Home

Technical!Report

1. Figure 19 SUNA spectra illustrating that a the optical path being blocked by broken O ring top figure b the sensor is working correctly bottom figure Figure 20 Example protective cage made of deer fencing and rebar Figure 21 Nitrate specific conductance and temperature data recorded every 30 minutes at DR3 Figure 22 Observations of stream flow temperature nitrate and specific conductance at DRKR measured every 30 minutes during base flow and storms October 2010 Figure 23 Nested watershed behavior of nitrate concentrations measured every 30 minutes under base flow conditions and during a storm vi 34 39 40 42 49 50 50 Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Tables Characteristics of study watersheds Observed average pump flow rates with tap water at 19 5 C Head lift and tubing length at each station Current drawn for sensor station components Calculation of battery replacement schedule Sizing of solar panel using http www batterystuff com solar calculator html Summary of wiring connections for CR10X and sensor set up vil Page 13 14 15 23 Abstract Real time water quality sensors were deployed at six stream stations in Dead Run Baltimore MD beginning in Fall 2010 This report summarizes initial experiences design variable considerations and lessons learned from one year of deployment Satlantic SU
2. Table 2 Program 02 0 0000 Execution Interval seconds Table 3 Subroutines 1 Beginning of Subroutine P85 1 1 Subroutine 1 Subroutine to turn on Suna pause Sna_secON seconds turn on pump pause Pmp_secON seconds take a measurement and then turn off pump and Suna Suna is on for Sna_secON Pmp_secON seconds Pump is on for Pmp_secON seconds 2 Z F x 10 n P30 Zero Suna counter 1 0 F 2 00 Exponent of 10 3 17 Z Loc Sna countl1l 17 is the location vve chose for the SUNA elapsed time count 20 is the location 3 Z F x 10 n P30 Zero pump counter 1 0 F 2 00 Exponent of 10 3 20 ZiLocl l Pmp countl for the pump elapsed time count 4 Set Port s P20 pump ON C5 is the channel location for 1 9991 C8 C5 nc nc nc high 2 9999 1 nc nc nc nc PUMP This loop executes once per second It exits when Pmp_count gt Pmp_secON 5 Beginning of Loop P87 Delay to run pump before taking measurement 1 0 Delay 2 0 Loop Count The datalogger 6 Excitation with Delay P22 programming requires a 1 1 Ex Channel finite delay to be specified 2 0 Delay VV Ex 0 01 sec units for the excitation channel 3 100 Delay After Ex 0 01 sec units to become active The 4 0 mV Excitation pump stays on for a specified amount of time 7 Z Z 1 P32 before SUNA is switched 1 20 ZLoc Pmp_count oe 62 8 If X lt gt Y P88 1 20 XLoc Pmp_count 2 3 gt 3 19
3. A30R3012 HCR 687 685 799 637 293 795 580 031 370 328 7102002 507 765 193 844 232 777 294 101 189 502 641 344 472 476 189 170 189 138 276 902 391 417 543 570 295 455 777 609 9224T87 274 356 300 888 57 Hoffman http www hoffmanonlin e com Home Depot Home Depot Home Depot Home Depot Home Depot USGS HIF Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot McMaster Carr Home Depot Home Depot Home Depot 753 60 8 97 4 50 2 00 13 96 4 98 11 00 3 98 3 73 0 98 1 21 1 98 1 36 1 21 1 66 1 27 3 36 4 73 5 64 1 69 4 99 9 05 2 65 5 41 6 51 Angle slotted galvanized metal xx ft Hex nuts 3 8 Hex bolts 3 8 Cut washers 3 8 Plywood 15 32 or 2 2 ft x 4 ft Wood screws for attaching relays ravens to plywood FH PH 6 x 3 4 Wood screws for attaching terminal connectors to plywood FH PH 6 x 1 Shims pack 655 449 655 570 300 896 251 356 251 380 879 282 58 Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot Home Depot 9 43 0 11 0 18 0 13 9 72 3 77 3 97 1 35 Appendix C Program for CR10X datalogger CR10X William J Davies USGS This program was written by Bill Davies of USGS and 09 03 10 modifie
4. Figure 11 taken from field sites where the sensors are to be deployed and established a correlation between specific conductance and chloride for use in subsequent analysis Figure 12 24 The purpose of the three trials using nitrate standards shown in Figure 8 was to test for memory effects in the sensor Each sequence was composed of low concentration nitrate standards followed by higher concentrations then low again In all cases the concentrations ranged from 0 01 mg L to 10 mg L The results show good agreement between IC and SUNA measurement of the standards Sequence 1 y 0 0247 0 945x R 0 999 Sequence 2 y 0 0721 0 932x R 0 998 Sequence 3 l y 0 0595 0 945x R 0 998 SUNA Value mg N L 0 2 4 6 8 10 12 IC Value mg N L Figure 8 SUNA vs IC for several nitrate standards low high sequences To test for ambient temperature effects that would be expected for the SUNA under field conditions the sensor was placed in a freezer at 15 C and in an incubator at 6 C 20 C and 40 C SUNA measurements of four nitrate standards 0 01 0 1 1 and 10 mg L were compared to those run on the IC at room temperature as would be the case for samples brought back to the lab The results Figure 9 indicate that temperature had little effect on the sensor nitrate readings compared to analyzing the samples at room temperature using the IC Because the SUNA is an optical probe it was desirable to
5. We expect to 1 21 Doo if Flag 1 is Low get no more than 1 tip in 2 2 9 Call Subroutine 9 seconds this serves as a flag for QA QC 2 Batt Voltage P10 measure battery volts 1 3 Loc Batt_Volt On 9 29 10 changed multiplier from 1 to 01 since this is how it is in the old program 3 Pulse P3 Read rain gages For the rain gage the 1 2 Reps multiplier slope chosen is 2 1 Pulse Channel 1 0 01 referring to 1 tip 3 2 Switch Closure All Counts 0 01 inches of rainfall The 4 22 Loc Rain 1 offset is the y intercept 5 01 Multiplier which is zero for our case 6 0 0 Offset sum the 2 rain gages so when we do the lt gt comparison we see if either gage tipped I m always nervous about doing or lt gt comparisons with floating point numbers so leave rain as integer tips Multiply by 01 before storage 4 Z X Y P33 22 XLoc Rain1 23 YLoc Rain 2 24 Zocl Rain Sum This sums the values of precipitation at the tvvo rain gages If either rain gage has tipped the value WN FR is recorded This avoids 5 If X lt gt F P89 1 24 XLoc 1 Rain Sum recording zeros for periods 2 2 lt gt of no rain 3 0 F 4 30 Then Do E On 9 29 10 took out P37 which Bill had because other program did not have this 6 Do P86 1 10 Set Output Flag High Flag 0 0On 9 27 10 Julia changed Array ID here to 100 was previously 300 7 Set Active Storage Area P80 45717 60
6. with deionized water making sure there are no bubbles on the SUNA s surfaces The SUNA must be connected to a computer and SUNACom opened The SUNA must be set to the RS 232 full binary or full ASCII mode not SDI 12 mode On the main SUNACom Dashboard Window Update Calibration should be clicked and then the prompts followed When completed a graph will appear that shows the percent error at each wavelength between the new and previous calibration If there are wavelength s with high error the plastic wrap can be removed and the process repeated If both calibrations were performed correctly this second calibration update should show little 47 error 4 4 Field installation of Ravens The magnetic hockey puck antenna is mounted on top of the junction box and its cable threaded through the hole in the bottom of the box The power cord is connected to the modem and its black wire is inserted into any G slot ground of the CR10X that is located the lower right corner of the panel and the red wire into the 12V CR10X slot directly next to the G slot The modem is connected to the interface using the 10873 serial cable RS 232 The interface is connected to the CR10X using the SC12 cable This is illustrated by the figure on page 6 of the instruction manual Campbell Scientific Inc 2006 The indicator lights on the modem should be checked The power light and network light should be a solid green A solid signal light indica
