Lead Investigators:

Robert Gilbert, Richard Ambrose, William Kaiser

Overview

Urbanization continues to degrade the function and services of riparian zones.  Described as the “Urban Stream Syndrome,” characteristics of this degradation include alterations to stream hydrology, geomorphology, chemistry, and biology.  Although symptoms of the Urban Stream Syndrome correlate to catchment imperviousness and drainage connectivity, the symptoms are often a result of complex interactions and many of the responses are inconsistent.  For example, stream baseflow magnitude, sedimentation, and algal biomass all have inconsistent responses to urbanization and are all influenced by multiple anthropogenically altered physical and/or chemical conditions. One possible factor leading to the inconsistency of algal biomass response is the inadequacy of the sampling methods currently employed to assess algal dynamics of urban streams.  We are developing and integrating methods of urban stream monitoring to provide a good understanding of what biologically relevant, physicochemical stream conditions are causing algal impairment at a given site. 

Approach

Two new methods of monitoring streams is being developed to perform stream characterization: high-resolution spatiotemporal sampling and an integrative, field-based algal bioassay.  These two methods will be performed in an integrated protocol along side to traditional stream sampling methods. In preparation to implement this new protocol, monthly high-resolution spatiotemporal sampling of Medea Creek, an urbanized stream, was performed using Networked Informechanical Systems (NIMS) technology for one year (June 2005 to June 2006).  This provides a greater understanding of the daily and seasonal patterns of stream constituent variations at the scale of a stream cross-section.  In addition, an outplant-based algal bioassay is currently being developed to provide a biologically meaningful measure of stream conditions.  The bioassay integrates the in-stream physicochemical conditions over an exposure period without the influence of stream conditions before the study period.  This summer, both these methods will then be combined with traditional sampling methods (biological and physicochemical) in an iterative design that will utilize the benefits of each method.  This iterative design will then be applied to streams in the Malibu Watershed to assess which stream factors are leading to algal impairment at a given location.  The result will be a more complete understanding of the driving forces behind algal impairment in urban watersheds and a focus for management practices.

Systems/Experiments

1) Seasonal, diel and cross-sectional variation at Medea Creek, a small urban influenced stream
Monthly deployments of the NIMS-RD system were performed over Medea Creek on a small parcel of undeveloped land at the corner of Kanan Rd. and Cornell Rd. in Augora Hills. The NIMS-RD system was set up over the creek and logging commenced as soon as the system was ready. Sample points were assigned every 20 cm for each of the vertical transects.  The resulting 20 cm by 20 cm grid (64 points) required 25 minutes for each scan.  The logging of pH, conductivity, temperature (Hydrolab 4a from Hach Environmental), nitrate (ISUS nitrate sensor from Satlantic), and depth (HOBO pressure sensor from Onset Corp.) was done for 24 hours.

The NIMS-RD system remained in operation for a full 24 hours (approximately 2 scans per hour).  The system was monitored during the duration of the 24-hour period. Once a full 24 hours of sampling finished, the NIMS-RD system was disassembled and the site cleared of equipment. This process was repeated approximately monthly from June 2005 to June 2006.

The data set from the monthly scans demonstrated temporal and spatial variations on all scales measured: daily, monthly, and cross-sectional.  All parameters collected showed strong diurnal patterns.  Moreover, the diurnal fluctuations changed seasonally. For example, temperature diurnal patterns at Medea Creek resembled sinusoidal functions.  However, the magnitude of the fluctuation was lower in the winter months than the spring, fall and summer months.  In addition to magnitude alterations, the entire function shifted up or down, tracking changes in seasonal air temperature (Figure 1a)

 
Figure 1: Monthly diel patterns of (a) Temperature and (b) pH

The pattern of pH also loosely resembled a sinusoidal function; however, the characteristics differed from temperature.  Compared to temperature, the peaks were sharper and earlier in the afternoon while the bottoms of the functions were flatter. Overall, pH was higher in the winter and lower in the summer and early fall.  In the winter months, the afternoon decrease in pH was sharper and the nighttime low was flatter and longer lasting compared to the spring and summer months.  The warmer months were characterized by much slower and longer downward afternoon slopes.  The nighttime low periods were shorter and do not exhibit much of a steady low (Figure 1b).
           
Parameters also show cross-sectional patterns that changed over the 24-hour sampling periods.  For example, in September conductivity showed a trend of higher conductivity at the stream edges and lower conductivity in the stream center at the early part of the sampling (late afternoon).  This pattern reversed during the nighttime hours (higher conductivity in the stream center); it switch back to the original pattern by the afternoon of the next day (Figure 2).  Visually, the change in these spatial patterns seems to correspond with the direction of the slope of the 24-hour temporal function.  In other words, the cross-sectional pattern, when the diel pattern has a negative slope, is the inverse of the cross-sectional pattern when the diel pattern has a positive slope.

This analysis of the NIMS-RD suggested that cross-sectional variation at the Medea Creek site sampled was created by differential rates of upstream water exchange.  The water at the edge of the stream tracked the temporal change slower than the center of the stream and, due to longer residence time, was more influenced by the local microclimate than the higher velocity, well-mixed stream center.  Although the variation observed was small, this observation suggests that the stream reaches with heterogeneous pool-riffle morphology may have greater variations in physicochemical conditions then homogenous stream reaches. Pools, which have slower moving water, will be more subject to the local microclimate than a riffle.  The conditions in a riffle will reflect the microclimatic conditions less and would be more representative of an average of the conditions upstream of the riffle.

