Macrophytic, epipelic and epilithic primary production in a semiarid Mediterranean stream

Introduction

Important ecological questions concerning the energy flow in stream ecosystems require estimates of algal primary production (Morin, Lamoureux & Busnarda, 1999). Such estimates of primary production have been made in lotic systems across biomes and latitudes in North America (Bott et al., 1985; Mulholland et al., 2001) and in most regions of the world (Lamberti & Steinman, 1997). However, primary production studies in Mediterranean-climate regions (areas surrounding the Mediterranean Sea, parts of western North America, parts of west and south Australia, southwestern South Africa and parts of central Chile) are largely lacking. Rainfall seasonality and variability are the principal attributes of the Mediterranean-type climate, most precipitation resulting from a few major storm events that may produce flooding (Gasith & Resh, 1999). Annual precipitation is highly variable among Mediterranean areas, usually ranging between 275 and 900 mm, while certain Mediterranean-climate regions fall into the category of semiarid regions (between 200 and 500 mm; UNESCO, 1979).

The importance of primary production in Mediterranean streams is generally expected to increase with increasing gradients of aridity (Gasith & Resh, 1999), following the pattern found in an analysis of 30 streams over a large geographic range (Lamberti & Steinman, 1997), where GPP increased significantly with mean temperature and declined with total precipitation. To our knowledge, there are few data on primary production in Mediterranean streams. Four streams in the Spanish Mediterranean area have been studied: La Solana and Riera Major, in the NE of the Iberian Peninsula with 800–1000 mm mean annual precipitation (Guasch & Sabater, 1994, 1998); Montesina stream, in the south (Mollá, Malchik & Casado, 1994) with 527 mm of mean annual precipitation; and Chicamo stream (Suárez & Vidal-Abarca, 2000), located in southeastern Spain, with a semiarid climate (250 mm mean annual precipitation). Early estimates of primary production and respiration in Chicamo stream (Suárez & Vidal-Abarca, 2000) were higher than those reported in the other Mediterranean streams of wetter areas, but were similar to those obtained in the warm desert stream, Sycamore Creek (Busch & Fisher, 1981; Grimm & Fisher (1984).

Although autotrophy may be important in arid and semiarid streams, the patchiness of aquatic vegetation, its seasonal growth pattern and the contribution of periphyton and macrophytes to autochthonous sources during the year may differ. The objectives of this paper were to determine the contribution of different algal assemblages to annual primary production in a semiarid Mediterranean stream (Chicamo stream), to confirm its autotrophic metabolism, to identify climatic and environmental variables that might govern primary production in this stream, and to compare metabolic rates with those of other Mediterranean and desert streams.

Study area

Chicamo stream is a saline and intermittent tributary of the Segura River located in southeast Spain (38°N latitude), in the most arid area of the province of Murcia. The climate is characterised by a mean annual precipitation below 300 mm and a mean annual temperature of 18°C. A long warm and dry summer season is interrupted by spring and autumn rains, the latter followed by a short and temperate winter. Annual and interannual variations in rainfall can lead to a high discharge variability, with extreme conditions of flooding and drying. The highest flows normally occur in early autumn, followed by a second peak in spring.

Chicamo stream is a 4th-order stream, draining a sedimentary watershed of 502 km². Surface water flow is intermittent in 10 km of its 59.4 km total length. The natural cover on the watershed is open Mediterranean scrub, although much is dedicated to citrus and horticultural crops.

Studies were conducted in a permanent, braided, shallow and unshaded reach, 100 m length, typical of the middle section. The substrate consisted of consolidated and impermeable marls, with deposits of gravels and sand in the erosional zones of runs, and silts in pools and depositional zones. Chicamo stream water is hyposaline and hard (7.6 g L⁻¹ mean salinity, 219.5 mg L⁻¹ mean alkalinity), well oxygenated and rich in nutrients, especially nitrates and ammonium (Vidal-Abarca et al., 2000). Riparian vegetation is sparse because of frequent floods. There are no trees, but isolated shrubs such as Phragmites australis, Tamarix canariensis and Juncus maritimus are present.

