Urbanization and agriculture increase exports and differentially alter elemental stoichiometry of dissolved organic matter (DOM) from tropical catchments

Science of the Total Environment 550 (2016) 785–792

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Urbanization and agriculture increase exports and differentially alter elemental stoichiometry of dissolved organic matter (DOM) from tropical catchments Björn Gücker a,⁎, Ricky C.S. Silva a, Daniel Graeber b, José A.F. Monteiro a, Iola G. Boëchat a a b

Applied Limnology Laboratory, Federal University of São João del-Rei, São João del-Rei, Brazil Department of Bioscience, Aarhus University, Denmark

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• We investigated land-use impact by comparing human-impacted with natural catchments. • Pasture catchments had lower stream DOM concentration, with lower C:N and C:P. • Agricultural catchments had higher DOM export, with lower C:P. • Urban catchments had higher concentration and export, with lower C:N and higher C:P. • Urbanization exerted the strongest impacts and should be a management priority.

a r t i c l e

i n f o

Article history: Received 11 August 2015 Received in revised form 24 January 2016 Accepted 24 January 2016 Available online xxxx Editor: D. Barcelo Keywords: DOM export Land use Pasture Agriculture Urbanization Organic carbon Elemental stoichiometry

a b s t r a c t Many tropical biomes are threatened by rapid land-use change, but its catchment-wide biogeochemical effects are poorly understood. The few previous studies on DOM in tropical catchments suggest that deforestation and subsequent land use increase stream water dissolved organic carbon (DOC) concentrations, but consistent effects on DOM elemental stoichiometry have not yet been reported. Here, we studied stream water DOC concentrations, catchment DOC exports, and DOM elemental stoichiometry in 20 tropical catchments at the Cerrado–Atlantic rainforest transition, dominated by natural vegetation, pasture, intensive agriculture, and urban land cover. Streams draining pasture could be distinguished from those draining natural catchments by their lower DOC concentrations, with lower DOM C:N and C:P ratios. Catchments with intensive agriculture had higher DOC exports and lower DOM C:P ratios than natural catchments. Finally, with the highest DOC concentrations and exports, as well as the highest DOM C:P and N:P ratios, but the lowest C:N ratios among all land-use types, urbanized catchments had the strongest effects on catchment DOM. Thus, urbanization may have alleviated N limitation of heterotrophic DOM decomposition, but increased P limitation. Land use—especially urbanization—also affected the seasonality of catchment biogeochemistry. While natural catchments exhibited high DOC exports and concentrations, with high DOM C:P ratios in the rainy season only, urbanized catchments had high values in these variables throughout the year. Our results suggest that urbanization and pastoral land use exerted the strongest impacts on

⁎ Corresponding author at: Applied Limnology Laboratory, Department of Geosciences, Campus Tancredo Neves, Federal University of São João del-Rei, 36307-352 São João del-Rei, Minas Gerais, Brazil. E-mail address: [email protected] (B. Gücker).

http://dx.doi.org/10.1016/j.scitotenv.2016.01.158 0048-9697/© 2016 Elsevier B.V. All rights reserved.

786

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

Neotropical catchments Cerrado savanna Atlantic forest Rivers Streams

DOM biogeochemistry in the investigated tropical catchments and should thus be important targets for management and mitigation efforts. © 2016 Elsevier B.V. All rights reserved.

1. Introduction DOM absorbs biologically harmful ultraviolet as well as photosynthetically active radiation, eliminates toxicants, interferes with water treatment for human consumption, and is an important energy and nutrient source for aquatic microorganisms (Findlay and Sinsabaugh, 2003; Kim and Yu, 2005). Thus, DOM plays a key-role in regulating river biological processes and ecosystem services (Findlay and Sinsabaugh, 2003). Moreover, dissolved organic carbon (DOC, the main component of DOM) is considered to be the most important component of the soil-derived C flux into inland waters, currently estimated to 1.9 ± 1.0 Pg C yr−1, and is thus an important component in the global carbon cycle (Regnier et al., 2013). Terrestrial sources, i.e. riparian litter fall and soils of headwater catchments, contribute considerably to riverine DOM, both in temperate and tropical regions (Billett et al., 2006; Morel et al., 2009; Wiegner and Tubal, 2010; Zurbrügg et al., 2013). DOC concentrations of natural tropical rivers are often thought to be low compared to temperate rivers, and typically range between 3 and 10 mg L−1 (Lewis et al., 1995), but can be much higher in tropical blackwater streams (McClain et al., 1997). Concentrations of DOC are highest on the rising limbs of hydrographs in some tropical rivers, but independent of hydrology in others (Lewis et al., 1986; Lewis and Saunders, 1989; Saunders and Lewis, 1989). Tropical forests can be an important source of labile DOM, such as amino acids, to rivers, fueling high rates of heterotrophic metabolism, but the DOM composition of tropical rivers is also poorly characterized, in terms of both biochemical composition and elemental stoichiometry (Jaffé et al., 2012). In the few studies that characterized DOM quality in tropical rivers, it appeared to be related to vegetation cover (Yamashita et al., 2010), and deforestation or changes to invasive riparian vegetation are expected to increase DOM bioavailability (Wiegner and Tubal, 2010; Yamashita et al., 2010). Moreover, DOM exported from tropical catchments can be relatively rich in N (Jaffé et al., 2012). Imbalances in elementary stoichiometry between available DOM and its heterotrophic microbial consumers can limit microbial metabolism and thus DOM decomposition in aquatic systems (Sterner and Elser, 2002; Hessen et al., 2004; Cross et al., 2005). Therefore, high N contents of DOM in some tropical regions may contribute considerably to river metabolism. Heterotrophic bacteria can also supplement nutrient deficiency in DOM by using dissolved inorganic nutrients, and bacterial metabolism can be C limited or C-nutrient co-limited in larger tropical rivers (Benner et al., 1995). Thus, C biochemical characteristics of DOM, such as the abundance of alkyl-C and aromatic-C (Jaffé et al., 2012), may become important for microbial metabolism in these systems. A better understanding on the processes that govern DOM composition and transport in tropical catchments is needed to assess the importance of DOM for stream ecosystem processes, both under natural conditions and human pressure. Approximately 40% of the globe's land surface is used for crop production and pasture (Alexandratos and Bruinsma, 2012). With future human population growth, food production must increase with intensification and expansion of agriculture (Alexandratos and Bruinsma, 2012; World Bank, 2013). Agricultural land use can affect both the concentration and the composition of riverine DOM in temperate catchments by affecting landscape features, the soil organic matter pool, hydrological flow paths, microbial DOM processing, and changes in riverine DOM production (Chow et al., 2007; Chen and Driscoll, 2009; Graeber et al., 2012; Stanley et al., 2012). DOM from agricultural sources