7. 1 1 Final Storage Area 1 2 100 Array ID 8 Real Time P77 42212 1 1111 Year Day Hour Minute Seconds midnight 0000 9 Sample P70 431631 reps refers to the number of 1 2 Reps rain gages for which to carry out 2 22 Loc Rain 1 the same commands in this case 7 2 10 End P95 Measure stage and record every 5 min 0On 9 27 10 Julia changed to every 5 minutes was every 15 11 If time is P92 1 0 Minutes Seconds into a This is an instruction to record the stage 2 5 Interval same units as above t zero defined as on the hour and 3 2 Call Subroutine 2 every 5 minutes using subroutine 2 Measure Suna every 30 minutes on the hour and half hour We can change this code later to allow changing the measurement interval without having to re program This is an instruction to record SUNA 12 If time is P92 data at 0 start and 30 minutes after 1 0 Minutes Seconds into a 0 This can be changed remotely as 2 30 Interval same units as desired above 3 1 Call Subroutine 1 Julia added on 9 39 10 taken from Bill s program Rain 501 V3 totalize tips on rain gage 1 From my understanding this is just so tips can be read when connected to the datalogger without downloading data 13 Z X Y P33 1 25 XLoc Rain_tot1 2 22 YLoc Rain_1 3 25 ZLoc Rain_tott1 totalize tips on rain gage 2 14 Z X Y P33 1 26 XLoc Rain_tot2 2 23 YtLoc Rain_2 3 26 ZLoc Rain_tot2 61
8. 1 72 coupler connection to aid in preventing the connection from breaking apart Before any trenching or pipe fitting we first determined the best location for the bilge pump to be placed in the stream We wanted the pumps to be as close to the enclosures as possible but placed in flowing water not pools and such that they could be protected in a storm i e braced with rebar and rocks These also needed to be placed in locations where the perforated pipe would be submerged during low flow conditions which may be as shallow as 3 or 4 inches in urban streams Typically the best locations were 2 or 3 feet from the bank At DR Franklinton and DR5 concrete aprons were present in the streams and the PVC pump housings could be butted up against the aprons and held in place by rebar Figure 17 At DR3 and DR4 the PVC runs along galvanized pipe that was previously deployed by USGS to protect the Accububble pressure transducer The PVC was secured to the pipe with wire ties A large tree with an exposed root system near the bank is present at DR1 we placed the PVC pipe through this root system for support At DR2 the stream is narrow near the testing location here we dug into the bank and used the earth as a support for the PVC 33 Figure 17 Pump housing deployment at DR Franklintown Once the housing location is determined a mock setup should be carried out by laying out the entire length of pipe from the bilge pump housing to the encl
9. 30 4 1 1 Enclosures 4 1 2 Trenching and pipe deployment 4 1 3 Tubing considerations 4 1 4 Solar panel deployment 4 1 5 Wiring 4 1 6 Grounding rod 4 2 Troubleshooting sensors 4 2 1 Blockage of SUNA optical path 4 2 2 Pump under oversizing 4 2 3 Tubing kinks 4 2 4 Leaf trash and periphyton clogs 4 2 5 Batteries dying solar controller and datalogger problems 4 2 6 Ice 4 2 7 Interpretation of error codes and erroneous data values 4 2 8 Testing pumps 4 2 9 Priming pumps 4 2 10 Relay shorting 4 2 11 Backflow bubbles 4 3 Maintenance 4 3 1 Power 4 3 2 Tubing 4 3 3 Pump replacement 4 3 4 Cleaning solar panels 4 3 5 Cleaning sensors 4 3 6 Site check after storms 4 3 7 SUNA calibration 4 4 Field installation of Ravens 5 Example data sets 6 Summary and recommendations References Appendices Appendix A Calculated head losses H hL Az for 4 5 A and 6 A pumps at 0 C and 30 C using Equations 1 3 for flow rates observed at 12 V and 10 8 V Appendix B Parts list Appendix C Program for CR10X datalogger 30 32 35 35 36 37 37 37 38 41 41 43 43 44 44 45 45 45 46 46 46 46 46 46 47 47 48 48 51 52 53 53 55 59 Figures Figure 1 Study watersheds for sensor project a Dead Run 14 1 km b Gwynns Falls 171 km2 and nested subwatersheds Figure 2 Pump flow rate dependence on battery voltage power watts voltage volts current amps Figure 3 Effect of tubing diameter
10. 5 A pump with a stainless steel impeller that has an expected service life of approximately 400 hours 250 days for our application We have deployed these pumps as an alternative to the Rule pumps to test their durability Of the 6 Wale pumps initially installed one pump stopped working after 30 days Another pump needed to be replaced after 50 days because it was having difficulty self priming Others have lasted for 150 days without any problems 2 2 3 Y configuration and location of sensors The water entering the protective enclosure via tubing from the pump is split between the SUNA and YSI using a Y connection so that the sensors are taking readings of the same water sample the YSI takes one reading whereas the SUNA takes multiple 9 readings and averages them Because the YSI flow cell is open to the atmosphere the top of the flow cell must be placed above all parts that contain water above all tubing and above the SUNA flow cell If it is lower than these components it will be at a higher pressure and will leak from the top regardless of the flow rate It is also important to make sure that the outflow tubing does not become crimped or clogged If this happens backpressure can result causing the YSI flow cell to leak from the top 2 2 4 Filtering Due to the potential for turbidity interference with the SUNA reading USGS Brian Pellerin personal communication July 2010 recommended filtering the water in line before
11. Port 4 connects to the common ground An excitation signal is sent to Port 3 at a time specified in the datalogger This allows the power supply at Port 2 to power the device connected to Port 1 When the signal is removed from Port 3 the power supply from Port 2 is no longer connected to Port 1 and the device shuts off The typical labeling of each port is the convention used here Labeling varies by manufacturer and may differ for various countries 2 4 Data logging Our choice of datalogger was dictated by those already in place for other projects at the chosen sensor sites We operate a set of streamflow gages in cooperation with the USGS MD DE DC Water Science Center In order for water level data from these stations to be transmitted via Raven cellular modem to the USGS office and ingested into the USGS data management system that office currently requires use of Campbell Scientific CR10X dataloggers for compatibility with its existing system Since our nitrate and conductivity sensors are co located at stations where these dataloggers are already in place we chose to use the CR10X sensors for this project This model of datalogger has been retired from Campbell Scientific but is available to us for rental from the USGS Hydrologic Instrumentation Facility If we were conducting the project without this constraint we would recommend use of the CR1000 datalogger which is currently available from Campbell Scientific A datalogger pr
12. USGS for the purpose of housing Accububble pressure transducers 12V batteries rest on the bottom of the box We lined the interior sides and back of each box with pieces of in thick plywood so as to be able to mount the sensors Raven datalogger and SunSaver on the inner faces of the box Figure 13 Figure 13 Interior of Hoffman box 30 Two holes were drilled in the bottom of each box to allow placement of the inflow and outflow vinyl tubing from the sensors The outflow tubing is shielded by a 1 5 in PVC pipe running vertically downward from the box and then horizontally trenched and buried to direct the outflow water away from the sensor station and onto the ground downstream of the pump intake Holes are also needed in the Hoffman box for the Raven antenna the grounding wire and the solar panel wires Foam sealant was used around all holes to keep moisture out to the extent possible At the one station where a Hoffman box was not available we chose to use a 55 gallon drum as the sensor enclosure Figure 14 The drum is bolted to part of the wingwall of a bridge culvert the top is removable and lockable and a sensor package can be slid into this enclosure This had previously been designed to hold an ISCO sampler for use in another project Ken Belt USFS personal communication 2010 For this set up we designed a cage made out of galvanized steel strips bolted together and further reinforced with a piece of plywood along on