Aquatic primary producers in the study site include the macrophyte Chara vulgaris (with epiphytic algae) in pools, an extensive diatom assemblage on fine sediments, including Nitzschia, Amphora, Navicula, Gyrosigma and Pleurosigma among the most abundant genera (D. Ros, per. com.), and epilithic periphyton dominated by the cyanobacteria Calothrix and the chlorophyte Kentrosphaera facciolae, although in spring the filamentous green alga Cladophora glomerata predominates.

Fine benthic particulate organic matter (FPOM) is the principal benthic fraction (about 70% of the total) while coarse particulate organic matter (CPOM) represents only 6% of the total because of the scarcity of riparian vegetation and the consequent low input of allochthonous materials (Martínez et al., 1998). Dissolved organic carbon (DOC) is the principal organic carbon source flowing in the Chicamo stream, particularly when flash-floods occur (Vidal-Abarca et al., 2001).

Methods

Metabolism experiments were conducted six times between July 1998 and May 1999 to record the annual variability. Macrophyte, epipelic and epilithic primary production and community respiration were measured using oxygen change in clear plastic chambers (25.5 × 13.8 × 8 cm) placed in the stream. Because of the low discharge and current in the stream, we did not use a pump for water recirculation. Chamber temperature never exceeded stream temperature by more than 3 °C. Four replicate estimates of production and respiration were obtained from each assemblage.

The macrophyte, C. vulgaris, and epipelic samples were collected 15 days before the experiment, tranferred to plastic trays (21.5 × 12 × 2 cm) and placed on the streambed contiguous with the sediment surface for colonisation to continue. At the start of the experiment, the trays were removed from the stream and transferred to the chambers with minimal disturbance to the community during removal. For the epilithon, the stones used for each incubation were collected from the stream immediately before the experiment started.

All the chambers were filled with stream water, closed and placed in situ. Each measurement of net photosynthetic rate was made over a 3-h period (2 h dark and 1 h light) during the morning. Chambers for dark incubations were covered with aluminium foil. Changes in O₂ concentration were detected using standard Winkler titration scaled to 10 mL sample sizes taken with a needle and syringe. Winkler determinations of dissolved oxygen (DO) were made before and after the dark incubations, and after the light incubations. Water temperature in the stream and in each chamber was recorded at each sampling time.

Gross primary production (GPP) was calculated by adding the change in DO measured during the dark incubation to the change measured during the light incubation. Because respiration rates include the metabolism of heterotrophs such as microbes and insects as well as autotrophs, respiration is termed community respiration (CR). Oxygen generation was converted to carbon fixation assuming a photosynthetic quotient of 1.2 (Bott, 1997a).

After the incubation experiments, C. vulgaris was removed from the chambers and stored on ice until it was washed with tap water in the laboratory and weighed. Three samples of 3 g of fresh macrophyte with its epiphitic algae were taken from each chamber to determine chlorophyll a (Chla) concentration by spectrophotometry, following extraction in boiling 90% ethanol as described by Biggs (1995). Concentrations were converted to algal carbon biomass using a factor of 30, which is suitable for communities in nutrient-rich, unshaded environments (Vollenweider, 1974; Bott et al., 1997). The rest of the macrophyte was dried at 60 °C to constant dry weight (DW) and then ashed at 450 °C for 4 h to estimate the content of ash free dry weight (AFDW) in each chamber. Triplicate sediment cores of 2.7 cm diameter were also collected from each epipelic chamber for Chla and AFDW analyses. The rocks were removed and returned to the laboratory for determination of surface area and Chla and AFDW content. Rock surface area was determined by tracing the outline of the top of the rock on aluminium paper, cutting out the shape, and comparing its weight with the weight of a known area of paper. Periphyton on rocks was scraped, brushed and washed from the rock surface. The extracts were filtered onto Whatman A filters for Chla concentration and AFDW determinations. Autochthonous detritus standing crop was determined as the difference between AFDW measurements (assuming a 50% carbon content) and the periphyton biomass obtained from Chla concentrations.