              
       

Figure 2. Averaged 3-D Stream cross-sections of Conductivity in September 2005.  The small graph next to each plot is the diel pattern of conductivity.  The area between the two arrows is the period where each respective cross-section graphs was averaged.

2) Development of the in-field bioassay
In-situ algal biomass can be used to assess the biological components over the stream reach; however, because the algal biomass at the time of the study is influenced by conditions prior to the study, it may not be representative of the current physicochemical conditions. An algal bioassay, which only assesses conditions over the time of physical sampling, can be used in addition to the in-situ biomass assessment to help resolve some of the confounding temporal factors.  The algal bioassay will be able to make isolated short-term assessments by measuring the change in viability of algal monocultures in permeable, membrane-enclosed packets.  The algal packets will is placed in the stream only during the sampling period.  In order to use this method as a stream assessment tool, the following goals need to be met: (1) Suitable algal taxa need to be selected (2) The algal viability response for relevant stream factors needs to be quantified and (3) Effective use in-situ needs to be verified.

3) A multi-method approach to the characterization of stream water quality and algal dynamics
A protocol to elucidate algal-physicochemical relationships at a given site, which accounts for small-scale spatiotemporal variations, can be achieved by using high resolution mobile sensing, an algal bioassay, and static or single point physicochemical sensing.  High-resolution mobile sensing can be used to measure physicochemical conditions over the range of microhabitats as well as track diel changes in the physicochemical conditions.  However, the NIMS-RD is not compatible will all useful forms of measurements and can only sample over a single transect. Traditional sampling can be used to measure areas outside of the NIMS-RD transect(s) and perform measurements not practically taken using sensor suites. For the biological component, an algal bioassay can be used in addition to the in-situ biomass assessment to help resolve some of the confounding factors.

  

Figure 3. Temperature vs. light intensity for a UCLA botanical garden stream pool and glide when light intensity is greater then zero.  Data collected in August.  Residual temperature is calculated by subtracted the predicted values of a local regression (span = 0.4) from the raw temperature values.  The log base 10 of light intensity is used to normalize light data.

Only limited by suitable attachment points, the NIMS-RD is capable of performing any transect across or down the stream at a given site. Since each potential transect may sample different microhabitats, an optimization process for determining the NIMS-RD transect location at each site will be critical to capture the reach-scale variation.  The correlation between light intensity and water temperature can be used to quantify the microhabitats based on the physical properties of riffle-pool dynamics.  The temporal pattern in stream pool temperature is expected to have a stronger correlation with the light intensity pattern compared to a stream glide or riffle where water residence time is short.  Figure 3 shows the R2 between the light intensity and temperature when light intensity is > 0 for a glide and a pool in the UCLA Botanical Garden.  The temperature fluctuation in the glide showed little correlation to the light pattern, whereas the pool data showed a clear relationship. In this way, the reach-scale variation can be determined and a NIMS-RD transect can be set to sample the transect with the greatest variation.

Five headwater tributaries will be selected in the Malibu Watershed.  An array of 15 fixed sensors will be evenly distributed over a 50-meter reach at each site.  HOBO Pendent Logger (Onset Corp.) as well as algal packet outplants for the algal bioassay will be placed at each of the 15 points.  The HOBO Pendent Loggers will record temperatures and light intensity for the duration of the 72 hour experiment.  Three replicate packets for each of the three taxa will also be placed at each location and remain for the duration of the 3-day period.  On the beginning of the second day the light intensity and temperature data from the HOBO Pendent Loggers will be downloaded.  The correlation analysis will be applied to the first day’s data and a transect for the NIMS-RD system will be delineated over the sampling region using the R2s of the light intensity and temperature data.  The NIMS-RD's transect and sampling regime will be optimized to include the greatest range of microhabitats (R2 values) while maintaining good temporal resolution (i.e. minimizing time of a single scan).  The NIMS-RD system with the following sensor payload will perform continual scans over the optimized transect for the remaining 24-hours: nitrate (ISUS), PAR, temperature, conductivity, pH, oxidation-reduction potential, turbidity, and depth.  A single 24-hour NIMS-RD scanning period will enable single day staggering between sites.  This will result in the entire sampling to be completed in seven days.

Accomplishments

Experiment #1 has been completed.  A partial dataset has been recently published.  The writing process for the publication of the complete dataset is in progress.  Preliminary data has been collected for both experiment #2 and #3.

Future Directions

Finish publication of the dataset from experiment #1.  Continued lab testing and field-testing of experiment #2.  Perform experiment #3 in summer of 2007.

People

Faculty:

• William Kaiser UCLA Electrical Engineering,
• Richard Ambrose UCLA Environmental Health Science

Staff:

• Yeung Lam UCLA Electrical Engineering

Graduate Students:

• Robert Gilbert UCLA Environmental Health Science,
• Michael Stealey UCLA Electrical Engineering

Undergraduates:

• Yi-Hui Lee UCLA,
• Breanne Bornemann UCR