ≥ 0.4 in at least one of the different treatments a <1 year >1 year >5 years p Values b Mean number of species per m 2 25.20 25.67 20.27 .19 Total number of species 89 96 83 .11 Equativity of life-forms 0.88 0.93 0.77 .06 Species Life-forms <1 year >1 year >5 years Achyrocline sp. Rosette 0 0.40 0.20 Agrostis sp. Graminoid 0.13 0.40 0.07 Anemone decapetala Prostrate 0.40 0 0 Baccharis coridifolia Woody 0.73 0.73 0.53 Baccharis rupestris Woody 0.60 0.60 0.27 Baccharis tridentata Woody 0.73 0.40 0.13 Chaptalia modesta Rosette 0.53 0.20 0.2 Chevreulia acuminata Rosette 0.20 0.53 0.40 Clitoria cordobensis Prostrate 0.67 0.07 0.13 Croton sp. Woody 0 0.33 0.47 Cuphea sp. Prostrate 1 1 1 Cynodon sp. Graminoid 0.33 0.67 0.13 Cyperaceae Graminoid 0.40 0 0.20 Desmodium affine Prostrate 0.40 0.73 0.67 Desmodium sp. Prostrate 0.40 0.40 0.80 Elionurus muticus Tussock grass 1 1 0.67 Eupatorium bupleurifolium Woody 0.13 0.40 0.93 Euphorbiaceae Prostrate 0.87 0.13 0.27 Gamochaeta sp. Rosette 0.40 0 0.27 Hybanthus parviflorus Erect 0.27 0.67 0.20 Hypericum sp. Erect 0.40 0.53 0.07 Hyptis mutabilis Woody 0.20 0.40 0.60 Juncus sp. Graminoid 0.67 0.40 0.20 Unknown 3 (Myrtaceae) Woody 0.20 0.40 0.27 Lepechinia vesiculosa Woody 0 0.80 0.67 Lobelia nana Prostrate 0.67 0.53 0.67 Malvaceae Prostrate 0.33 0.40 0.33 Ophioglossum sp. Fern 0 0.26 0.47 Panicum ovuliferum Graminoid 0 0.47 0.33 Panicum sp. Graminoid 0 0.60 0.33 Paspalum notatum Graminoid 0.53 0.47 0.47 Paspalum sp. Tussock grass 0.87 0.47 0.67 Plantago sp. Woody 0.60 0.07 0.20 Polygala pulchella Erect 0.60 0.20 0 Polygala sp. Erect 0.20 0.40 0 Pteridium aquilinum Fern 0.47 0.27 0.07 Ranunculus praemorsus Erect 0.20 0.40 0.20 Richardia sp. Erect 0.53 0.13 0.13 Setaria sp. Graminoid 0.53 0.07 0 Stenandrium dulce Rosette 0.13 0.53 0.20 Stevia alpina Woody 0.80 0.67 0.13 Stevia yaconensis Woody 1 1 0.93 Tagetes filifolia Prostrate 0.40 0.20 0.07 Tessaria fastigiata Woody 0.40 0 0 Tibouchina sp. Woody 0.13 0.20 0.40 a <1 year: patches that were burned in the ongoing growing season, >1 year: burned the previous season, or >5 years: not burned for 5 years; mean number of species per m 2 . Total number of species and equativity of life-forms per treatment is also shown. b p values correspond to Kruskal–Wallis tests. 3523_book.fm Page 93 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC 94 Land Use Change and Mountain Biodiversity FIGURE 7.1 Variation of biomass according to different treatments. Live biomass, standing dead, litter, and total biomass (g m –2 ) in patches that were burned in the ongoing growing season (<1 year), burned in the previous season (>1 year), or not burned for 5 years (>5 years). Different letters denote significant differences at p < .05 according to an ANOVA test, and ns indicates no significant differences. The comparisons were made between treatments within each biomass category. FIGURE 7.2 Relative contribution of live biomass, standing dead, and litter to the total biomass in the different treatments: patches that were burned in the ongoing growing season (<1 year), in the previous season (>1 year), or not burned for 5 years (>5 years). Different letters stand for significant differences at p < .05 according to an ANOVA test. The comparisons were made between treatments within each biomass category. a a b ns ns ns a b c a b c 0 200 400 600 800 100 0 120 0 140 0 160 0 < 1 > 1 > 5 y e ar s Time since last fire event (y ears ) Dry weight (gm -2 ) Live biomass Standing dead Litter Total biomass c b a b a a cd bd ac 0 0. 2 0. 4 0. 6 0. 8 1 <1 >1 >5 Time since last fire event (years) Biomass (proportion) Live biomass Standing dead Litter 3523_book.fm Page 94 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC Fire, Plant Species Richness and Biomass in Mountain Grasslands of NW Argentina 95 burned during the ongoing growing season (<1 year) had 70 to 80% of their biomass as live biomass during December and January (Figure 7.3), whereas in the >1-year and >5- year treatments, this proportion hardly approached 50%. The contribution of live bio- mass to the total was higher in the <1-year patches almost throughout the year, which may represent substantial changes