is generally thought to be of low molecular weight and aromaticity, and of high redox state and lability (Stanley et al., 2012). The elemental composition, i.e. C:N:P stoichiometry, of aquatic DOM may also be shifted by land use towards lower C:N and C:P ratios, with subsequent effects on the metabolism of the heterotrophic microbial community (Sterner and Elser, 2002; Cross et al., 2005; Heinz et al., 2015). However, the above-mentioned patterns are synthesized from studies in temperate regions and studies on agricultural impacts on DOM in tropical catchments are rare. For example, a case study in Peruvian Amazonian headwater streams detected higher total and biodegradable DOC concentrations in 2nd-growth forest and pasture streams than in an undisturbed forest stream (Bott and Newbold, 2013). An inter-biome comparison of the effects of agriculture on stream DOM concentration and composition concluded that tropical regions might currently be less affected than temperate regions due to their more recent land and fertilizer use history (Graeber et al., 2015). An unprecedented wave of urban growth in developing—mostly tropical—countries is forecast until 2050, with potentially dramatic impacts on water availability and quality (McDonald et al., 2011). Urbanization can also increase stream water DOM concentrations (Aitkenhead-Peterson et al., 2009; Silva et al., 2011) and a large fraction of urban DOM is thought to be labile (Hudson et al., 2007). However, the composition of urban wastewater-derived DOM varies substantially depending on the kind of wastewater and the type of treatment process (Imai et al., 2002)—with more advanced treatment generally removing more bioavailable DOM—and may thus also vary substantially among different urban settings and countries. Further, urban sources contribute a wide range of synthetic organic chemicals to stream water DOM, including biocides, personal care products, pharmaceuticals, hormones, and flame retardants. Some of those chemicals can stress stream biota and communities, and alter ecosystem processes even at very low concentrations (Stanley et al., 2012). In conclusion, DOM plays a key-role in regulating ecosystem processes and services, such as microbial metabolism. DOM from anthropogenic sources is generally thought to be more labile and bioavailable, as e.g. indicated by low C:N and C:P ratios, than that from natural sources. Land-use change may considerably affect riverine DOM exports, but catchment responses may vary widely, depending on specific changes in terrestrial DOM sources and aquatic in situ DOM production (Stanley et al., 2012). Several aspects of land-use impacts are particularly unclear for tropical regions and cannot be simply derived from work in temperate regions, as tropical regions may differ considerably in biota, soil type, climate, agricultural practices, land-use patterns and applied wastewater treatment technology. Therefore, the main objective of this study was to shed light into the impacts of different land-use types on DOM export from tropical catchments. More specifically, we hypothesized that (1) natural catchments at the Cerrado–Atlantic forest transition had low stream water DOC concentrations, but high exports, due to high specific discharges. (2) Ratios of C:N and C:P of DOM were also expected to be low. We hypothesized that (3) deforested pasture and agricultural catchment had lower DOC concentrations and exports than natural catchments, due to soil organic matter depletion, but lower C:N and C:P ratios of DOM due to fertilizer use. (4) Urbanization was expected to increase DOC concentrations and exports and to decrease C:N and C:P ratios of fluvial DOM, due to massive inputs of poorly treated sewage from septic tanks. Moreover, we hypothesized that (5) year-round fertilizer use and irrigation in agricultural catchments and sewage discharge in urban catchments would counteract the natural seasonality in C:N and C:P ratios of stream water DOM, thereby

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

alleviating potential nutrient limitation of heterotrophic microbial metabolism. To test these hypotheses, we sampled 20 tropical headwater catchments in SE Brazil dominated by natural vegetation, pasture, crop production, and urban land cover. We compared DOC concentrations and catchments exports, as well as elemental DOM stoichiometry among land-use types and tested for relationships with specific stream discharge, the main variable characterizing seasonality in the investigated region. Further, we conducted a literature review on DOC concentrations, exports and DOM stoichiometry in natural and human-impacted Neotropical catchments in order to assess whether the land-use effects found in our study reflect a general pattern. 2. Methods 2.1. Study area The investigated headwaters are situated in the Rio das Mortes catchment (6619 km2 total catchment size) in the Brazilian Federal State of Minas Gerais. The Rio das Mortes is a fifth-order tributary to the Rio Grande in the upper Paraná basin, in the transition zone between the Brazilian Cerrado savannah and the Atlantic rainforest. The catchment has acidic, nutrient-poor soils that are rich in iron and manganese. Liming and NPK fertilizer application is common in agricultural areas. The predominant land cover in the Rio das Mortes catchment (Silva-Junior et al., 2014) is native vegetation (52.0% of total catchment area), followed by agriculture (pasture, 30.2%; crop production, 5.6%; open soil and burnt areas, 7.3%; and eucalypt plantations, 1.3%) and urban cover (urban areas, 1.2%; roads, 2.0%; railways, 0.2%; and mines, 0.1%). The mean annual precipitation of the region is 1400 mm yr−1 (data from the only weather station in the catchment, in the city of Barbacena). The tropical climate of the region is characterized by two main seasons; a warm and rainy summer (22nd September–20th March) and a mild and dry winter (21st March–21st September). Between 1961 and 2012, July was the coldest month (15.1 °C; multi-year mean value; data from Barbacena weather station) with the lowest mean monthly precipitation (15.2 mm). February was the warmest month (20.8 °C) and the highest mean precipitation occurred in January (274.1 mm). 2.2. Sampling sites and procedure We chose 20 first-order headwater stream catchments, representing the land cover categories natural vegetation (Nat), pasture (Past), intensive agriculture (crop production; Ag), and urban area (Urb). All studied streams were independent from each other (i.e. unnested), but some streams contributed to the same 2nd- or 3rd-order stream catchment (Fig. S1). Their catchment sizes ranged between 0.08 km2 (natural stream Correias; Table S1) and 11.9 km2 (urban stream Santo Antônio; Table S1). Remote sensing data (Silva-Junior et al., 2014) were validated with onsite surveys to avoid inconsistencies (e.g. animal breeding occurring in catchments that had 100% natural vegetation according to remote sensing), and we based our land-cover classifications and catchment choice on onsite survey data in such cases. Catchments classified as predominantly natural or agricultural (pasture and intensive agriculture) had this land cover in the entire 50 m riparian corridor and floodplain of the stream, and in N 60% of the entire catchment area. Catchments from these three land cover types (Nat, Ag, Past) were carefully chosen not to have any housing or sewage inputs. Additionally, natural catchments did not exhibit intensive agriculture at all, and had low percentages of cattle grazing in remote areas of the catchment only. The vegetation of natural catchments was mature Cerrado savannah or Semideciduous Atlantic forest (Table S1). Urban catchments exhibited direct urban impacts in the riparian zone, e.g. sewage discharge, impervious surface area, and housing. In all urban catchments, sewage originated from septic tanks without further sewage treatment. However, the contribution of urban area to total catchment