On each date, the area covered by each primary producer assemblage in the stream was measured in the study reach (100 m length) to extrapolate the metabolism rates estimated for each assemblage to the reach. Measurements of discharge, photosynthetically active radiation (PAR), air and water temperatures, conductivity, salinity and dissolved oxygen were made in the stream during the experiments. Two water samples were taken to determine suspended solids, ammonium, nitrate and soluble reactive phosphorus (SRP) concentration on each date. The samples were kept cool until their arrival at the laboratory, where they were filtered onto preashed and preweighed GF/F glass-fiber filters (Whatman) and oven-dried at 60 °C to constant dry weight. Nutrients were determined according to standard methods (American Public Health Association (APHA), 1992): nitrate by the cadmium reduction method, nitrites by sulfanylic acid colorimetry, ammonium by phenol nitroprussiate colorimetry and phosphate by ascorbic acid colorimetry.

The days on which the experiments were performed were sunny with mean PAR values during the experiments of more than 1000 μE m⁻² s⁻¹, except in August (914.78 μE m⁻² s⁻¹) and December (595.94 μE m⁻² s⁻¹).

At the stream reach scale, we estimated the net daily metabolism (NDM) and the GPPd⁻¹/CR₂₄ ratio by patch-weighting the assemblage-level metabolism values. Chamber measurements of community respiration were extrapolated to a 24-h period, assuming the respiratory rate during the night to be half daylight respiration because of the decrease in temperature:

and

Differences in gross primary production and respiration between primary producer assemblages were analysed using analysis of variance (one way anova) and Tukey's honest significant difference test (HSD). All data were log(x + 1) transformed to normalise distributions and equalise variance. Relationships of climatic and environmental variables to metabolic rates were examined using bivariate correlation and multiple linear regression approaches. Pearson correlation analysis was used to indentify relationships between single factors and metabolic rates. Finally, a stepwise multiple linear regression used all significant (P ≤ 0.05) correlated factors and the response variables (GPP and CR). Statistical analyses were performed using Systat (Wilkinson, 1996).

Results

Climatic and environmental conditions

Table 1 shows the mean values of physical and chemical parameters measured for the six experiment dates in Chicamo stream, together with monthly air temperature and precipitation recorded in a nearby metereological station. Total precipitation during the study period (from July 1998 to May 1999) was 143.5 mm, mainly concentrated in autumn and winter peaks. Discharge variation during the study period was closely related to daily precipitation. Discharge was very low (≤ 3.5 L s⁻¹) except following rainfall in December and March. The mean annual temperature was 18.2 °C with August being the hottest month, when the maximum temperature reached over 32 °C, while December was the coldest month, with the minimum recorded temperature being 4 °C. The mean water temperature was 20 °C (max = 30 °C in August; min = 7.5 °C in December).

Areal coverage and biomass of algal assemblages

Epipelic algae were the most widespread assemblage throughout the study period, representing more than 70% of the total wetted surface (Table 2). The coverage by C. vulgaris was peaked in October, though subsequent high winter flows buried it with sediments and reduced it to small patches. In March, the macrophyte was absent, but it started to grow again in May. Maximum epilithon cover was attained in December and minimum in October. C. vulgaris cover was negatively correlated with discharge and monthly precipitation (r = −0.92, P ≤ 0.005 and r = −0.84, P ≤ 0.05) while it was positively correlated with ammonium (r = 0.85, P ≤ 0.02). Epilithic algal cover showed a positive correlation with discharge and precipitation (r = 0.95, P ≤ 0.001; r = 0.89, P ≤ 0.005, respectively) and a negative corrrelation with conductivity (r = −0.76, P ≤ 0.05). The epipelic algal cover showed a negative correlation with the dissolved oxygen (r = −0.78, P ≤ 0.05).