13. YlLocl Pmp secON 1 4 31 Exit Loop if True 9 End P95 nc no change to c5 pump i stays on c4 is the channel 10 Set Port s P20 Turn Suna ON where He relay is 1 9999 8 5 nc nc nc nc connected that turns on 2 1999 4 C1 high nc nc nc the SUNA This loop executes once per second It exits when Sna_count gt Sna_secON 11 Beginning of Loop P87 1 0 Delay 2 0 Loop Count Again a delay is required to turn on the SUNA 12 Excitation with Delay P22 1 1 Ex Channel 2 0 Delay W Ex 0 01 sec units 3 100 Delay After Ex 0 01 sec units 4 0 mV Excitation 13 Z Z 1 P32 1 17 ZLocl Sna 14 If X lt gt Y P88 1 17 Loc Sna countl 2 3 gt 3 16 YLoc Sna_secON 4 31 Exit Loop if True 15 End P95 set default values to 7 if no comms first val is 6999 16 Beginning of Loop P87 1 0 Delay 2 3 Loop Count 7 is a flag set to mean no communication between SUNA and datalogger The 17 Z F x 10 n P30 default value is 6999 we 1 7 F changed this to 7 2 0 n Exponent of 10 3 12 ZLoc Sna_N_mgL 18 End P95 Measure Suna returns NO3 in micro moles NO3 mg L Avg Spectrum Dark frame value 19 501 12 Recorder P105 1 0 SDI 12 Address 2 0 Start Measurement aM 3 3 Port 4 11 Loc Sna_N_mM 5 1 0 Multiplier 6 0 0 Offset O is the SDI address for the SUNA the SUNA measures 4 parameters starting with NO3 in micromoles Me
14. against power outages that could cause interruption of data streaming The computer must remain turned on and LoggerNet must remain open in order to connect to a modem and download data Using LoggerNet a data collection schedule is enabled from this central computer We suggest that the status monitor window remain open This allows data downloading to occur and also allows all the networks that have been setup in LoggerNet to be easily monitored Additional details can be viewed using the Log Tool For our sensors LoggerNet is set up to connect and download data from the CR10Xs every hour 2 5 3 Data archiving and serving Once the data is downloaded using LoggerNet it is then transferred to the CUAHSI Consortium of Universities for the Advancement of Hydrologic Sciences Inc Observations Data Model ODM http his cuahsi org odmdatabases html and made available for publication using CUAHSI WaterML Web Services http his cuahsi org wofws html A number of different software components are used to format and transfer the data to the ODM and publish the data from the WaterML Web Service on the web A workflow showing the connected software components is depicted in Figure 5 The data from LoggerNet must first be transformed into a format that can be ingested into the ODM because the CR10X datalogger does not produce a time stamp that conforms to the date time data type in the ODM A Python script is used to convert the 18 CR10X time s
15. and temperature on friction loss at DR2 for the 6 amp Rule pump Figure 4 Sequence of events required for taking sensor measurements and storing the results Figure 5 System architecture for serving nitrate sensor data Figure 6 Prototype sensor data visualization website Figure 7 Wiring diagram for sensors Figure 8 SUNA vs IC for several nitrate standards low high sequences Figure 9 Temperature effects on nitrate measurement Figure 10 DOC effects on nitrate measurement Figure 11 Measurements of base flow nitrate concentrations from grab samples taken at station locations and analyzed in the laboratory using the SUNA and ion chromatography Figure 12 Correlation between specific conductance measured using the YSI sensor and chloride measured using ion chromatography for stream samples taken at Dead Run Franklintown Figure 13 Interior of Hoffman box Figure 14 55 gallon drum that can be used as an enclosure for the sensor package a with lid off b secured Figure 15 Cage constructed to hold sensor package inside 55 gallon drum Figure 16 Schematic of PVC pipe pieces used to protect pump and pump tubing V Page 12 19 20 21 25 26 26 28 28 30 31 32 33 Figure 17 Pump housing deployment at DR Franklintown Figure 18 Example field data illustrating that a the optical path being blocked by broken O ring top figure b the sensor is working correctly bottom figure
16. at all locations for the purpose of powering the bilge pump via a 12V battery One 30 watt solar panel was already in place at each station to power the Accububble pressure transducer as well as the datalogger and sensors A mounting bracket for the 30W panel allows the solar panel to be mounted against a wall horizontally or vertically or attached to a pole The bracket also allows the panel to be tilted with a range of motion of almost 90 We mounted the solar panels to existing galvanized steel poles previously deployed by USGS to hold satellite data communication antennas The panels were attached to the pole using two stainless steel hose clamps that could accommodate a 2 pipe Our goal was to mount the solar panels high enough to minimize the potential for vandalism At two of our stations the existing poles had to be extended which was accomplished by 35 attaching a section of 2 pipe to the top of the existing pipe using a coupler The coupler contained all female threads and was screwed to the top of the existing pipe which was already male threaded The extension pipe was then threaded into the top of the coupler In cases where a new pole is needed a hole needs to be excavated at least 2 ft deep and filled with fast drying concrete e g Sakrete 50 1b in which a 10 ft 2 in galvanized steel pole can be placed At one site our solar panels needed to be mounted against a bridge To do this masonry wedge anchor bol
17. debris and trash The copper mesh and cage are described in more detail in Section 4 2 4 4 1 3 Tubing considerations Tygon tubing was used to connect the pump to the sensors because it is flexible and resists kinking and breakage In several cases we used vinyl tubing as a substitute and this proved to be harder to work with All tubing that is pressure attached to barb fittings should be secured with small hose clamps for extra security While Tygon tubing is flexible and resists kinking and breakage it will show signs of biofouling over time as will most plastic tubing As an alternative food grade tubing which is more resistant to biofouling can be used however food grade tubing is stiff and difficult to run through the PVC pipe from the pump to the enclosure Copper pipe could be used to supply water to the sensors however this could also be difficult to run from the pump to the enclosure The choice of tubing should be made based on site accessibility with the understanding that anything other than copper or food grade plastic will show biofouling over time A Y connection can be used to split the water equally from the inlet tube to both sensors The Y split should be placed near the bottom of the box enclosure Once the Y connection is in place the tubing can be cut to length from each sensor to the Y connector 4 1 4 Solar panel deployment One 30 watt unbreakable solar panel manufactured by Power Up Solar was deployed