787

was rather small in most catchments, ranging from 12% to 74% (median: 43%). We took triplicate water samples in each of the 20 investigated headwater streams in four seasonal sampling campaigns in March, May, September, and December 2012. These three samples, taken in within a few minutes, were used for analytical quality control and their averages were used for subsequent statistical analyses. We transported the samples to the laboratory on ice and filtered them immediately through pre-rinsed organic matter-free glass fiber filters with a nominal pore size of 0.45 μm (Sartorius, Göttingen). Additionally, we measured stream discharge using an advective Doppler velocimeter (Flow-Tracker, SonTek, CA, USA) in each stream once during each sampling campaign directly before taking the water samples. In order to obtain specific discharge (SpQ) estimates, discharge values were divided by the drainage basin area upstream of the sampling station (ha), obtained by geoprocessing of the topographic map of the Rio das Mortes catchment (Silva-Junior et al., 2014). The investigated catchments were located in the seasonal tropics, which are characterized by a rainy season with frequent rain events and a dry season with little water export from catchments. Thus, we considered SpQ as the most adequate proxy variable for seasonality in the statistical analyses of our results. 2.3. Laboratory analyses Dissolved organic carbon concentrations of samples were quantified by high-temperature catalytic oxidation with non-dispersive infrared detection (Vario TOC Cube, Elementar, Hanau, Germany). All measured samples had concentrations considerably higher than the method's lower detection limit (50 μg DOC L−1). We performed nitrogen and phosphorus analyses with standard spectrophotometric methods (APHA, 2005) using a flow injection analysis system (FIAlab 2500, FIAlab, WA, USA) equipped with a 50 cm-long flow cell for lowconcentration nutrient analysis. Detection limits were 2.0 μg L−1 ammonium-N (NH4-N), 0.4 μg L−1 nitrate + nitrite-N (henceforth referred to as NO3-N), and 2.0 μg L−1 soluble reactive phosphorus (SRP). Briefly, the salicylate method was used for NH4-N, the cadmium reduction–sulfanilamide method for NO3-N, and the ascorbic acid–molybdate method for SRP analysis. We measured total dissolved nitrogen (TDN) and phosphorus (TDP) concentrations of samples as NO3-N and SRP as described previously, following persulfate digestion (APHA, 2005). Dissolved organic nitrogen (DON) and dissolved organic phosphorus (DOP) concentrations were subsequently calculated as DON = TDN − (NH4-N + NO3-N) and DOP = TDP − SRP. Instantaneous catchment exports of DOC (ExportDOC, in g C ha−1 d−1) were calculated for each sampling site and campaign as ExportDOC = DOC × Q ∕ A, where DOC is DOC concentration (g m−3), Q is stream discharge (m3 d− 1), and A is the drained catchment area upstream of the sampling station (ha). 2.4. Statistical analysis Data were analyzed using linear mixed-effect models. The response variables DOC, ExportDOC, as well as molar DOC:DON, DOC:DOP, and DON:DOP ratios were analyzed separately as a function of the fixed factors land use, specific discharge (SpQ, in L ha−1 s−1), and their interaction (land use × SpQ). We included different streams/catchments as a random factor in order to account for systematic variability among streams/catchments. The function “lme” of the package “nlme” within the software R was used for these analyses. Distributions of residuals were analyzed to select the most appropriate models (Zuur et al., 2009). To achieve homogeneous distributions of residuals, all response variables were log10-transformed prior to analysis. The generic formula for the final model in R was log10(response variable) ~ land_use ∗ SpQ, random = ~1|stream, weights = varIdent(form = ~1|stream). The effects of land use were tested in the linear mixed models by comparing

788

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

intercepts, i.e. the respective response variables at the theoretical level SpQ = 0 L ha−1 s−1, and the effects of SpQ on the response variables were tested by comparing slopes. Respecting the orthogonal number of comparisons, we used natural catchments (Nat) as controls, and tested for each response variable whether the slope and intercept of Nat differed from 0. We then tested the intercept and slope of each treatment (land use types Ag, Past and Urb) against the intercept and slope of the control Nat. 3. Results 3.1. DOC concentrations Stream water DOC concentrations ranged from 0.47 to 5.9 mg L−1 (median: 1.5 mg L−1; n = 20) in natural streams, and were lower in pasture streams (0.11–2.0 mg L−1, median: 0.84 mg L−1) and higher in urban streams (0.58–25.9 mg L− 1, median: 12.5 mg L− 1; Fig. 1, Tables 1 and S2). Specific discharge (SpQ) affected DOC concentrations in natural streams, and there were interaction effects between agricultural land use and SpQ, and between urban land use and SpQ, due to the fact that DOC increased with increasing SpQ in natural streams (flushing effect), but not in agricultural and urban streams (Fig. 1, Tables 1 and S2). 3.2. Instantaneous DOC exports Instantaneous DOC exports ranged from 0.13 to 273 g ha−1 d− 1 (median: 2.5 g ha−1 d−1) in natural catchments, and were higher in agricultural (0.40–88.4 g ha−1 d−1, median: 7.8 g ha−1 d−1) and urban catchments (21.6–1612 g ha−1 d−1, median: 71.9 g ha−1 d−1; Fig. 1, Tables 1 and S2). SpQ positively affected DOC exports in natural catchments, and there were interaction effects between all land use types and SpQ (Fig. 1, Tables 1 and S2), because DOC exports increased more strongly with increasing SpQ in pasture catchments, but more