in the seasonal pattern of forage availability. DISCUSSION Time since the last fire event affected the total aerial biomass, the proportion of live biomass, standing dead, and litter, and the contribution of the different life-forms, both in terms of bio- mass and life-form frequency. Nevertheless, species richness was similar among all treat- ments, and species composition showed only small variations. Our results differ from those of Collins (1987) and Pucheta et al. (1998), who TABLE 7.2 Relative contribution (mean and standard deviation) of the different life-forms to the live biomass Life-Forms <1 year a >1 year a >5 year a p -Values b Erect species 0.18 ± 0.02 0.08 ± 0.04 0.07 ± 0.03 .06 Ferns 0.10 ± 0.09 0.03 ± 0.03 0.02 ± 0.01 .67 Graminoids 0.05 ± 0.01 0.09 ± 0.01 0.06 ± 0.06 .30 Prostrate species 0.05 ± 0.02 0.05 ± 0.02 0.03 ± 0.01 .39 Rosettes 0.01 ± 0.01 0.01 ± 0.01 0.01 ± 0.01 .20 Tussock grasses 0.23 ± 0.04 0.26 ± 0.03 0.09 ± 0.09 .06 Woody species 0.37 ± 0.09 0.49 ± 0.05 0.72 ± 0.09 .03 a <1 year: patches that were burned in the ongoing growing season, >1: burned the previous season or >5 years: not burned for 5 years. b p values correspond to Kruskal–Wallis tests. TABLE 7.3 Seasonal dynamics of total live biomass, standing dead, and litter (mean and standard error) a Biomass Compartment December January February March August F b p <1 year Live 207.64 ± 16.18 371.77 ± 95.92 209.05 ± 29.84 262.85 ± 41.31 185.11 ± 34.51 1.81 0.21 Standing dead 23.18 ± 5.22 85.87 ± 43.11 100.80 ± 37.90 98.09 ± 23.25 252.39 ± 37.77 6.12 0.01 Litter 32.40 ± 11.27 49.55 ± 19.27 73.15 ± 18.21 38.72 ± 7.03 167.80 ± 19.38 10.80 0.002 >1 year Live 391.53 ± 66.42 288.44 ± 44.93 327.87 ± 25.26 445.95 ± 30.70 184.76 ± 50.86 13.22 0.001 Standing dead 196.02 ± 47.97 159.19 ± 20.78 172.43 ± 32.48 184.92 ± 59.82 462.35 ± 56.28 20.01 0.003 Litter 235.88 ± 80.12 175.20 ± 18.92 172.84 ± 20.62 209.41 ± 39.03 313.61 ± 62.52 1.38 0.32 >5 years Live 666.15 ± 52.82 542.86 ± 55.39 557.21 ± 94.12 611.96 ± 46.36 662.14 ± 80.68 1.80 0.22 Standing dead 142.35 ± 38.75 84.80 ± 35.89 92.57± 29.44 110.06 ± 29.17 298.33 ± 117.10 3.90 0.04 Litter 603.98 ± 112.18 409.74 ± 10.33 549.87 ± 65.30 367.83 ± 82.96 816.37 ± 84.52 5.88 0.05 a In areas that were burned in the ongoing growing season (<1 year), burned the previous season (>1 year), or not burned in 5 years (>5 years). b Boldened cells indicate significant differences at p < .05. 3523_book.fm Page 95 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC 96 Land Use Change and Mountain Biodiversity found that disturbances such as grazing and fire increased both species richness and diversity. In many cases, the increment in the number of species results from the colonization by exotic species or from the predominance of small- sized species, which are tolerant to disturbance (Belsky 1992; Pucheta et al. 1998). In our study site, we did not record exotic species, and because all our patches have a long history of grazing, most of these species may indeed be tolerant to disturbances. Fire and grazing may produce the same kind of selective pressure, and they can both favor fast-growing or small- sized species, especially tussock grasses and annuals. For these reasons, fire suppression may not cause compositional changes in areas such as our study site, where grazing occurs simul- taneously. Although species richness remained similar in the different treatments, we detected changes in the life-form spectrum and in the distribution of aerial biomass. The reduction in functional or species diversity as a consequence of a decrease in disturbance frequency has been observed in many cases (e.g. Pucheta et al. 1998; Valone and Kelt 1999) and is often attrib- uted to a strengthening in species competition. In the grasslands of Los Toldos, fire suppression caused an increase in the dominance of woody species; many of these species were present in burned plots, but they became more abundant and of a bigger size in plots that were not recently burned. Tussock grasses were favored by fire, but their contribution was reduced in >5-year treatment. This change in the domi- nance of woody species alters site flammability that