18. determine whether there could be any interference from either color due to dissolved organic carbon DOC from humic substances or turbidity Concerns had been raised by other groups U Florida USGS that both of these could possibly interfere with SUNA nitrate measurements The DOC of waters in our system typically does not exceed 15 mg L We used International Humic Substances Society IHSS Suwanee River Fulvic Acid standards to prepare various controlled concentrations of DOC ranging from 0 1 to 15 mg L and 25 formulated a matrix of mixtures with nitrate standards ranging from 0 to 10 mg N L The results in Figure 10 indicate that this range of DOC from humic substances had little effect on nitrate concentrations measured with the SUNA vs the IC 12 25 15 C y 0 0548 0 9733x R e1 10 6 y 0 0189 0 98 R 0 995 8 SUNA Value mg N L gt 200 y 0 0335 0 954x R 0 999 40 C y 0 0252 0 972x R 1 0 2 4 6 8 10 12 IC Value mg N L Figure 9 Temperature effects on nitrate measurement 12 0 1 mg L DOC i y 0 101 1 25x R 0 999 8 1 0 mg L DOC y 1 38E 04 1 24x R 0 995 SUNA Value mg N L m gt 5 0 mg L DOC y 2 ATE 02 1 25x R 0 968 15 0 mg L DOC y 6 43E 02 1 19 R 0 999 2 4 6 8 10 12 IC Value mg N L Figure 10 Effects of DOC from humic substances on nitrate measurement 26 To evaluate
19. front of the pump housing and 2 making a cage upstream of the pump housing using plastic deer fencing Figure 20 Figure 20 Example protective cage made of deer fencing and rebar The rock barriers can divert leaves away from the pump housing while still allowing for water to enter it Determining the optimal arrangement is a trial and error process We found that after storm events the rocks often moved and a new barrier had to be built A deer fencing cage can be built with 2 ft rebar spaced about 1 5 to 2 ft apart driven into the streambed until the rebar is about 10 above the water level at baseflow The deer fencing can be pulled tightly against the rebar and attached with zip ties Rebar can also be also laid against the streambed and attached to the deer fences to prevent debris from flowing under the cage If the cage is built near a stage measuring device such as a pressure transducer it must be ensured that as the cage collects debris the water level does not back up such that stage measurements are erroneously increased As with pumping sizing problems identified in the previous section if water does not reach the flow cells owing to problems with clogs of the pump intake the sensors will not receive water and therefore will not record water quality One sign of a pump 42 intake clog is that sensors for nitrate and specific conductance both go toward zero but the temperature continues to be recorded by the YSI The da
20. is the goal to overcome the problems with turbidity interference on nitrate readings Geoff Maclntyre personal communication December 2011 The final phase of laboratory testing included an initial round of measurement of stream samples at the sensor test sites taken manually under base flow conditions using the IC and the SUNA on samples in the laboratory Figure 11 With the exception of one point results for nitrate concentration using both methods show good agreement A correlation between specific conductance and chloride for samples taken at DRKR is shown in Figure 12 This information was required for mathematical analysis being conducted for this project e g VerHoef et al 2011 27 3 0 2 5 2 0 Villa Nova 2 Carroll Park 15 5 a 1 0 ad Run Franklintown 0 50 y 0 083691 0 97628x R 0 98934 0 0 0 0 0 50 1 0 1 5 2 0 2 5 3 0 IC Value mg N L Figure 11 Measurements of base flovv nitrate concentrations from grab samples taken at station locations and analyzed in the laboratory using the SUNA and ion chromatography 1 4 1 2 1 0 11 10 10 YSI SC mS cm a 10 20 10 0 40 0 20 y 0 154 0 00415x R 1 000 0 0 0 50 100 150 200 250 300 BES Weekly Sample IC Value CI mg L Figure 12 Correlation between specific conductance measured using the YSI sensor and chloride measured using ion chromatography for stream samples taken at Dead Run Franklint
21. paragraph in the discussion of LoggerNet If no IP address or the wrong IP address is shown in the Status window then the user should go 17 to the Misc group and enter the static IP address in the Force static IP value The phone number for the SIM card is entered under the new value cell for phone number Next AT Verbose Mode should be set to 0 numeric when using CR10X dataloggers Then under Cellular window APN is set to I2GOLD Once all these values are entered write can be clicked again to program these changes to the modem When the modem is finished being programmed disconnect can be chosen to end the session 2 5 2 Base station communication with datalogger Campbell Scientific dataloggers can be set up and configured using the Campbell Scientific program LoggerNet The connection type selected for communication with the datalogger should be IP Port connection The internet IP address is the assigned static IP address of the SIM card plus the port number Per the discussion above the assigned port number becomes 3001 after the Campbell Scientific template has been loaded and programmed into the modem A dedicated computer in a central location is recommended for connecting to the sites and downloading data This ensures that the downloading of data is not interrupted by other uses of the computer We also recommend that the computer be connected to an uninterruptible power supply UPS to protect
22. that section 16 gage wire was used to connect all components Red wire was used as power positive black was used as ground negative and blue was used as excitation channel or data transmission wire In terms of wiring order this is somewhat arbitrary but we found that the following order worked well for our set up 1 connections between the relays and datalogger 2 connection of the bilge pump and SUNA to the relays 3 connection of the YSI and other SUNA wires to the datalogger and 4 connection of the Raven to the datalogger The SunSaver solar controllers should be connected last Per the SunSaver manual the 36 battery should first be connected to the SunSaver followed by the solar panel wires and lastly the load Each port is labeled on the SunSaver 6 It is important to not touch any exposed wires together other than the following 1 Ground wires can touch and multiple wires can go in the same ground port on the datalogger 2 The YSI power and Raven power can go in the same 12V port or in separate 12V ports on the bottom right of the datalogger 3 The positive load for the SunSaver and relay port 2 on the SunSaver and relay used for the datalogger and touch and go in the same 12V power in port on the top right of the datalogger These connections are all shown in Figure 7 It is important to note that the wire connected to Port 1 on the relay should NOT be disconnected or reconnected if the system is supposed to be runn
23. the sensors Our testing of the USGS recommended filter system see parts list in Appendix B resulted in identification of two problems 1 increased head drop due to resistance to flow and 2 significant flushing time required to purge the filter housing Because the filter consists of a fine 0 2 micron mesh it greatly reduces the flow rate of the system since the filter provides resistance to flow To compensate for this effect a larger pump would be required to maintain a desired flow rate which would draw more power The volume of the recommended filter was 1 8 liters A concern with using such a large filter is that water would be stored in the filter between runs This water would need to be fully purged for a subsequent trial to obtain a water quality value that was not averaged with water from the previous trial Laboratory trials using water spiked with known concentrations of NaCl showed that it took between 4 and 5 minutes of pumping time to completely purge the filter For this reason the filter apparatus was not used during the field trial 2 2 5 Determination of on off time for pump and sensors Operational characteristics and requirements of the sensors as well as sampling requirements at the site need to be factored into determining the pumping cycle required We noted above that for the first test deployment we wanted to take measurements every 30 minutes We chose to turn the pump on before each measurement as oppos