Table 1 Effects of land use (Urb = urban, Past = pasture, Ag = intensive agriculture), specific discharge (SpQ), and interactions of both (×) on dissolved organic carbon (DOC) concentration and export, and molar elemental ratios of dissolved organic matter. Upward and downward pointing arrows indicate positive and negative effects respectively (P ≤ 0.05, single arrow; P b 0.01, two arrows). See Table S2 for complete model statistics. Ag DOC DOC export DOC:DON DOC:DOP DON:DOP

Past

Urb

SpQ

Ag × SpQ

↓

↑↑ ↑↑ ↓↓ ↑↑ ↑↑

↑ ↑↑

↓ ↓↓

↑ ↓

↓↓ ↓↓

Past × SpQ ↑↑ ↑

↑↑

Urb × SpQ ↓↓ ↓↓ ↓↓

DON = dissolved organic nitrogen, DOP = dissolved organic phosphorus.

weakly in urban and agricultural catchments than in natural ones (Fig. 1). 3.3. DOM elemental stoichiometry DOC:DON ratios ranged from 5 to 238 (median: 26) in natural streams, and were lower in pasture (4–72, median: 7) and urban streams (4–17, median: 6; Fig. 2, Tables 1 and S2). SpQ did not affect DOC:DON ratios in natural streams, but there was an interaction effect between pasture and SpQ due to a significant positive relationship between DOC:DON and SpQ in pasture streams (Fig. 2, Tables 1 and S2). DOC:DOP ratios ranged from 43 to 2526 (median: 140) in natural streams (Fig. 2), and were lower in pasture (16–554, median: 63), but higher in urban streams (47–3171, median: 752; Fig. 2, Tables 1 and S2). SpQ positively affected DOC:DOP ratios in natural streams and there was an interaction effect between urban land use and SpQ due to the absence of a SpQ dependency of DOC:DOP ratios in urban streams (Fig. 2, Tables 1 and S2). DON:DOP ratios ranged from 0.2 to 92 (median: 8) in natural streams, and were higher in urban streams (13–631, median:

Fig. 1. Dissolved organic carbon (DOC) concentrations and exports in natural (Nat), intensive agricultural (Ag), pasture (Past) and urban catchments (Urb). Boxes in boxplots show the 1st quartile, median and 3rd quartile and whiskers represent an estimate of the 95% confidence interval for the difference between two medians (graphs on the left-hand side). In the linear mixed-effect models with the factors land use and specific discharge (graphs on the right-hand side), the shaded areas represent 95% confidence bands. See Tables 1 and S2 for model statistics.

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

789

Fig. 2. Molar elemental ratios of dissolved organic matter in stream water from natural (Nat), intensive agricultural (Ag), pasture (Past) and urban catchments (Urb). Boxes in boxplots show the 1st quartile, median and 3rd quartile and whiskers represent an estimate of the 95% confidence interval for the difference between two medians (graphs on the left-hand side). In the linear mixed-effect models with the factors land use and specific discharge (graphs on the right-hand side), the shaded areas represent 95% confidence bands. DOC = dissolved organic carbon, DON = dissolved organic nitrogen, DOP = dissolved organic phosphorus. See Tables 1 and S2 for model statistics.

100). SpQ did not affect DON:DOP in natural streams and there were no interaction effects between SpQ and land use (Fig. 2, Tables 1 and S2).

4. Discussion

In conclusion, our results supported our first hypothesis that natural catchments at the Cerrado–Atlantic forest transition had low stream water DOC concentrations. However, contrary to our predictions, DOC exports were also low (hypothesis 1). Ratios of DOM C:P were low as predicted (hypothesis 2), but DOM C:N ratios were high and may point to stoichiometric N limitation of microbial metabolism in the studied tropical streams.

4.1. DOM concentrations, exports and elemental stoichiometry in natural catchments

4.2. Land-use effects on DOC concentrations and exports

The investigated natural secondary mountain streams at the Cerrado–Atlantic forest transition had relatively low DOC concentrations compared to other tropical streams (see Text S1 for a comprehensive discussion of DOM in the investigated natural catchments). Average annual DOC exports were within the lower to intermediate range of literature values for tropical catchments, and were estimated to account for 0.01 to 0.33% of terrestrial NPP only (Text S1). These values may be typical for the tropics, but are low compared to temperate and boreal regions. Natural stream DOC:DON ratios were in the upper range of values reported for tropical streams and rivers, and may suggest a stoichiometric N limitation of microbes metabolizing DOM in these streams. DOC:DOP and DON:DOP ratios of natural streams were within the range of ratios reported for several Amazonian streams and low compared to ratios reported for larger tropical rivers, and may indicate a relatively weak or absent stoichiometric P limitation of microbes metabolizing DOM (Text S1).