Mean annual biomass values for macrophyte, epipelic and epilithic assemblages were 25.26, 5.05 and 4.62 gC m⁻², respectively (Table 2). C. vulgaris registered the highest biomass values on all the dates except in March when it was absent. Its biomass peak occurred in December, and the minimum in summer. It was positively correlated with the standing crop of detritus and ammonium concentration (r = 0.96, P ≤ 0.001 and r = 0.80, P ≤ 0.05, respectively).

The epilithic assemblage registered its highest biomass value in March, when C. glomerata was the dominant alga. A second peak of epilithon biomass occurred in summer, when the dominant alga was the cyanobacterium Calothrix. Epilithon biomass also showed a positive correlation with the standing crop of detritus (r = 0.88, P ≤ 0.01). The peak of epipelon biomass was reached in spring. Epipelon biomass was positively correlated with alkalinity (r = 0.83, P ≤ 0.05).

At the reach level, epipelic biomass accounted for the highest portion of biomass on all dates, except in October when C. vulgaris represented 76% of the total biomass of the reach. Differences in algal biomass between assemblages were significant on all dates (Table 3).

Gross primary production and respiration rates

The mean gross primary production rate was 2.05 ± 0.42 gC m⁻² h⁻¹ and the mean respiration rate 0.52 ± 0.12 gC m⁻² h⁻¹, resulting in a mean net primary production rate of 1.53 ± 0.31 gC m⁻² h⁻¹. Respiration represented 25.36% of mean gross primary production.

At the assemblage level, C. vulgaris showed the highest mean metabolic rates (Table 2), followed by the epilithic and the epipelic assemblages. Summer and spring were the most productive seasons (Fig. 1a). The maximum values measured corresponded to epilithon in July (5.46 gC m⁻² h⁻¹) and C. vulgaris in May (5.01 gC m⁻² h⁻¹).

Gross primary production rates showed highly significant differences between assemblages, except in December (Table 3). Macrophyte (Chara) production exceeded epilithic and epipelic production for most of the year (Fig. 1a), significantly so in August, October and May (Table 3). Epipelic and epilithic production rates were similar except in July, when epilithic production reached its highest value (Fig. 1a).

Respiration and gross primary production were highly correlated for all the assemblages (r = 0.9, P ≤ 0.005 for C. vulgaris; r = 0.83, P ≤ 0.02 for epipelic algae, and r = 0.97, P ≤ 0.001, for epilithic algae). However, no significant correlations were found for any algal assemblage between metabolic rates and biomass.

Community respiration rates followed a similar pattern to the gross primary production rates (Fig. 1b). The highest respiration rates were also found in spring and summer, and the lowest in December. Significant differences were found between assemblages except in October and December (Table 3). C. vulgaris respiration was significantly greater than that of the periphyton assemblages in August and May. Differences between epipelic and epilithic respiration were not significant, except in July.

Examination of the GPP/Chla ratio throughout the year (Fig. 1c) showed the most photosyntethic efficient assemblage to be the epilithon (mean 37.44 gC g⁻¹ Chla h⁻¹) which reached its highest value in May (137.59 gC g⁻¹ Chla h⁻¹). Significant negative correlation was found between GPP/Chla and epilithon biomass. The epipelic algae and C. vulgaris were less efficient, although they showed greater efficiency in summer than during the rest of the year.

The mean annual GPP and CR of the reach studied were 285.19 ± 72.24 and 76.51 ± 20.12 gC h⁻¹, respectively, with the mean annual NPP being 208.68 ± 52.12 gC h⁻¹. At the reach level, the epipelic assemblage was the principal contributor to autochthonous production in the stream, representing 84% of total production and 86% of total respiration (Table 2). Its contribution to GPP varied during the year from 71.23% in December to 96.83% in March. C. vulgaris and epilithon production represented only 8.96 and 7.05% of the total, although the contribution of the macrophyte and its epiphitic algae was more important in summer and autumn, while the epilithon showed a maximum in December. Patch respiration followed a similar pattern to patch GGP.