might reduce fire frequency in these plots in the future. In addition to the changes in life-form con- tribution, there was a change in the distribution of aerial biomass. Fire reduced the total bio- mass by more than two-thirds (1303 gm –2 in areas not burned for 5 years compared to 431 gm –2 in areas that were recently burned), and the amount of litter was reduced in a similar way. This reduction in aboveground biomass that is associated with changes in the contribu- tion of different life-forms results in changes in vegetation structure that may alter soil cover. Modification of soil cover can, in turn, affect FIGURE 7.3 Relative contribution of live biomass to the total biomass throughout the year in patches that were burned in the ongoing growing season (<1 year), in the previous season (>1 year), or not burned for 5 years (>5 year). 0 0.2 0.4 0.6 0.8 1 Dec Jan Feb Mar Aug Biomass (proportion) <1 year >1 year >5 years 3523_book.fm Page 96 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC Fire, Plant Species Richness and Biomass in Mountain Grasslands of NW Argentina 97 erosion hazards that may have further implica- tions on the hydrology and nutrient dynamics of the system (Hofstede et al. 1995). The dif- ferential allocation of biomass to the distinct biomass compartments and, especially, the vari- ation in the amount of dead material that reaches the soil, can alter the decomposition rate and, consequently, the nutrient pools (Hobbs et al. 1991). Unfortunately, due to the lack of sound information, at present we can only hypothesize about these effects in the grasslands of Los Toldos. On the other hand, the short-term effect of fire on forage availabil- ity to herbivores in this site is easier to appre- ciate. Patches that were burned in springtime had more than 70% of their total biomass as live biomass in the following summer (December and January). Therefore, fire modifies the sea- sonal dynamics of aerial biomass and changes forage availability at least for the summer period, when livestock is brought up to these mountain grasslands. The availability of green forage, especially in the form of highly palat- able grasses, is particularly important for cattle after a period when they have had access only to low-quality winter forage. This means that cattle obtain, in proportion, more green biomass per bite in the patches that were recently burned. This can explain why these patches are often preferred (Coppock and Detling 1986; Hobbs et al. 1991). Consequently, the propor- tion of live biomass can have important effects on livestock energy budgets and determine their local movements. Importantly, fire promotes more palatable life-forms (grasses instead of woody species), and this makes the effect of fire even more meaningful to herbivores. Our results indicate that changes in the fire frequency strongly affect vegetation dynamics in the montane grasslands of northwest Argen- tina. However, it is worth pointing out some limitations of this study. First, we were unable to find areas that were not burned and, there- fore, we lacked a true control for our treatments. Our conclusions refer to the effects of a change in the fire frequency from once a year to once in 5 years. We do not know if there is a threshold after which a reduction in fire frequency pro- duces no further changes in plant communities, so we cannot say if our >5-year treatment patches represent a transitional or a steady state. Second, our sample size was rather small, espe- cially with regard to species composition. This is why we gave major emphasis to the results referred to biomass distribution, the variability of which seems to have been sufficiently accounted for by our samples. Third, an assumption of the present study is that our sam- ple patches experience a similar grazing pres- sure. Even though there are no fences or other obstacles, and livestock have free access to all patches in the study area, which are also very close to one another, animals, as mentioned ear- lier, may prefer recently burned grasslands. As a consequence, these patches may receive higher grazing pressure. All these limitations have to be taken into account when