Unlike inorganic nutrient concentrations that can be negatively affected by deforestation in the tropics (Neill et al., 2001; Figueiredo et al., 2010), stream water DOC appears to generally increase as a result of land-use change to pasture, crop production and urban area (Table 2). Potential mechanisms include high allochthonous DOC exports from pasture and agricultural areas to streams due to the degradation of organic matter stocks and root exudation in crops, high primary productivity in streams suffering riparian clear-cutting and associated high instream DOC production, and high DOC exports to streams by untreated or poorly treated sewage discharge from urban areas (Williams et al., 1997; Ometo et al., 2000; Daniel et al., 2002). Corresponding with that pattern, urban streams in our study showed considerably increased DOC concentrations. However, pasture streams in our study had lower DOC concentrations than natural streams. Pasture catchments in the Rio das Mortes catchment, and probably throughout the seasonal and wet tropics (Williams et al., 1997), are often heavily

790

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

Table 2 Dissolved organic matter concentrations, exports, and stoichiometry of small Neotropical streams. System

DOC (mg L−1)

ExportDOC (g ha−1 d−1)

DOC:DON

DOC:DOP

DON:DOP

3 montane tropical streams, Luquillo Forest, Puerto Rico (McDowell and Asbury, 1994) 6 pristine streams, Rio Tempisque basin, NW Costa Rica (Newbold et al., 1995) 2 pristine headwater streams, Central Amazon, Brazil (McClain et al., 1997) – Clearwater stream – Blackwater stream nd 2 order pristine Cerrado stream, Central Brazilian Plains (Markewitz et al., 2006) – 1998/1999 – 1999/2000 3 low order streams, Solimões River, Central Amazon, Brazil (Williams and Melack, 1997) – Reference stream – Partially deforested (upstream) – Partially deforested (downstream) 2 small streams, Piracicaba basin, São Paulo, SE Brazil (Ometo et al., 2000) – Pasture stream – Agricultural stream (sugar cane) 4 streams, Rondônia, Brazilian Amazon (Neill et al., 2001) – Forest streams – Pasture streams Stream in the Piracicaba basin, São Paulo, SE Brazil (Daniel et al., 2002) – Headwater – Treated sewage affected – Raw sewage affected Pasture stream, Paragominas, East. Brazilian Amazon (Markewitz et al., 2004) 4 low-order streams, Paragominas, East. Brazilian Amazon (Figueiredo et al., 2010) – Pristine streams – Mixed land use streams 9 streams, Central Brazilian Plains (Silva et al., 2011) – Pristine streams – Rural streams – Urban streams 3 headwater streams, Madre de Dios basin, Peruvian Amazon (Bott and Newbold, 2013) – Pristine forest stream – 2nd-growth forest stream – Pasture stream 20 streams, Rio das Mortes basin, SE Brazila – Natural streams – Agricultural streams – Pasture streams – Urban streams

1.4–2.1 0.5–1.2

90.3–257 52.1–118

10.6–19.1 14.7

– –

– –

4.5 38.1

32.1 247

27.9 59.7

– –

– –

8.1 2.4

86.3 34.4

78.6 30.2

– –

– –

0.9 1.0 1.9

– 95.7 182

23.5 14.3 22.6

304 231 376

12.9 16.1 16.7

2.9 5.8

– –

– –

– –

– –

– –

– –

– –

– –

4.1–41.7 32.1–60.4

3.9 6.6 10.1 1.3

– – – 2.1

– – – 11.4

– – – –

– – – –

1.3 0.7–4.2

– –

34.6 10.6–37.0

– –

– –

1.3 1.3 1.6

– – –

8.2 9.5 6.2

– – –

– – –

0.8 7.9 2.7

– – –

12.3 24.3 28.2

205 1625 516

16.7 66.9 18.3

0.9–4.3 0.7–2.0 0.6–0.9 9.9–14.6

1.5–72.8 3.1–39.6 2.0–14.8 72.6–1076

24.3–77.0 14.7–36.2 8.2–21.0 5.7–8.9

81.1–791 54.7–212 90.1–185 557–1281

6.6–24.6 6.9–16.7 12.3–20.7 81.9–220

DOC = dissolved organic carbon, DON = dissolved organic nitrogen, DOP = dissolved organic phosphorus. a This study (ranges of stream annual averages among the 5 streams of each land cover category).

affected by surface erosion, leading to the loss of surface soil organic matter and diminished DOC concentrations in soil solution. Moreover, aboveground harvesting of corn, wheat and beans (Table S1) leaves root biomass in soils that may contribute to higher DOC emissions from agricultural than from pasture catchments. Interestingly, the low DOC concentrations in pasture streams did not lead to diminished instantaneous catchment DOC exports, because they were compensated for by higher specific discharges in these streams; probably due to low soil infiltration and evapotranspiration, and thus low water losses to groundwater and the atmosphere in pasture catchments (Hunke et al., 2014). Conversely, agricultural catchments did not show increased stream water DOC, but had higher instantaneous DOC exports than natural streams due to higher specific discharges. However, the massively increased instantaneous DOC exports from urban catchments were mainly due to high DOC concentrations, especially at low flow (Fig. 1), probably resulting from missing or poor sewage treatment by the septic tanks commonly used, and often not properly maintained in the region. There is no literature data available on land-use impacts on Neotropical catchment DOC exports, to which we could compare our results (Table 2). However, in a study of seven forested and five agricultural streams on the tropical island of Hawai'i, there was no evidence that anthropogenic activity affected catchment DOC exports (Michaud and Wiegner, 2011). Data from other climate zones suggest strong positive effects of agricultural and urban land use on DOC exports in some catchments, but no or weak effects in others (Aitkenhead-Peterson et al., 2009; Graeber et al., 2012; Stanley et al.,

2012). Impacts of pasture on DOC exports appear to be less consistent than those of crop production, but pasture management regimes, i.e. N fertilizer use and artificial drainage, appear to be governing factors for DOC exports (Nelson et al., 1996; McTiernan et al., 2001). In boreal and temperate regions, peatland and wetland saturated organic soils can be more important for stream DOC exports than agriculture (Mattsson et al., 2009). To sum up, our third hypothesis was only partially confirmed and pasture catchments had lower DOC concentrations than natural catchments as predicted, while agricultural catchments showed no differences in stream water DOC concentrations. Contrary to our predictions, higher specific discharges resulted in similar or higher DOC exports in pasture and agricultural catchments, respectively, than in natural ones (hypothesis 3). As expected (hypothesis 4), urbanization led to massive increases in stream water DOC concentrations and exports, due to poorly treated sewage from septic tanks in the studied tropical catchments. 4.3. Land-use effects on DOM elemental stoichiometry and its implications for DOM bioavailability to heterotrophic bacteria There is very little work available about land-use effects on DOM stoichiometric ratios in tropical streams, and the few available studies do not show a clear trend regarding effects on DOC:DON ratios (Table 2). However, deforestation and pastoral land use appear to increase C:P and N:P ratios of stream water DOM (Table 2), thus pushing