Relationships between metabolic rates and climatic and environmental variables

The primary production of C. vulgaris showed a positive correlation with ammonium (r = 0.92, P ≤ 0.005), standing crop of detritus (0.82, P ≤ 0.01), and mean monthly air temperature (r = 0.76, P ≤ 0.05), and was negatively correlated with monthly precipitation (r = −0.82, P ≤ 0.05). Ammonium concentration was the best single predictor of GPP in the simple regression analysis, explaining 81% of its variation, although the best multiple regression model included detritus standing crop and monthly air temperature, which explained 92% of the variation (Table 4).

Table 4. Results of stepwise multiple regression analysis for rates of gross primary production (GPP) and community respiration (CR) for each algal assemblage and for mean assemblage rates (n = 6 for each regression)

Community	Dependent variable	Independent variable	r	P
C. vulgaris	GPP	Detritus standing crop	0.582	0.0146
		Mean monthly air temperature	0.342	0.0223
		Full model	0.924	0.0097
	CR	Mean monthly air temperature	0.507	0.0683
Epipelic	GPP	Mean monthly air temperature	0.629	0.0095
		Nitrate		0.0016
		Ammonium		0.0034
		PAR		0.0041
		Full model	0.999	0.0018
	CR	Mean monthly air temperature	0.980	0.0299
		Dissolved oxygen		0.0279
		Conductivity		0.0429
		Full model	0.999	0.0066
Epilithon	GPP	Suspended solids	0.690	0.0253
	CR	Suspended solids	0.753	0.0157
Mean assemblages	GPP	Mean monthly air temperature	0.638	0.0352
	CR	Mean monthly air temperature	0.712	0.0018
		Conductivity		0.0032
		PAR		0.0053
		Nitrates		0.0219
		Full model	0.999	0.0029

The respiration rate of C. vulgaris was positively correlated with mean monthly air temperature (r = 0.78, P ≤ 0.05), and negatively with dissolved oxygen concentration (r = −0.76, P ≤ 0.05) and monthly precipitation (r = −0.77, P ≤ 0.05). Mean monthly air temperature was the only significant predictor of macrophyte community respiration, explaining 51% of its variation (Table 4).

Epipelic production and respiration were also positively correlated with mean monthly air temperature (r = 0.84, P ≤ 0.05, and r = 0.99, P ≤ 0.001, respectively) and water temperature (r = 0.78, P ≤ 0.05, and r = 0.98, P ≤ 0.001, respectively). Respiration rate was also positively correlated with conductivity (r = 0.92, P ≤ 0.005) and negatively with monthly precipitation (r = −0.89, P ≤ 0.005), discharge (r = −0.77, P ≤ 0.05) and dissolved oxygen (r = −0.766, P ≤ 0.05). Mean monthly air temperature was the best single predictor of the epipelic production and respiration, explaining 63 and 98%, respectively, of its variation. Multiple regression analyses provided a predictive model of GPP that included, besides monthly air temperature, nitrate, ammonium concentration and PAR as predictor variables, accounting for 99.99% of the variation (Table 4). The best multiple regression model for epipelic respiration (r = 99.99%) included monthly air temperature, dissolved oxygen and conductivity as independent variables (Table 4).

However, epilithon GPP and CR only showed a positive correlation with the suspended solid concentration (r = 0.87, P ≤ 0.05, and r = 0.78, P ≤ 0.05, respectively). This independient variable explained 69% of epilithon production and 75% of variation in its respiration rate, and it was the only predictor variable in the multiple regression analysis (Table 4).

