Pre- and post-flood retention efficiency of nitrogen in a Sonoran Desert stream

The objectives of this study were 1) to compare Sonoran Desert streams with other streams in terms of retention efficiency of nitrate; 2) to examine the effects of a flood on nitrate retention and to determine which factors control nitrate retention in the surface stream subsystem in Sycamore Creek, Arizona; and 3) to compare the short-term nutrient addition technique with computations based upon natural nutrient gradients. From June to September 1995, we did 8 short-term nitrate and chloride additions (4 additions before and 4 after a flood) in a 240-m reach to measure nitrate uptake length as an index of surface stream retention efficiency of nitrate. We also calculated nitrate uptake lengths based on a natural downstream decline in nitrate concentration, using data from the addition dates and from previous studies. Nitrate uptake lengths measured in Sycamore Creek were short (<I20 m) compared to published values from other streams, indicating a high retention efficiency of nitrate in this nitrogen-limited stream. A midsummer flood caused a 2-fold decrease in retention efficiency of nitrate in the reach (i.e., nitrate uptake length increased from 61 to 124 m); however, this change was within the range of variation measured before the flood. Rapid algal recovery (23 d), the dramatic decrease in discharge, and a large transient storage zone may account for the apparent high resilience of nutrient retention efficiency to disturbance. Most of the temporal variation in nitrate uptake length during the study period was attributed to changes in the algal assemblage. In particular, retention efficiency of nitrate decreased when nitrogen fixers were abundant. Uptake lengths calculated from additions were always shorter than those from natural nitrate declines, supporting our hypothesis that nutrient uptake lengths from short-term nutrient additions reflect gross, rather than net, nutrient uptake. Uptake rates from short-term additions and from natural declines of nitrate over post-flood succession showed a similar temporal pattern, but the ratio between them increased late in succession. This result suggests that, during late successional stages, nutrient release processes became more important than nutrient uptake processes, a prediction that is consistent with the ecosystem succession and nutrient retention hypothesis. Key zwrds: nitrate, nutrient uptake length, nutrient retention, disturbance, transient storage zone, algae, resilience. All ecosystems are open to some extent, exon surface-stream nutrient retention. Actual changing materials and energy with their surmeasurements of spiraling length for stream roundings. Streams and rivers are especially ecosystems are rare because of the difficulty of open ecosystems dominated by downstream quantifying inputs and outputs of all comparttransport. Yet streams also retain and transform ments in the ecosystem; however, uptake length, some of h s transported material through varthe average downstream distance traveled by a ious physical, chemical, and biological mechadissolved molecule before removal from the wanisms. Nutrients in transport may change drater column, has been proposed as an estimator matically in form and in concentration, dependof retention efficiency of nutrients in the surface ing on the hydrologic route followed (i.e., as restream (Stream Solute Workshop 1990). Shortflected in water residence time) and on the term nutrient additions have been used increasbiologically mediated changes that occur within ingly over the past 16 y to measure nutrient updifferent subsystems. The net effect of flow take length. path, heterogeneity, and biological processing Most reports of uptake length are for small, determines the extent to which the ecosystem temperate forest streams. In such systems, hetretains transported material. The spiraling conerotrophic processes dominate nutrient cycling cept (Newbold et al. 1981) provides a theoretical because the closed canopy causes light limitaand mathematical framework to estimate the intion of primary production (Mulholland et al. fluence of biotic and abiotic instream processes 1985). Further, because phosphorus is believed 806 E. MARTIET AL. [Volume 16 to be the limiting nutrient in these streams, most studies to date have focused on phosphorus uptake length (Mulholland et al. 1985, Munn and Meyer 1990, D'Angelo and Webster 1991, Marti and Sabater 1996). In contrast, desert streams are limited by nitrogen, not light. The stream channel is wide as a result of floods, the riparian canopy does not cover the channel, and thus light is not limiting to primary producers. As a result, autotrophs play an important role in nutrient dynamics. In this paper we report nitrate uptake lengths for Sycamore Creek, a Sonoran Desert stream in central Arizona. Nutrient uptake length varies as a function of channel morphology (Marti and Sabater 1996), organic matter standing stocks (D'Angelo and Webster 1991) and inputs (Mulholland et al. 1985), degree of nutrient limitation (Mum and Meyer 1990), season (Mulholland et al. 1985, D'Angelo and Webster 1991, Marti and Sabater 1996, Valett et al. 1996), and hydraulic conductivity of stream-bed sediments (Valett et al. 1996). Despite intense interest in disturbance as a force influencing stream communities and stream ecosystem structure and function (e.g., Resh et al. 1988), the influence of disturbance on nutrient uptake length has rarely been investigated (but see Aumen et al. 1990). Here we compare nitrate uptake length for a single reach of Sycamore Creek before and after a small summer flash flood in 1995. We also examine post-flood changes in nitrate uptake length calculated from data in Fisher et al. (1982), and compare these changes to the 1995 post-flood successional sequence. Several assumptions are involved in the use of short-term nutrient additions to measure uptake length (Stream Solute Workshop 1990). First, characteristics of the channel and of flow are assumed to be uniform for the study reach. Second, assumptions associated with the nutrient uptake length calculations are: 1) additions of nutrient should elevate concentration onlv withn the range where the nutrient uptakeconcentration relationship is linear; and 2) release of nutrients to the stream water does not significantly affect nutrient concentration during the experiment. The latter assumption suggests that nutrient uptake rates calculated from uptake lengths based on solute additions are closer to gross uptake than to net uptake (i.e., uptake minus release); however, this has not been demonstrated conclusively (but see Mulholland et al. 1990). In Sycamore Creek and in many other Sonoran Desert streams, dramatic downstream declines in nitrate are common, particularly below in-channel springs (Grimm et al. 1981) or hyporheic discharge zones (i.e., points where subsurface water enters the surface stream; Valett et al. 1990, 1994). In this study, we also calculated nitrate uptake length from natural, longitudinal declines in nitrate concentration. llus decline reflects the net result of uptake and release processes occurring in the surface stream. We compared these values with those from short-term nutrient addition experiments. We hope this comparison will provide a better understanding of the meaning of these concepts. The objectives of this research were to answer three major questions. First, how does retention efficiency of nitrate in Sycamore Creek in particular and desert streams in general compare with other streams? To answer th s question, we measured uptake length in Sycamore Creek using short-term nutrient addition experiments and calculated uptake length for a variety of streams using data from Grimm et al. (1981). Second, how do floods affect nitrate uptake length? We used data from nutrient addition experiments done before and after a flood and published data on post-flood nutrient retention to address this question. A secondary objective of this part of the study was to aetermine whether hydrologic or biotic factors were dominant in determining changes in uptake length over successional time. Third, we asked, what is measured by the short-term nutrient addition method? We hypothesized that nutrient additions reflect gross uptake, and predicted that uptake length calculated from natural declines in nutrient concentration would be longer than uptake length measured using nutrient additions.