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

heterotrophic metabolism of DOM possibly towards P limitation, considering that these land-use types are usually not associated with increases in stream water phosphate concentration that microbes could use to supplement P deficiency of DOM. Interestingly, our study confirmed this pattern for urban streams, but not for agricultural and pasture streams. This was unexpected as sewage is thought to be the major source of DOP to urban and rural streams (Aitkenhead-Peterson et al., 2003). We also found considerably lower DOC:DON ratios in urban streams, suggesting that urban land use may have alleviated N limitation, but may have increased, or even caused, P limitation of heterotrophic DOM decomposition. Additionally, urban DOM in the investigated region—and probably in large parts of the tropics—is derived from septic tanks and may thus be highly bioavailable (Hudson et al., 2007). Pasture streams showed both lower DOC:DON and DOC:DOP ratios than natural streams, thus pastoral land use may have diminished N limitation of DOM decomposition, but also P limitation, if present. In combination with higher DIN and phosphate concentrations in the studied pasture streams compared to natural streams (Table S1), these results point to an eutrophication effect, probably due to fertilizer use and cattle excrements on pasture areas. Agricultural streams showed only lower DOC:DOP ratios than natural streams, but no changes in elemental ratios involving N. Therefore, it appears unlikely that changes in DOM stoichiometry due to agricultural land use could have caused changes in the N limitation of DOM decomposition probably occurring in the studied seasonal tropical streams. However, high DIN concentrations in some of the studied agricultural streams (Table S1), probably due to fertilizer use, could still have subsidized heterotrophic microbial metabolism and thus have affected DOM decomposition. Bioavailability tests of DOM including experimental N and P supplementation would be the next step to investigate the potential landuse effects on tropical stream DOM quality suggested by our results on DOM elemental stoichiometry. We conclude that deforested pasture and agricultural catchments had lower DOM C:N ratios than natural catchments as predicted (hypothesis 3), but that lower C:P ratios were only found in agricultural catchments. Urbanization led to decreases DOM C:N ratios as expected (hypothesis 4), but surprisingly, C:P ratios of fluvial DOM increased due to urbanization in the studied tropical catchments. 4.4. Effects of specific discharge The investigated catchments were located in the seasonal tropics, which are characterized by a rainy season with high stream discharge and a dry season with little water export from catchments. Therefore, specific discharge (SpQ), i.e. the amount of water exported by streams per catchment area, represents an appropriate proxy variable for seasonality in the studied catchments. SpQ positively affected DOC concentration, export and DOC:DOP ratios (Table 1) in the investigated natural catchments, indicating that there was a natural seasonality in DOM concentration, export and P content. Dilution effects on inorganic nutrients, i.e. the dilution of stream water with less concentrated water, are common in natural tropical catchments (Grimaldi et al., 2004; Bücker et al., 2011), and the flushing effect we encountered for DOC may represent the export of soil organic matter decomposed during the previous dry season (Williams et al., 1997). As predicted (hypothesis 5), land use had profound effects on the seasonality in catchments biogeochemistry; however, effects on DOC concentrations and exports were stronger than those on DOM elemental stoichiometry. Urban land use profoundly changed the observed seasonal patterns by leading to whole-year-round high DOC concentrations, exports and DOC:DOP ratios, potentially altering the seasonality in P limitation of DOM decomposition in the studied tropical streams. Agricultural and pastoral catchments also differed in DOM seasonality from natural catchments; agricultural catchments showed a weaker and pasture catchments a stronger seasonality in DOC exports than

791

natural catchments. In agricultural catchments, this difference was due to the absence of the DOC concentration flushing effect observed in natural catchments, whereas in pasture catchments the stronger seasonality appeared to be a hydrologic effect, not caused by differences in DOC concentration seasonality (Table 1, Fig. 1). Finally, pasture catchments showed seasonality in DOC:DON ratios, not observed in natural catchments. 5. Conclusions The present study is the first to demonstrate detailed effects of different land-use types on catchment DOM biogeochemistry in tropical catchments and shows that land use—especially urbanization—profoundly affected stream water DOC concentration and stoichiometry, as well as catchment DOC exports in the studied Neotropical catchments. Pasture streams had lower DOC concentration, and DOC:DON and DOC:DOP ratios than natural streams. Catchments with intensive agriculture had higher DOC exports and lower DOC:DOP ratios than natural catchments. Urbanized catchments exhibited the highest DOC concentrations and exports among all land-use types, as well as the highest DOC:DOP and DON:DOP ratios, but the lowest DOC:DON ratios. Therefore, urbanization may have alleviated N limitation of heterotrophic DOM decomposition, but increased P limitation. The reported effects are important, not only because they may cause water quality problems in water scarce regions, but also because tropical riverine DOM exports are an important component in the global carbon cycle. The observed changes in DOM export and seasonality may also affect heterotrophic metabolic rates, and thus CO2 evasion, both in the studied river system and in downstream reservoirs. With the projected future population increases and the expansion and intensification of agriculture in the tropics, sewage treatment, soil conservation and sustainable fertilizer application—issues often ignored in tropical regions—should thus be important targets for catchment management and mitigation efforts. Acknowledgments This research was funded through the DONCOPRA project by the Foundation for Research Support of the State of Minas Gerais (FAPEMIG, CRA-80/10), including a M.Sc. fellowship to R.C.S. Silva, and through the research network REHMANSA by the Funding Authority for Research and Projects (FINEP, 01.12.0064.00). D. Graeber was supported by a grant from the Danish Centre for Environment and Energy, Aarhus University. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.scitotenv.2016.01.158. References Aitkenhead-Peterson, J.A., McDowell, W., Neff, J., Stuart, E., Robert, L., 2003. Sources, production, and regulation of allochthonous dissolved organic matter inputs to surface waters. In: Findlay, S.E.G., Sinsabaugh, R.L. (Eds.), Aquatic Ecosystems: Interactivity of Dissolved Organic Matter. Academic Press, London, pp. 25–70. Aitkenhead-Peterson, J.A., Steele, M.K., Nahar, N., Santhy, K., 2009. Dissolved organic carbon and nitrogen in urban and rural watersheds of south-central Texas: land use and land management influences. Biogeochemistry 96, 119–129. Alexandratos, N., Bruinsma, J., 2012. World Agriculture Towards 2030/2050: The 2012 Revision. Food and Agricultural Organization of the United Nations, Rome. APHA, 2005. Standard Methods for the Examination of Water and Wastewater. American Public Health Association, Washington, DC. Benner, R., Opsahl, S., Chin-Leo, G., Richey, J.E., Forsberg, B.R., 1995. Bacterial carbon metabolism in the Amazon River system. Limnol. Oceanogr. 40, 1262–1270. Billett, M.F., Deacon, C.M., Palmer, S.M., Dawson, J.J.C., Hope, D., 2006. Connecting organic carbon in stream water and soils in a peatland catchment. J. Geophys. Res. Biogeosci. 111, G02010. Bott, T.L., Newbold, J.D., 2013. Ecosystem metabolism and nutrient uptake in Peruvian headwater streams. Int. Rev. Hydrobiol. 98, 117–131.