In general, the mean rates of gross primary production and respiration of algal assemblages were positively correlated (P ≤ 0.05) with mean monthly air temperature (r = 0.88 and r = 0. 87, respectively), and respiration was also negatively correlated with monthly precipitation (r = −0.79) and dissolved oxygen (r = −0.83). The results obtained by simple regression analysis using the mean values for the assemblages as a whole, indicated that 64% of the variation in mean GPP and 71 % of the variation of mean CR could be explained by the mean monthly air temperature alone. For mean CR, a significant multiple regression model including, besides mean monthly air temperature, conductivity, PAR and nitrate concentration, explained 99.99% of its variation (Table 4).

Net ecosystem metabolism and GPP : CR₂₄ ratio

Daily net metabolism in the reach was positive on all the dates studied, ranging from 2.23 gC m⁻² day⁻¹ in March to 16.61 gC m⁻² day⁻¹ in July, with a mean value of 7.03 gC m⁻² day⁻¹ (Table 2).

The maximum rate of daily GPP was 26.41 gC m⁻² day⁻¹ in July, and the minimum 5.15 gC m⁻² day⁻¹ in December. The rates of daily benthic respiration were much lower than those of GPP, with values ranging from 2.67 to 10.22 gC m⁻² day⁻¹. Mean daily GPP and CR₂₄ were 13.70 and 6.88 gC m⁻² day⁻¹, respectively. The GPP : CR₂₄ ratios on all dates were higher than 1, with a mean value of 1.9, a maximum value of 2.7 being reached in July.

Discussion

The exceptionally high rates of primary production found in Chicamo stream are only possible as a result of high temperatures, high light availability and the intensive internal recycling of nutrients. Mean monthly air temperature was the best single predictor of mean GPP and mean CR in Chicamo stream, explaining more than 63 and 71% of their variation, respectively. Morin et al. (1999), using multiple regression models fitted to predict the primary production of stream periphyton, found that the production of algae increased with both Chla standing crop and water temperature, although Chla was most strongly correlated with primary production, accounting for 65% of the variation, while temperature accounted for a much smaller portion (4%) of the variability. Other comparative studies of streams from different biomes found that temperature explained 33 (Bott et al., 1985) and 38% (Sinsabaugh, 1997) of the variation in R, although Mulholland et al. (2001) found no evidence for any effect of water temperature on R or GPP. On a local-scale, Uehlinger, Konig & Reichert (2000) during one annual cycle of a Swiss river, found that R was significantly related to water temperature, although temperature explained only 22% of its variation. Temperature interacts with other environmental factors, such as nutrients, gases, metabolites, development stage, trophic interactions and, especially, light, although the individual effect of each factor is not clear (DeNicola, 1996). In Chicamo stream, high temperatures were associated with high irradiance, low or null precipitation, low discharge, high conductivity, low dissolved oxygen levels and high ammonium concentrations.

Although photoinhibition usually ocurrs at irradiances greater than 600 μmol m⁻² s⁻¹ (Hill, 1996), in our study, exposure of the experimental chambers to higher natural light levels did not appear to cause any photoinhibition. Regarding photosynthesis-irradiance responses, benthic algae inhabiting “high-light” environments (such as open sites in clear streams) are probably under considerable selective pressure to develop mechanisms that reduce the potentially damaging effects of high irradiances, whether in the form of accessory pigments (carotenids) or sheath pigments such as scytonemin (Garcia-Pichel & Castenholz, 1991). On the other hand, high irradiance at the surface of benthic algal mats may inhibit photosynthesis by surface cells, while shaded subsurface cells receive only saturating or subsaturating levels of irradiance. Photosynthesis by underlayers at high surface irradiance may compensate for inhibition in surface layers, so there is no evidence of photoinhibition at community level (Hill, 1996).

In our study, PAR was a secondary predictor variable of both epipelic GPP and mean CR in the multiple regression models. Of the nutrients examined, only the ammonium concentration was significantly related to C. vulgaris GPP and the nitrate and ammonium concentration to epipelic GPP.

