Research Methods

Water Quality Parameters

During satellite overpass days, water samples were collected between the time window of 10:15 AM - 1:15 PM. Six liters of water were collected from each beach site at the coordinates listed below. Sample bottles were labeled with the date and location, kept on ice, and transported back to the lab within three hours. Generic temperature and pH probes, as well as a turbidity meter, were used to record values on-site at the time of collection. Turbidity measurements taken on-site were assumed to be the “true” turbidity values and are hereafter referred to as “on-site probe turbidity”.

A Hydrolab Multiparameter Sonde was used to measure pH, turbidity, conductivity, and % dissolved oxygen for each sample in lab. Turbidity and pH were measured using more than 1 device in our project in order to gather more comprehensive data

Click here to see the instrument used in lab and learn more about its application.

The Thermo Scientific Orion AQUAfast AQ3010 Turbidity Meter was taken to each sampling site to obtain turbidity values in the field. Turbidity values were reported by the meter in Nephelometric Turbidity Units (NTU). A clean, dry sample vial was handled by the top cap and rinsed with 10 mL of sample water three times. On the fourth fill, the vial was capped with 10 mL of sample water and wiped dry with a soft, lint-free cloth. A thin film of silicone oil was then applied and wiped around the vial. The vial was then inserted into the sample well of the meter for measurement by aligning the arrow on the meter with the arrow on the outside of the vial. The meter was then turned on, the “Read/Enter” button was pushed, and the measured value appeared several seconds later.

Three 1.5 µm filters were used to measure TSS for each site. Before use, filters were washed, baked for sterilization, and stored in a desiccator until needed. Initial weights of each filter were recorded immediately before running sample water through them. Then, a funnel manifold was set up with suction and filters were sealed on with deionized water. Between 100 mL-600 mL of sample water was passed through each filter (final calculations of weight accounted for the volume used). The filters were removed with tweezers, placed on individual aluminum pans, and dried for 24 hours in an oven. Lastly, filters were cooled in a desiccator and final weights were recorded.

To assess light absorbance, A spectrometer paired with the UV Express software were used to obtain light absorbance data. A baseline reading was established by adding 1 mL of MilliQ water to a glass cuvette in the spectrometer. After getting the baseline, a sampling cuvette was prepared and filled with 1 mL of sample water. Each sample was run from a frequency of 1100 nm to 190 nm, and the sample cuvette was rinsed 3 times with MilliQ water between each sample. After testing all of the samples, the data was exported from the software and saved.

Phosphorus:

Prepared sample tubes with the PhosVer 3 Phosphate reagent were labeled for each sample site. The “Phosphorus, Reactive (Orthophosphate)” test was used on the LAMBDA 365 UV/Vis Spectrophotometer with a light shield in cell compartment 2. 5.0 mL of sample water was added to the appropriate Test ‘N Tube Vial, capped, and mixed. The vial was wiped down to remove fingerprints and inserted into the 16 mm round cell holder to zero the spectrophotometer. One PhosVer 3 Phosphate Powder Pillow was added to the sample, capped, and shaken for 20 seconds. After two minutes, the tube was placed inside the spectrophotometer. The detectable range of PO43- for this test is from 0.06 to 5.00 mg/L PO43-.

Nitrate:

All square sample cells were cleaned with detergent and rinsed with deionized water before use. The “351 N, Nitrate LR” test was selected on the LAMBDA 365 UV/Vis Spectrophotometer. A 25 mL graduated mixing cylinder was filled with 15 mL of sample water, and one NitraVer 6 Reagent Powder Pillow was added. The cylinder was closed, vigorously shaken for 3 minutes, and then left to rest for 2 minutes. Once the timer sounded, 10 mL of the solution was added to a clean square sample cell. One NitriVer 3 Reagent Powder Pillow was then added to this cell and gently shaken for 30 seconds. The sample cell was then left to rest for a reaction period of 15 minutes during which time a pink color appeared if nitrate was present in the sample. While the 15 minute reaction period elapsed, a new square sample cell was filled with 10 mL of the sample water and ran to zero the spectrophotometer. After 15 minutes, the prepared sample was run through the spectrophotometer. The detectable range of NO3-- N for this test is from 0.01 to 0.50 mg/L NO3--N.

Ammonia-Nitrogen:

All square sample cells were cleaned with detergent and rinsed with deionized water before use. The “385 N, Ammonia, Salic.” test was selected on the LAMBDA 365 UV/Vis Spectrophotometer. One square sample cell was filled with 10 mL of the sample while another was filled with 10 mL of deionized water. The contents of one Ammonia Salicylate powder pillow was added to each square sample cell. The square sample cells were then shaken enough to dissolve the reagent and left to rest for 3 minutes. Once the 3 minutes elapsed, one Ammonia Cyanurate powder pillow was added to each sample cell, shaken until the reagent was dissolved, and then left to rest for 15 minutes. If ammonia was present, a green color would develop as the reaction occurred. Once 15 minutes had elapsed, the blank sample cell with the deionized water was wiped clean, inserted into the cell holder of the spectrophotometer, and zeroed. The square sample cell was then wiped, inserted, and read to show results in mg/L of NH3–N. The detectable range of NH3–N for this test was from 0.01 to 0.50 mg/L NH3–N.