24. variable Pmp secON So the total on time for the suna before a measurement is Sna_secON Pmp_secON seconds 47 Z F x 10 n P30 115 F 2 00 Exponent of 10 3 16 ZLoc Sna_secON Pump warm up time in seconds Jason changed F to 50s 48 Z F x 10 n P30 Default pump purge time in seconds 1 50 F 2 00 Exponent of 10 3 19 ZLoc Pmp_secON 49 Set Port s P20 Thi 1 9997 C8 C5 nc nc nc output y lt 2 7999 C4 C1 output nc nc nc v n 50 End P95 End Subroutine 51 Beginning of Subroutine P85 1 98 Subroutine 98 Julia added on 9 30 0 subroutine to do the SDI 12 sensor thing as Bill says in Rain SDI V3 Has to be subroutine 98 Basically allows tips to be seen when connected to the datalogger without downloading data 52 501 12 Sensor P106 1 1 SDI 12 Address 2 0204 Time Values 3 27 Loc Rain sdi11 put rain totals into locations for SDI 12 output 53 Z X P31 1 25 XLoc Rain_tot1 67 2 27 ZilLocl Rain sdi11 54 Z X Y P33 1 25 XLoc Rain_tot1 2 29 Y Loc RnSD Tot1 3 29 21 1 RnSD Tot11 55 Z X P31 1 26 XLoc Rain_tot2 2 28 21 sdi21 56 Z X Y P33 1 26 XLoc Rain_tot2 2 30 YLoc1 RnSD Tot21 3 30 Z Loc1 RnSD Tot21 srezero totalizers 57 Z F x 10 n P30 10 0 F 2 00 Exponent of 10 3 25 ZiLoc nRain tot11 58 Z F x 10 n P30 1 0 0 F 2 00 Exponent of 10 3 26 ZLoc Rain_tot2 59 End P95
25. 0X e 6 12V Power In on CR10X SunSaver 6 for bilge pump e 1 Battery for bilge pump negative e 2 Battery for bilge pump positive 3 Solar panel for bilge pump negative 4 Solar panel for bilge pump positive 5 10 6 2 for bilge pump 2 7 Parts chart costs and itemization An itemization of all components used to build a sensor station excluding tools is provided in Appendix B 3 Laboratory testing Two activities comprised laboratory testing 1 chemical analyses to compare ion chromatography with the SUNA and identify possible interferences with the sensor memory effects temperature colored DOC turbidity and 2 a dry run of the hydraulics wiring data logging and communication systems to ensure that all items were working properly in advance of deployment 3 1 Chemical QA QC The weekly stream chemistry sampling that has proceeded for 13 years in the Baltimore Ecosystem Study employs ion chromatography for anion water quality analysis In order to evaluate whether the optical nitrate sensor will provide acceptable quality data as compared to the IC analysis we ran a series of laboratory tests to evaluate the IC versus the SUNA for 1 three trials of 7 concentrations of nitrate standards Figure 8 2 temperature effects Figure 9 3 effects of dissolved organic carbon from humic substances Figure 10 and 4 effects of turbidity In addition we analyzed samples
26. 18 2011 01 32 00113 44 October 18 2011 02 02 0010 91 October 18 2011 02 02 00 October 18 2011 02 02 00 13 35 October 18 2011 02 32 00 0 91 October 18 2011 02 32 00 October 18 2011 02 32 00 13 28 DR5 Date and Time EST Stage ft Discharge cfs October 18 2011 00 00 00 0 450 372 October 18 2011 00 05 00 0 450 1 0 372 October 18 2011 00 10 00 0 450 0 372 October 18 2011 00 15 0010 450 0 372 October 18 2011 00 20 00 0 450 0 372 October 18 2011 00 25 00 0 450 0 372 Site Information Site ID DR 5 DR5 USGS Site ID 01589312 Location 39 17 45 2 N 76 44 38 7 W Baltimore County Hydrologic Unit 02060003 on right bank at upstream side of culvert on Black Friars Road 1 1 mi north of Catonsville 1 7 mi southwest of Woodlawn and 1 8 mi west of Baltimore City Period of Record November 2010 to present Remarks Records for this station are managed by the Center for Urban Environmental Research and Education University of Maryland Baltimore County For any inquiries or to report malfunctions regarding the sensors contact Julia Miller julia umbc edu Funding Funding for the operation of this station is provided by National Science Foundation Web Design Roxanne Sanderson UMBC Conducted in partnership with USGS and the Baltimore Ecosystem Study Figure 6 Sensor data visualization website 20 Upon navigatin
27. 250 00 1 667 25 295 00 576 00 157 50 76 50 76 50 122 00 260 00 108 00 72 00 35 69 44 00 189 00 159 00 292 50 90 00 20 00 15 00 88 31 Wiring CR10X modem monthly Campbell rental Scientific CRYDOM Relay Solid state D1D12 Double rovv terminal block 4 POS 1 Double row terminal block 6 POS 1 Wire connectors 22 18 AWG Spade vinyl Stud 8 10 100 Wire black 16 AWG 24 ft Wire red 16 AWG 24 ft Wire blue 16 AWG 24 ft Electrical tape 3 4 x 30 ft Heat shrink tubing 3 8 3 16 Heat shrink tubing 1 2 1 4 Aim n flame II Hydraulics 12V Submersible Pump 6 amp 12V Submersible Pump 5 amp 12V GP In line water pump 5 amp Tygon tubing 3 8 in 50 ft clear Vinyl tubing 1 2 OD 3 8 ID x 10 ft Filter housing 3 4 12 222 Opaque w double plugs O ring for filter housing Mounting bracket MC 1 Kit Filter FP 02 10 A PP Pleated 12CS Couplings 3 8 barb 10 pack Y connectors Barbed Y 3 8 10 pack Hose clamps 006 SS clamp 3 8 x 7 8 dia 10 pack Hose clamps 036 SS clamp 1 3 4 x 2 3 4 dia 10 pack 1DTL7 296 187 296 748 299 850 710 914 710 797 710 900 225 835 542 415 542 458 233 989 P Congo P Amazon GP1692 2DCA3 702 229 PL 150512 285 81 005 276 13 501 3XUY7 3XVY2 602 044 602 048 56 USGS Grainger Home Depot Home Depot Home Depot Home Depot Home Depot Home Depo
28. Center for Urban Environmental Research and Education University of Maryland Baltimore County Preliminary Assessment of Real Time Sensor Deployment in Baltimore Urban Watersheds CUERE Technical Report 2011 001 December 2011 Jason VerHoef Claire Welty Julia Miller Melissa Grese Michael P McGuire Roxanne Sanderson Sujay Kaushal and Andrew J Miller Preliminary Assessment of Real Time Sensor Deployment in Baltimore Urban Watersheds CUERE Technical Report 2011 001 December 2011 Jason VerHoef Claire Welty Julia Miller Melissa Grese Michael P McGuire Roxanne Sanderson Sujay Kaushal and Andrew J Miller University of Maryland Baltimore County Center for Urban Environmental Research and Education 1000 Hilltop Circle Technology Research Center Baltimore Maryland 21250 This material is based upon work supported by the National Science Foundation under Grant No 0854307 C Welty Pl as well as NSF grants EEC 0540832 and DEB 1027188 and NOAA grants NA070AR4170518 and NA100AR4310220 This document is available in pdf format for download from http vvvvvv umbc edu cuere BaltimoreVVTB Any opinions findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the vievvs of the National Science Foundation Please cite this publication as VerHoef J C Welty J Miller M Grese M P McGuire R Sanderson S Kaushal and A I Mil
29. End Program nput Locations 1Sta ID 101 2 Version 000 3 Batt Volt111 4 Abb stage111 5 Abb_units 000 6 000 7 YSI Temp 122 8YSI_SC 111 9 YSI_Stage 100 68 10 000 11Sna_N_mM 101 12 Sna_N_mgL111 13 Sna_avg 5000 14 Sna_dark 000 15 100 16 Sna_secON111 17 Sna_count 112 18 000 19 Pmp_secON 111 20 Pmp_count 112 21 000 22Rain 1 531 23Rain 2 1731 24 Rain Sum 111 25 Rain tot1132 26 Rain tot2132 27 Rain sdi1111 28 Rain sdi2101 29 RnSD Tot1111 30 RnSD Tot2111 31 000 32 000 33 000 34 000 35 000 36 000 37 000 38 000 39 000 40 000 41 000 Program Security 0000 0000 0000 Mode 4 Final Storage Area 2 0 CR10X ID 0 CR10X Power Up 3 69 CR10X Compile Setting 3 CR10X RS 232 Setting 1 DLD File Labels 0 Final Storage Labels 0 100 5717 1 Year_RTM 2212 1 Day_RTM 1 Hour_Minute_RTM 1 Seconds RTM 2 Rain 1722 31631 2 Rain 2723 3 200 31535 4 Year RTM 311 4 Day RTM 4 Hour Minute 5 YSI Temp 7 7133 6 YSI_SC 8 2817 7 Sna N mgl 12 9486 8 300 22962 9 Year RTM 26036 9 Day_RTM 9 Hour_Minute_RTM 10 Abb_stage 4 28326 11 Batt_Volt 3 25135 70