Measurements of GPP and CR showed within-year differences between assemblages that reflected the phenological characteristics of the primary producers studied in relation to seasonal variations of temperature and precipitation. In addition, proximate factors such as resource availability, ecophysiology, life history characteristics, stress, allelophathy, competition and predation, are causes of differential species performance (Stevenson, 1996). In unshaded streams, the flood disturbance regime is perhaps the fundamental factor determining habitat suitability and the spatial and temporal patterns of benthic algae, and is a major biomass loss mechanism (Biggs, 1996). Floods in the Chicamo stream increase the habitat suitability for epilithon and nutrient availability in the water column, but have a negative effect on the production of C. vulgaris which is buried by transported sediment. Although epipelic algae biomass was largely removed by flooding, recolonisation was rapid and biomass and production values returned to predisturbance levels in less than one month. The availability of algal propagules and rapid growth rates make epipelic algae resilient in the face of flooding.

In Chicamo stream, the greater extent of epipelic assemblages, the high abundance of gathering collectors and the low diversity of algal grazers (Martínez et al. 1998) suggest that neither the macrophyte C. vulgaris nor epilithic algae are an important food source, although epipelic algae with detritus derived from sloughing periphyton are. High rates of primary production in Chicamo stream support correspondingly high rates of secondary production of aquatic macroinvertebrates. The annual production estimates of the gathering-collector Caenis luctuosa presented the highest value ever reported for Caenis species (6.35 gDW m⁻² year⁻¹; Perán, Velasco & Millán, 1999).

A comparison of the primary production values obtained in this study with those reported for other desert and Mediterranean streams (Table 5) shows that our rates of GPP were higher than those reported earlier from other streams, including desert streams which had higher GPP than a variety of streams from different biomes and locations throughout the world (Naiman, 1976; Lamberti & Steinman, 1997; Mulholland et al., 2001). The expected pattern of increasing GPP with increasing mean annual temperature and decreasing total precipitation was observed (Fig. 2), as found by Lamberti & Steinman (1997) in a multi-stream comparison study.

Net metabolism measurements in the Mediterranean streams also point to differences in the degree of autotrophy as aridity increases. Autotrophic metabolism prevailed throughout the year in Chicamo stream, the most arid stream, but only during some periods of the year (principally spring) in the rest of the streams (Guasch & Sabater, 1994, 1998; Molláet al., 1994). The mean net metabolism in Chicamo was 7.03 gC m⁻² day⁻¹, resulting in the stream being a net producer of organic carbon.

Early estimates of daily GPP and CR in Chicamo stream made by the one station-open system method (Suárez & Vidal-Abarca, 2000) were lower than our estimates made using the chamber method, because the former only reflects a particular local community. The chamber method employed in our study provides a better estimate of stream metabolism because the contribution of the different algal assemblages present is recorded and can be extrapolated to the total reach. Bott et al. (1978), in a comparison of methods for measuring primary productivity and community respiration in streams, found that open methods provided lower NPP estimates than the chamber methods. Therefore, inter and intra stream differences should be interpreted with caution because comparative analyses of primary production are difficult because of the large differences in analytical methods, conversion factors, communities and sampling frequencies employed (Wetzel & Ward, 1992).

In conclusion, Chicamo stream is a net producer of organic carbon, with autotrophic metabolism throughout the year, epipelic algae being the most important primary producer in the stream. Temperature dominates the control of mean GPP and mean CR in the stream, perhaps because it integrates many features relevant to primary production. In addition, the frequency and intensity of flood disturbance has a considerable influence on the biomass and production of different algal assemblages. Nutrient concentrations, especially of ammonium and nitrates, appeared to be a secondary determinant of GPP.

Thus, Chicamo stream resembles desert and prairie streams as regards the dominant role of autochthonous organic matter as the base of the food web.