792

B. Gücker et al. / Science of the Total Environment 550 (2016) 785–792

Bücker, A., Crespo, P., Frede, H.G., Breuer, L., 2011. Solute behaviour and export rates in Neotropical montane catchments under different land uses. J. Trop. Ecol. 27, 305–317. Chen, X., Driscoll, C.T., 2009. Watershed land use controls on chemical inputs to Lake Ontario embayments. J. Environ. Qual. 38, 2084–2095. Chow, A.T., Dahlgren, R.A., Harrison, J.A., 2007. Watershed sources of disinfection byproduct precursors in the Sacramento and San Joaquin Rivers, California. Environ. Sci. Technol. 41, 7645–7652. Cross, W.F., Benstead, J.P., Frost, P.C., Thomas, S.A., 2005. Ecological stoichiometry in freshwater benthic systems: recent progress and perspectives. Freshw. Biol. 50, 1895–1912. Daniel, M.H.B., Montebelo, A.A., Bernardes, M.C., Ometto, J.P.H.B., de Camargo, P.B., Krusche, A.V., Ballester, M.V., Victoria, R.L., Martinelli, L.A., 2002. Effects of urban sewage on dissolved oxygen, dissolved inorganic and organic carbon, and electrical conductivity of small streams along a gradient of urbanization in the Piracicaba river basin. Water Air Soil Pollut. 136, 189–206. Figueiredo, R.O., Markewitz, D., Davidson, E.A., Schuler, A.E., Dos, S., Watrin, O., de Souza Silva, P., 2010. Land-use effects on the chemical attributes of low-order streams in the eastern Amazon. J. Geophys. Res. Biogeosci. 115, G04004. Findlay, S., Sinsabaugh, R.L., 2003. Aquatic Ecosystems: Interactivity of Dissolved Organic Matter. Academic Press, London. Graeber, D., Gelbrecht, J., Pusch, M.T., Anlanger, C., von Schiller, D., 2012. Agriculture has changed the amount and composition of dissolved organic matter in Central European headwater streams. Sci. Total Environ. 438, 435–446. Graeber, D., Boëchat, I.G., Encina-Montoya, F., Esse, C., Gelbrecht, J., Goyenola, G., Gücker, B., Heinz, M., Kronvang, B., Meerhoff, M., Nimptsch, J., Pusch, M.T., Silva, R.C.S., von Schiller, D., Zwirnmann, E., 2015. Global effects of agriculture on fluvial dissolved organic matter. Sci. Rep. 5, 16328. Grimaldi, C., Grimaldi, M., Millet, A., Bariac, T., Boulegue, J., 2004. Behaviour of chemical solutes during a storm in a rainforested headwater catchment. Hydrol. Process. 18, 93–106. Heinz, M., Graeber, D., Zak, D., Zwirnmann, E., Gelbrecht, J., Pusch, M.T., 2015. Comparison of organic matter composition in agricultural versus forest affected headwaters with special emphasis on organic nitrogen. Environ. Sci. Technol. 49, 2081–2090. Hessen, D.O., Ågren, G.I., Anderson, T.R., Elser, J.J., de Ruiter, P.C., 2004. Carbon sequestration in ecosystems: the role of stoichiometry. Ecology 85, 1179–1192. Hudson, N., Baker, A., Reynolds, D., 2007. Fluorescence analysis of dissolved organic matter in natural, waste and polluted waters — a review. River Res. Appl. 23, 631–649. Hunke, P., Mueller, E.N., Schröder, B., Zeilhofer, P., 2014. The Brazilian Cerrado: assessment of water and soil degradation in catchments under intensive agricultural use. Ecohydrology. Imai, A., Fukushima, T., Matsushige, K., Kim, Y.-H., Choi, K., 2002. Characterization of dissolved organic matter in effluents from wastewater treatment plants. Water Res. 36, 859–870. Jaffé, R., Yamashita, Y., Maie, N., Cooper, W.T., Dittmar, T., Dodds, W.K., Jones, J.B., Myoshi, T., Ortiz-Zayas, J.R., Podgorski, D.C., 2012. Dissolved organic matter in headwater streams: compositional variability across climatic regions of North America. Geochim. Cosmochim. Acta 94, 95–108. Kim, H.-C., Yu, M.-J., 2005. Characterization of natural organic matter in conventional water treatment processes for selection of treatment processes focused on DBPs control. Water Res. 39, 4779–4789. Lewis Jr., W.M., Saunders III, J.F., 1989. Concentration and transport of dissolved and suspended substances in the Orinoco River. Biogeochemistry 7, 203–240. Lewis Jr., W.M., Saunders III, J.F., Levine, S.N., Weibezahn, F.H., 1986. Organic carbon in the Caura River, Venezuela. Limnol. Oceanogr. 31, 653–656. Lewis Jr., W.M., Hamilton, S.K., Saunders III, J.F., 1995. Rivers of northern South America. In: Cushing, C.E., Cummins, K.W., Minshall, G.W. (Eds.), Ecosystems of the World: Rivers and Stream Ecosystems. Elsevier Science B.V, Amsterdam, pp. 219–256. Markewitz, D., Davidson, E., Moutinho, P., Nepstad, D., 2004. Nutrient loss and redistribution after forest clearing on a highly weathered soil in Amazonia. Ecol. Appl. 14, 177–199. Markewitz, D., Resende, J.C.F., Parron, L., Bustamante, M., Klink, C.A., Figueiredo, R.D.O., Davidson, E.A., 2006. Dissolved rainfall inputs and streamwater outputs in an undisturbed watershed on highly weathered soils in the Brazilian cerrado. Hydrol. Process. 20, 2615–2639. Mattsson, T., Kortelainen, P., Laubel, A., Evans, D., Pujo-Pay, M., Räike, A., Conan, P., 2009. Export of dissolved organic matter in relation to land use along a European climatic gradient. Sci. Total Environ. 407, 1967–1976.