30. NA nitrate optical sensors and YSI 600 LS specific conductance and temperature sensors were placed in locked housings at USGS sites where water level data were already being recorded Considerations for system hydraulics power data logging communication systems laboratory testing field site construction and troubleshooting tips are presented Example data sets are shown and recommendations for system improvements are made An itemization of all parts required for deployment is provided viii 1 Introduction 1 1 Objectives The purpose of this technical report is to document the steps undertaken to deploy real time nitrate and conductivity sensors in selected watersheds of the Baltimore Ecosystem Study Long Term Ecological Research project http beslter org The sensor work was funded by National Science Foundation as part of the WATERS Testbed Phase 2 studies A long term goal of this work is to quantify the travel time for solutes to travel through watersheds and predict how this is affected by variability in land use The specific research questions addressed in this project include e How can sources timing and fluxes of solutes from groundwater to surface water vary as a function of land use ultra urban suburban exurban forest and stream position headwater vs downstream e How do transport time scales and subsurface flowpaths vary with flow regime base flow vs storms and antecedent conditions e How can information f
31. actor L is the length of the tubing L D is the diameter of the tubing L V is the average fluid velocity L t Q A Q is the volumetric flow rate IL t A is the cross sectional area of tubing L 4 g is the acceleration due to gravity IL t Frictional head losses are proportional to fluid velocity and tubing length and inversely proportional to tubing diameter The friction factor f is a function of the Reynolds number and can be calculated using the Blasius equation 0 25 where Re is the Reynolds number mes 3 Re and v is the kinematic viscosity of the fluid 2 The head lift and length of tubing required at each of the six Dead Run stations is shown in Table 3 We calculated the frictional head losses for at 0 C and 30 C and at fully charged and low voltage conditions for both pumps Appendix A Table 3 Head lift and tubing length at each station Station Head lift Tubing length required required ft ft DRKR 9 0 21 DR1 9 8 18 DR2 8 0 45 DR3 7 3 19 DR4 7 4 36 DR5 9 2 26 Both pumps are rated by the manufacturer as being able to overcome 32 ft of total head loss elevation head lift head losses due to friction and minor losses at their rated orifice openings The 6 amp pump has an orifice opening of 0 75 in outer diameter the 4 5 amp pump has an orifice opening of 0 5 in outer diameter Since the length of the tubing and the lift at each site are fixed the tub
32. asure YSI 20 Z F x 10 n P30 Set SC to default value of 7 1 7 F 2 00 Exponent of 10 7 again is the flag for 3 8 ZLoc YSI_SC no communication 21 501 12 Recorder P105 1 0 SDI 12 Address 2 0 Start Measurement aM 3 2 Port 4 7 Loc YSI_Temp 5 1 0 Multiplier 6 0 0 Offset 015 the SDI address for the YSI YSI measures 2 parameters starting temperature we only selected these 2 parameters within YSI Ecowatch 22 If X lt gt F P89 if no coms from YSI set Temp to 7 default value 1 7 X Loc Temp 2 4 lt 3 100 F 4 30 Then Do 23 Z F x 10 n P30 1 7 F 2 00 Exponent of 10 3 7 Z Loc YSI Temp 24 End P95 64 OFF Suna and Pump 25 Set Port s P20 Turn Pump and Suna OFF 1 9990 8 5 nc nc nc lovv 2 0999 4 1 lovv nc nc nc CS and C4 are now both changed to low Store Suna and YSI Data to turn off pump 26 Do P86 and SUNA 1 10 Set Output Flag High Flag 0 27 Set Active Storage Area P80 431535 1 1 Final Storage Area 1 2 200 Array ID On 9 27 10 Julia changed time to include year ssas On10 08 10 Jason changed to allow SUNA to be high resolution 28 Real Time P77 4311 1 1110 Year Day Hour Minute midnight 0000 29 Sample P70 7133 1 1 Reps 2 7 Loc YS Temp Temp is stored at lovv resolution default 30 Resolution P78 1 1 High Resolution SC is stored at high resolution but this is not necessa
33. ate values A low value or negative value in the SUNA output during base flow such as 0 003 mg L of nitrate means that there is no nitrate in the sample In our case since we know that our streams all contain significant concentrations of nitrate a low or negative value likely means that there is no water going past the sensor i e the pump is not adequately supplying water to the sensors or there is a blockage in the water line 4 2 8 Testing pumps Many times when a sensor is not collecting accurate data this is due to the pump not 44 supplying water to the sensor To quickly check whether the pump is working properly the positive wire on the pump connected to Port 1 on the relay can be removed and touched to the positive terminal on the battery the relay and battery designated for the pump when the pump is not running Since the system has a common ground once the pump wire touches the positive battery terminal the pump should activate and water should run into the enclosure If this does not happen one of the following may be a occurring 1 there is an issue with the pump due to low batteries 2 the pump is not priming or 3 the PVC pipe is clogged with debris Care should be taken with this procedure because as mentioned in Section 4 1 5 if the positive pump wire attached to Port 1 is disconnected or reconnected while the system is running the relay will short out The relay will no longer act as switch to turn the p
34. atural fluctuation Our second deployment showed a much more smoothly varying nitrate signal which led us to suspect that something was wrong with the Franklintown deployment Investigating the SUNA at DR Franklintown revealed that there was a broken O ring 37 partly blocking the path of the optical sensor The SUNA flow through cell contains two small O rings that fit in circular grooves on the top and bottom of flow cell These O rings provide a tight interference fit but do not fit well and often break when the flow cell is inserted into the SUNA Once we realized that these could break upon flow cell insertion we took greater care to ensure that the O rings were not flopping out of the grooves designed to hold them We found that it takes a great deal of finesse to get the flow cell into the SUNA without breaking these O rings wetting the O rings before inserting the flow cell helps to make insertion easier Even if the flow cell is inserted without the O rings breaking they frequently break when the flow cell is removed Satlantic subsequently provided us with better fitting O rings but precaution with inserting them is still advised Mechanical modifications to the O ring grooves in the flow cells have reportedly improved the design Geoff Maclntyre personal communication 2011 This problem can be identified most easily by inspection of a graph of the streamed sensor data Figure 18 an example from DR3 shows that the diurnal nitrate s
35. connected using a series of small pieces such that the 2 end housing the bilge pump could be screwed on and off Figure 16 for the bilge pump to be easily accessed for maintenance or troubleshooting 32 2in Cap 2in Pipe to Protect Pump 2in z 2in Female 2in Plug ainto 1 with drill holes Threaded Threaded 3 reducer wi rill holes cut from pipe Adapter Adapter 1 in Pipe to Project Tubing Figure 16 Schematic of PVC pipe pieces used to protect pump and pump tubing A 2 male threaded adapter is pressure connected to the 2 PVC piece using PVC cement This can then be connected to the 2 female threaded adapter which needs to be reduced to fit the 1 72 pipe containing the plastic tubing The 2 female adapter and the 2 to 1 72 reducer cannot directly be connected since they have the same inner and outer diameters but they can be connected via a 2 plug The 2 pipe 2 male threaded adapter and a 2 cap to seal the pipe comprise the bilge pump housing assembly which can be screwed on and off The 2 female adapter 2 plug 2 to 1 72 reducer and 1 pipe are all glued together and remain stationary at the site We have experienced that the PVC can break apart at the connection between the 1 pipe and the 2 to 1 coupler connection shown in Figure 16 during flashy storms As an added measure of security we added a self tapping set screw to the 1 pipe and the 2 to