considering our conclusions. To overcome these inherent difficulties, we are currently carrying out a con- trolled experimental study with a bigger sample size in the Los Toldos grasslands, which aims at separating the effects of fire and grazing. The preliminary results of this new experimental setup, which has been running for more than 2 years, seem to support the findings reported here. SUMMARY Fire and grazing are the most common distur- bances in the mountain grasslands of northwest Argentina. They can affect species composition and richness, determine dominant life-form, and the general structure of the community. This work aims to determine the effect of burn- ing on species richness, vegetation structure, and aerial biomass distribution in the grasslands of northwest Argentina that are subjected to grazing. We performed a comparative study at Los Toldos, Salta, Argentina (22º30 S, 64º50 W) at 1700 masl and surveyed patches that differed in the time since the last fire event. We considered three treatments: patches that were burned during the ongoing growing sea- son (in spring 2000), burned the previous sea- son, or not burned for at least 5 years. Treat- ments did not cause differences in species richness, and caused only small changes in spe- cies composition. The equativity of life-forms showed a tendency to decrease with fire sup- pression, with woody species becoming more 3523_book.fm Page 97 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC 98 Land Use Change and Mountain Biodiversity dominant in plots that were not recently burned. Total biomass and the proportions of live bio- mass, standing dead, and litter varied among treatments. Fire caused a reduction in total bio- mass, but increased the contribution of live bio- mass and encouraged the development of more palatable growth forms (mainly grasses). Patches that were burned during the ongoing growing season had 80% of their biomass as live biomass in December and January. In these months, livestock are moved from forests at lower altitudinal levels to these highland grass- lands. This modification in the seasonal dynam- ics of aerial biomass may represent a substantial change in the pattern of forage availability, especially at this time of the year. ACKNOWLEDGMENTS We are grateful to phytogeography students of Universidad Nacional de Tucumán for the assis- tance during fieldwork. The manuscript bene- fited from suggestions from three anonymous reviewers and from colleagues from LIEY. International Foundation for Science and Fun- dación PROYUNGAS provided financial sup- port for this study. References Belsky, A. (1992). Effects of grazing, competition, disturbance and fire on species composition and diversity in grassland communities. Journal of Vegetation Science, 3: 187–200. Bianchi, A. (1981). Las precipitaciones del Noroeste Argentino. INTA, Salta, Argentina. Brown, A.D., Grau, H.R., Malizia, L.R., and Grau, A. (2001). Argentina. In Kappelle, M. and Brown, A.D. (Eds.), Bosques Nublados del Neotropico. INBIO, Costa Rica. Cabrera, A. (1976). Regiones Fitogeográficas Argen- tinas. ACME, Buenos Aires, Argentina. Collins, S. (1987). Interaction of disturbances in tallgrass prairie: a field experiment. Ecology, 68: 1243–1250. Collins, S., Knapp, A., Briggs, J., Blair, J., and Stein- auer, E. (1998). Modulation of diversity by grazing and mowing in native tallgrass prai- rie. Science, 280: 745–747. Coppock, D.L. and Detling, J.K. (1986). Alteration of bison and black-tailed prairie dog grazing interaction by prescribed burning. Journal of Wildlife Management 50:452–455. Crutzen, P.J., Delany, A.C., Greenberg, J., Haagen- son, P., Heidt, L., Lueb, R., Pollock, W., Seiler, W., Wartburg, A., and Zimmerman, P. (1985) Tropospheric chemical composition measurement in Brazil during the dry sea- son. Journal of Atmospheric Chemistry 2:233–256. De Baro, L., Neary, D., and Folliot, P. (1998). Fire Effects on Ecosystems. John Wiley & Sons, New York. Diaz, S., Acosta, A., and Cabido, M. (1992). Mor- phological analysis of herbaceous commu- nities under different grazing regimes. Journal of Vegetation Science, 3: 689–696. Eckholm, E.P. (1975). The deterioration