McClain, M.E., Richey, J.E., Brandes, J.A., Pimentel, T.P., 1997. Dissolved organic matter and terrestrial lotic linkages in the Central Amazon Basin of Brazil. Global Biogeochem. Cycles 11, 295–311. McDonald, R.I., Green, P., Balk, D., Fekete, B.M., Revenga, C., Todd, M., Montgomery, M., 2011. Urban growth, climate change, and freshwater availability. Proc. Natl. Acad. Sci. U. S. A. 108, 6312–6317. McDowell, W.H., Asbury, C.E., 1994. Export of carbon, nitrogen, and major ions from three tropical montane watersheds. Limnol. Oceanogr. 39, 111–125. McTiernan, K.B., Jarvis, S.C., Scholefield, D., Hayes, M.H.B., 2001. Dissolved organic carbon losses from grazed grasslands under different management regimes. Water Res. 35, 2565–2569. Michaud, J., Wiegner, T., 2011. Stream nutrient concentrations on the windward coast of Hawai'i island and their relationship to watershed characteristics. Pac. Sci. 65, 195–217. Morel, B., Durand, P., Jaffrezic, A., Gruau, G., Molenat, J., 2009. Sources of dissolved organic carbon during stormflow in a headwater agricultural catchment. Hydrol. Process. 23, 2888–2901. Neill, C., Deegan, L.A., Thomas, S.M., Cerri, C.C., 2001. Deforestation for pasture alters nitrogen and phosphorus in small Amazonian streams. Ecol. Appl. 11, 1817–1828. Nelson, P.N., Cotsaris, E., Oades, J.M., 1996. Nitrogen, phosphorus, and organic carbon in streams draining two grazed catchments. J. Environ. Qual. 25, 1221–1229. Newbold, J.D., Sweeney, B.W., Jackson, J.K., Kaplan, L.A., 1995. Concentrations and export of solutes from six mountain streams in northwestern Costa Rica. J. North. Am. Benthol. Soc. 14, 21–37. Ometo, J.P.H.B., Martinelli, L.A., Ballester, M.V., Gessner, A., Krusche, A.V., Victoria, R.L., Williams, M., 2000. Effects of land use on water chemistry and macroinvertebrates in two streams of the Piracicaba river basin, south east Brazil. Freshw. Biol. 44, 327–337. Regnier, P., Friedlingstein, P., Ciais, P., Mackenzie, F.T., Gruber, N., Janssens, I.A., Laruelle, G.G., Lauerwald, R., Luyssaert, S., Andersson, A.J., 2013. Anthropogenic perturbation of the carbon fluxes from land to ocean. Nat. Geosci. 6, 597–607. Saunders III, J.F., Lewis Jr., W.M., 1989. Transport of major solutes and the relationship between solute concentrations and discharge in the Apure River, Venezuela. Biogeochemistry 8, 101–113. Silva, J.S.O., da Cunha Bustamante, M.M., Markewitz, D., Krusche, A.V., Ferreira, L.G., 2011. Effects of land cover on chemical characteristics of streams in the Cerrado region of Brazil. Biogeochemistry 105, 75–88. Silva-Junior, E.F., Moulton, T.P., Boëchat, I.G., Gücker, B., 2014. Leaf decomposition and ecosystem metabolism as functional indicators of land use impacts on tropical streams. Ecol. Indic. 36, 195–204. Stanley, E.H., Powers, S.M., Lottig, N.R., Buffam, I., Crawford, J.T., 2012. Contemporary changes in dissolved organic carbon (DOC) in human dominated rivers: is there a role for DOC management? Freshw. Biol. 57, 26–42. Sterner, R.W., Elser, J.J., 2002. Ecological Stoichiometry: The Biology of Elements from Molecules to the Biosphere. Princeton University Press, Princeton, NJ. Wiegner, T.N., Tubal, R.L., 2010. Comparison of dissolved organic carbon bioavailability from native and invasive vegetation along a Hawaiian river. Pac. Sci. 64, 545–555. Williams, M.R., Melack, J.M., 1997. Solute export from forested and partially deforested catchments in the central Amazon. Biogeochemistry 38, 67–102. Williams, M.R., Fisher, T.R., Melack, J.M., 1997. Solute dynamics in soil water and groundwater in a central Amazon catchment undergoing deforestation. Biogeochemistry 38, 303–335. World Bank, 2013. World Development Report 2014: Risk and Opportunity — Managing Risk for Development. World Bank, Washington, DC. Yamashita, Y., Maie, N., Briceño, H., Jaffé, R., 2010. Optical characterization of dissolved organic matter in tropical rivers of the Guayana Shield, Venezuela. J. Geophys. Res. Biogeosci. 115, G00F10. Zurbrügg, R., Suter, S., Lehmann, M.F., Wehrli, B., Senn, D.B., 2013. Organic carbon and nitrogen export from a tropical dam-impacted floodplain system. Biogeosciences 10, 23–38. Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed Effects Models and Extensions in Ecology with R. Springer, New York.