36. d by Julia Miller and Jason VerHoef of UMBC The program is intended for use with some or all of wiring the sensors listed to the left P1 Rain Gage 1 P2 Rain Gage 2 C1 Stage SDI 12 C2 YSI SD 12 C3 Suna SD 12 C4 Suna Control C5 Pump Control This list shows the channels on the CR10x for the various sensors Subroutines 1 subroutine to turn on pump turn on Suna measure Suna measure YSI turn off pump turn off Suna 2 Measure stage There are two main subroutines in the program Data Output ID 300 Julia and Mike switch on 9 27 10 Day of the year time Array ID 300 is assigned to the stage Accububble 200 is for the YSI Battery Voltage and SUNA 100 is for the tipping bucket rain gage Array ID 200 Day of the year stime temperature 5 Nitrate mg L ID 100 Julia and Mike switched on 9 27 10 so all rain gauge data had an Array ID of 100 year day stime hhmm stime sec Rain gage 1 tips Rain gage 2 tips 59 Table 1 Program 01 2 Execution Interval seconds On 9 29 10 changed execution interval was 30 seconds change to 2 seconds since this is what the current rain gauge program has If flag 1 is low reset everything to defaults 01 is the execution interval This will set default values if there is a temp power command 2 seconds is the loss recording interval for 1 If Flag Port P91 precipitation
37. e ISUS and came on the market for 10 000 less than the ISUS just as this project was in its formative stages One potential concern with the Satlantic optical sensor technology is possible interference from turbidity as will be discussed later in this report Sensors for robust specific conductance measurement have been on the market for many years we chose the YSI 600LS http www ysi com for conductivity temperature measurement based on our long term experience with YSI as a vendor of dependable equipment and on USGS recommendations The ability to measure water level with this particular sensor makes it expensive this feature was chosen so as to be able to use the sensors for future projects where water level measurements are required at sites where such data is not already being collected Neither sensor has internal data logging capability we made this choice owing to our intended external data logging system for these and other co deployed sensors Both the SUNA and YSI 600 LS sensors are designed to be submerged in natural waters However owing to our location in an urban area where both vandalism and damage by flash flood storm debris can result in loss of sensors deployed directly in streams we chose to secure the sensors in locked shelters in the floodplain adjacent to each stream location and to pump water to flow cells connected to the sensors This choice made the logistics more complicated than would be required for direct senso
38. e flow cell Lab testing showed that one minute of pumping was adequate to purge the YSI flow cell at the beginning of a subsequent run When the pump is activated it only takes a few seconds for water the reach the sensors from the stream so this factor is negligible in the determination of the system timing Since the YSI flow cell must be purged the system begins by running the pump for 50 seconds After this purging time an excitation signal is sent to the YSI and SUNA to take readings Since the SUNA takes longer than the YSI to obtain a reading the excitation channels remain on for both sensors until the datalogger obtains a data value from the SUNA Once this occurs the excitation channels for the sensors are turned off the YSI goes back to standby mode and the SUNA shuts off The pump is also shut down at this time The SUNA will always attempt to take readings when powered on If the SUNA cannot obtain an accurate reading for any reason various error codes will be sent to the datalogger The datalogger interprets this error code as a data value and will shut down the system The series and order of events required for system operation are summarized in Figure 4 11 On the hour or half hour on pump s Turn pump off Turn on SUNA SUNA Take YSI Take measurement neasurement Figure 4 Sequence of events required for taking sensor measurements and storing the results 12 2 3 Power 2 3 1 Battery sizin
39. e side Figure 15 The sensors are mounted in the interior of the cage using zip ties and bolts the Raven datalogger and SunSaver control panel are mounted onto the plywood facing outward for ease of access The only constraints on the size of the cage are the circumference and height of the drum as well as the need to fit two 12V batteries into the enclosure Figure 14 55 gallon drum that can be used as an enclosure for the sensor package a with lid off b secured 31 Figure 15 Cage constructed to hold sensor package inside 55 gallon drum 4 1 2 Trenching and pipe deployment In order to protect the bilge pump deployed in the stream and the Tygon tubing connecting the pump to the sensors located in the housing on the stream bank the pump and tubing system were enclosed in a length of connected PVC pipe pieces The PVC pipe system was then buried in a channel trenched in the soil so as to conceal and protect the pipe The bilge pump was housed in 2 ID inner diameter PVC pipe Four rows of holes were drilled along an upper half circumference of one end of the 2 PVC pipe segment to allow for water flow into the pipe and contact the bilge pump intake A lower portion of the pipe circumference was left unperforated so as to avoid sucking in sediment when the pump was active and placed against the streambed The 2 pipe containing the pump was attached to 1 PVC pipe containing the Tygon tubing These were
40. ed Water Resources Research 44 W09416 DOI 09410 01029 02007WR006360 VerHoef J R C Welty J Miller M McGuire M Grese S Kaushal A J Miller J M Duncan P M Groffman L E Band and R M Maxwell 2011 Analysis of High Frequency Water Quality Data in the Baltimore Ecosystem Study LTER Abstract H53J 1546 Presented at the 2011 Fall Meeting of the American Geophysical Union San Francisco CA December 5 9 2011 52 Appendix A Calculated head losses H h Az for 4 5 A and 6 A pumps at 0 C and 30 C using Equations 1 3 for flow rates observed at 12 V and 10 8 V Station Tubing Required Length ft Lift Az ft DRKR 21 9 0 DR1 18 9 8 DR2 45 8 0 DR3 19 7 3 DR4 36 7 4 DR5 26 9 2 0 C 6 amp pump D 3 4in D 5 8in D 1 2in 3 8in 12V 10 8 V 12V 10 8 V 12V 10 8 V 12V 10 8 V Station h ft H ft H ft h ft H ft h ft H ft h ft H ft h ft H ft h ft H ft h ft H ft DRKR 1 5 11 0 76 9 8 3 7 13 1 8 11 11 20 5 2 14 41 50 20 29 DR1 1 3 11 0 65 10 3 1 13 1 5 11 9 0 19 4 4 14 35 45 17 27 DR2 3 3 11 1 6 9 6 7 8 16 3 9 12 23 31 11 19 89 97 44 52 DR3 1 4 8 7 0 68 8 0 3 3 11 1 6 8 9 9 5 17 4 7 12 37 45 18 26 DR4 2 6 10 1 3 8 7 6 3 14 3 1 10 18 25 8 9 16 71 78 35 42 DR5 1 9 11 0 94 10 4 5 14 2 2 11 13 22 6 4 16 51 60 25 34 30 C 6 amp pump D 3 4in 5 8
41. ed to leaving the pump running continuously Running the pump continuously is certainly feasible and would need to be taken into account in calculating the power budget The YSI is always on it has standby and data acquisition modes When the datalogger sends a signal to excite the YSI it switches from standby to data acquisition mode and instantly takes a reading of the parameters chosen during setup of the sensors for our 10 case specific conductance and temperature Once the reading is taken it is sent to the datalogger and stored Multiple readings can be taken if desired this is specified through programming of the datalogger The SUNA is either on or off there is no standby When the datalogger sends a signal to excite the SUNA to turn it on the SUNA takes one measurement every 2 seconds The measurements are averaged over the total measurement interval and that value is sent to the datalogger No warm up is needed unless the SUNA will be running for minutes at atime Running the SUNA for 15 seconds to obtain about 6 light samples is sufficient the first three seconds are for taking a dark frame sample according to Satlantic Geoff Maclntyre personal communication June 2011 When the pump is turned off water will drain out of the flow cells and tubing system by gravity Owing to the design of the YSI flow cell some water will remain in the bottom of the flow cell between runs the lower port is located above the bottom of th