Microbiology and IDEXX Methods

PBS Preparation: To begin, Phosphate Buffered Saline (PBS) Preparation is completed by weighing out 2.5 g disodium hydrogen phosphate, 0.58 g sodium dihydrogen phosphate, and 8.5 g sodium chloride which is then mixed into 1 L of reagent grade water and autoclaved at 121℃ for 15 minutes.

Cefotaxime Antibiotic Preparation (50 mg/mL): 250 mg of cefotaxime sodium salt and 250 mg antibiotic solution were added to 5 mL of reagent grade water. The cefotaxime antibiotic solution was divided into 50 μL aliquots and stored in PCR tubes at -20 ℃.

mTEC Agar Preparation: To prep mTEC agar plates, 9.12 grams of modified mTEC agar is weighed out and funneled into a 1000ml Kimax bottle with one large stir bar. Per 9.12 grams of mTEC agar add 200ml of reagent grade water and stir with heat for 5-10 minutes till all agar powder has been dissolved. Autoclave your dissolved mTEC agar mixture and an emppty 500ml Erlenmeyer flask with small stir bars and a strong mark at 100ml at 121 ℃ for 15 minutes. Wait for agar to cool to 40-50 ℃ before adding 100ml to the 500ml Erlenmeyer flask and spike with 8 μL of cefotaxime antibiotic stock solution. The antibiotic spiked media was poured into 50 mm plates (4 mL per plate) and labeled with “(+).” For the control plates without antibiotic, the remaining 100 mL of mTEC agar media was poured into 50 mm plates (4 mL per plate) and labeled “(-).”

Membrane Filtration: A 0.45 μm membrane filter was placed in a filter cup, followed by a 0.20 μm membrane filter. Then, 30 mL of PBS followed by 3 sequential dilutions of sample water were added and vacuumed through the filter cup. The sides of the filter cup were rinsed with 5-10ml of PBS, and the 0.20 µm membrane was discarded. Finally, the 0.45 µm filter was transferred onto an mTEC agar plate. This process was completed 4 times for each sample site, producing two antibiotic spiked (+) plates and two control (-) plates.

Plate Counting Analysis: After incubation, a countable dilution was used and compared with the number of purple dots (bacteria CFU) on each plate in order to obtain a value for the average E. coli CFUs/100mL of sample water. hi

By dividing the total number of cefotaxime-resistant E. coli for each sample by the total of E. coli for each sample, we obtained the proportion of E.coli that is resistant to cefotaxime. hi

IDEXX Method:

The IDEXX method for enumerating FIB was performed according to the manufacturer’s instructions. Total coliforms (TC) and E. coli, antibiotic-resistant coliforms and ESBL-E. coli, and enterococci were measured for each sample site, equating to 3 solutions prepared for analysis per site. Solutions were prepared by adding chemical substrates (Colilert and Enterolert reagent packets) to a 1:10 dilution ratio of sample water to MilliQ water.

This dilution ratio was based on historical ocean water quality data and allowed for countable results. To measure antibiotic-resistant coliforms and ESBL-E. coli, 100 µL of cefotaxime antibiotic syrup was added to one of the solutions. The final solutions were then poured into multi-well trays and sealed for incubation.

After incubating the Colilert trays for 18-22 hours at 35°C and the Enterolert trays for 24 hours at 41°C, the samples were examined for the presence of yellow color or fluorescence. Coliforms were indicated by yellow wells, while E. coli and enterococci were indicated by fluorescence under UV light. A most probable number (MPN) table provided by IDEXX was used to convert the number of wells positive for fluorescence or yellow color to the MPN of bacteria (CFU/100 mL). This value was then compared to manual colony forming unit (CFU) counts from the pour-plate method.

Remote Sensing

Python 3.10.2, Anaconda Navigator, and ACOLITE (a generic atmospheric correction module) were used to analyze satellite images over our sample sites. First, remote sensing images were downloaded from USGS Earth Explorer from the Sentinel-2 satellite or Landsat-8 satellite for the date of interest. Cloud cover was restricted to 60%. The image was downloaded as a “L1C Tile in JPEG2000.” These images were then run through the ACOLITE interface to extract a .tif file. In order to obtain turbidity data, this file was run through a Python script to extract mean turbidity (in FNU) and standard deviation for each sample site.

Similar to the ACOLITE workflow, remotely sensed images were imported into Google Earth Engine. On this interface, a Python script returns turbidity data (in NTU) for the sample area. Using the coordinates of each sample site, turbidity values were recorded and compared to ACOLITE results to investigate potential correlation.

Due to the impact seasonal weather could have on measured values, the relationship between the measured parameters and upwelling indices was also explored in our models. Data was obtained from NOAA’s SWFSC Environmental Research Division using their “Upwelling Index, 33N 199W 6-hourly” database.