of mountain environments. Science, 189: 764–770. Grau, H., Gil-Montero, R., Villalba, R., Carilla, J., Araoz, E., Masse, G., and Membiela M. Submitted. Environmental history and forest regeneration dynamics in a degraded valley of NW Argentina cloud forests. In Juvik, J., Bruijnzeel, S., and Scatenna, F. (Eds.). Fo r- ests in the Mists, Ecology and Conservation of Tropical Montane Cloud Forests. Univer- sity of Hawaii Press. Grau, A. and Brown, D. (2000). Development threats to biodiversity and opportunities for conser- vation in the mountain ranges of the upper Bermejo River Basin, NW Argentina and SW Bolivia. Ambio, 29: 445–450. Grau, H. (2005). Dinámica de bosques en el gradi- ente altitudinal de las yungas. In Arturi, M.F., Frangi, J., and Goya, J.L. (Eds). Ecología y Manejo de los bosques naturales Argentinos. Univ. EDULP (Ediciones de la Universidad Nacional de la Plata). Argen- tina. Hobbs, T., Schimel, D., Owensby, C., and Ojima, D. (1991). Fire and grazing in the tallgrass prai- rie: contingent effects on nitrogen budgets. Ecology, 72: 1374–1382. Hofstede, R., Mondragón, C., and Rocha, C. (1995). Biomass of grazed, burned and undisturbed Páramo grasslands, Colombia. Arctic and Alpine Research, 27: 1–12. Lloret, F. and Vila, M. (2003). Diversity patterns of plant functional types in relation to fire regime and previous land use in Mediterra- nean woodlands. Journal of Vegetation Sci- ence, 14: 387–398. 3523_book.fm Page 98 Tuesday, November 22, 2005 11:23 AM Copyright © 2006 Taylor & Francis Group, LLC [...]... extensive livestock systems in páramo In Baslev, H and Luteyn, J.L (eds.) Páramo: An Andean EcoSystem Under Human Influence Academic Press, London Vogl, R.J., (1 974 ) Effects of fire on grasslands In Kozlowski, T.T and Ahlgren, C.E (Eds.) Fire and Ecosystems Academic Press, New York Valone, T and Kelt, D (1999) Fire and grazing in shrub-invaded arid grassland community: independent or interactive ecological... In Brown, A and Grau, H (Eds.), Investigación, Conservación y Desarrollo en Selvas Subtropicales de Montaña Proyecto de Desarrollo Agroforestal, LIEY, Tucumán, Argentina Rusch, G and Oesterheld, M (19 97) Relationship between productivity and species and functional group diversity in grazed and nongrazed Pampas grassland Oikos, 78 : 519–526 Schmidt, A.M and Verweij, P.A (1992) Forage intake and secondary... M.G., Zygmunt, B., Christensen, S.W., Dale, V.H., and Graham, R.L (1988) Indices of landscape pattern Landscape Ecology, 1: 153–162 Pucheta, E., Vendramini, F., Cabido, M., and Diaz, S (1998) Estructura y funcionamiento de un pastizal de montaña bajo pastoreo y su respuesta luego de la exclusión Revista de la Facultad de Agronomía, La Plata, 103: 77 –92 Copyright © 2006 Taylor & Francis Group, LLC 99...3523_book.fm Page 99 Tuesday, November 22, 2005 11:23 AM Fire, Plant Species Richness and Biomass in Mountain Grasslands of NW Argentina Malizia, A (2003) Host tree preference of vascular epiphytes and climbers in a subtropical montane cloud forest of Northwest Argentina Selbyana, 24: 196–205 McCune, B and Mefford, M.J (19 97) PC-ORD Multivariate Analysis of Ecological Data Version 3 0 MjM Software Design,... Serengeti Ecological Monographs, 55: 259–294 Milchunas, D and Lauenroth, W (1993) Quantitative effects of grazing on vegetation and soils over a global range of environments Ecological Monographs, 63: 3 27 366 Oesterheld, M., Loreti, J., Semmartin, M., and Paruelo, J (1999) Grazing, fire, and climate effects on primary productivity of grasslands and savannas In Walker, L (Ed.), Ecosystems of disturbed . impor- tant land uses in many montane grasslands (Eckholm 1 975 ). Particularly in Andean grass- lands, extensive cattle grazing is often com- bined with burning of the natural vegetation (Schmidt and. 19. 27 73.15 ± 18.21 38 .72 ± 7. 03 1 67. 80 ± 19.38 10.80 0.002 >1 year Live 391.53 ± 66.42 288.44 ± 44.93 3 27. 87 ± 25.26 445.95 ± 30 .70 184 .76 ± 50.86 13.22 0.001 Standing dead 196.02 ± 47. 97 159.19. & Francis Group, LLC 90 Land Use Change and Mountain Biodiversity mountain grasslands are scarcely represented in the protected areas of Argentina, and little is known about their functioning.