42. es we show Table 5 calculation of how long one battery will last for the 6 amp pump 6 7 days For cases where the batteries are charged in the lab for swap out this requires the additional expense of two extra batteries plus two battery chargers see Appendix B This cost plus the time labor required for changing out batteries should be considered when weighing against the option for solar panel purchase and installation for battery 13 recharge Table 5 Calculation of battery replacement schedule The required battery capacity is calculated as follows Since the battery should not be discharged battery past 80 full charge capacity the capacity is scaled by 80 80 is a convention used by the battery industry to account for tolerance errors of the connected devices power ratings C A T 0 8 where required battery capacity amp hrs A amps T time in use hours In our case we wanted to turn the 6 amp pump on for 60 sec 48 times per day 6 amps 60 sec 1 hr 3600 sec 48 measurements day 0 8 6 amps 0 80 hr day 0 8 6 0 amp hours day A 40 amp hour battery will therefore last for 40 amp hrs 6 amp hrs day 6 7 days before needing replacement When fully charged battery voltage will be greater than or equal to 12 6V If the battery charge is allowed to fall below 12V it must be recharged to a minimum voltage of 12 6V Battery charge should never be allowed to fall below 11V Falling bel
43. esign installation and managing of the peripherals to obtain the data in an unattended manner The deployment required background or acquired knowledge in fluid mechanics electronics computer programming data communication systems information management aquatic ecology and biogeochemistry and significant know how with field deployments We could not have completed this effort without a team approach to cover the required knowledge base Although the solar panels and transmitting of data by cellular modem added significantly to the cost of deployment we would not be able to keep the system running otherwise We cannot stress enough how important it is to continually look at the data remotely not looking at the numbers but rather graphing the data to ensure that the system is working as intended The solar panel recharging of the batteries saves on labor and worry of swapping out batteries on a schedule The system hydraulics presented the greatest challenge namely with the pump intake clogging owing to various types of biological matter and trash as well as the limited lifespan of the pumps We stress that groups wanting to try to do sensor deployments must go into this endeavor with a full understanding of the complexities to be encountered and expenses involved We hope that this report can help guide users in such an undertaking 51 References Campbell Scientific Inc 2006 Campbell Scientific Inc Instruction Manua
44. for one week at all six stations when all stations were functional and after the problem with the O rings had been resolved The diurnal signal is apparent at all stations and is most pronounced at DR1 Mean nitrate concentrations at DR1 and DRS two first order watersheds are about double those of the other stations including DR2 which is also a first order watershed of approximately the same size as DR1 and DRS The observed behavior of the nested watersheds reveals interesting behavior that is not just a function of scale but rather is likely also a function of land use and land cover Figure 23 also illustrates the washout of the diurnal signal during storm recovery Previously discussed problems with periphyton clogs of the pump intakes are also shown by Figures 21 22 and 23 49 Dead Run Franklintown a Temperature C Specific Conductance mS cm Nitrate mg L problem with sensor 0 0 0 10 16 0 16 00 16 00 16 00 16 00 16 00 16 00 16 00 16 00 12O0ct 13Oct 140ct 150ct 16 1 17Oct 18Oct 19Oct 200ct Time EST Figure 22 Observations of stream flovv temperature nitrate and specific conductance at DRKR measured every 30 minutes during base flovv and storms October 2010 Nitrate 2 5 2 0 a Nitrate mg L 0 50 Periphyton clogs uu 11 27 10 11 28 10 11 29 10 11 30 10 12 1 10 12 2 10 12 3 10 12 4 10 Date Figure 23 Nested vvatershed behavior of nitrate concentrations
45. g The first requirement for any power budget is to determine what the current draw is of all pieces of equipment These requirements can be determined from manufacturer documentation The current drawn for our sensor components is listed in Table 4 Table 4 Current drawn for sensor station components Device Current drawn amps Rule Pump 4 5or6 VVale Pump 3 5 SUNA 0 63 YSI lt 0 001 CR10X datalogger lt 0 001 quiescent to 0 46 processing Our field logistics did not include the option for AC power supply in all cases DC via 12V batteries was the best option available In order to avoid having to continuously swap out batteries our system was designed to charge the batteries in situ from solar panels Owing to the significant current draw from the bilge pumps 3 5 4 5 or 6 amps depending on the pump chosen we chose to run the pump from one 12V battery and all sensors and datalogger from a second 12V battery with each battery charged by a dedicated solar panel The size of each 12V battery was chosen so as to maximize the amp hour rating while allowing the battery to fit inside the existing enclosures and be portable such that it could be carried by project personnel from a field vehicle to the site over uneven terrain We chose 40 amp hour gel batteries by MK see parts list Appendix B each of which weighs 32 pounds and measures 7 75 L x 6 63 W x 6 88 H Because one site was initially powered exclusively by batteri
46. g to the web interface a graph automatically displays specific conductance nitrate temperature and discharge data for the DRKR DR Franklintown watershed Dropdown boxes to the right of the graph allow one to change the site e g DRS in Figure 6 and the number of days visible on the graph Data from the graph are displayed in a table below the graph At the bottom of the page site information includes site ID location and contact information 2 6 Wiring The wiring of the sensors batteries and solar panels to the datalogger is shown in Figure 7 SunSaver 6 for Sensors and CR10X G 12V Solar Battery Load G Earth Ground Power In CR10X 2 1 6 5 4 3 P1 G P2 G C8 C7 C6 C5 C4 C3 C2 C1 G 12V 12V Solar Panel 3 4 Relay SunSaver 6 for Bilge Pump 2 1 Solar Battery Load ra mi Gi 5 x 1 1 1 RH Bilge Pump Trt HB Panel Data for YSI Povver Ground Excitation Channel Data for SUNA Note the colors of shipped SUNA wires are white for ground and black for power Figure 7 VViring diagram for sensors The solar panels and batteries are connected to the SunSavers vvhich have ports labeled for each One SunSaver can be used for the datalogger and sensors and the other SunSaver for the pump The 12V povver in and the ground on the datalogger connects to the load terminals on the one SunSaver 21 Two relays are used one for the SUNA and one for the bilge pump Port 2 on
47. gher turbidity during storm events Therefore we recommend that the SUNA be cleaned after storms Subsequent to recession of storm flow the 2 inch PVC at some sites becomes covered with sediment While the PVC may still fill with water and the pump may still be supplying water the sediment needs to be removed after storm events Over time sediment can get into the pump and although the pump may be working with the PVC covered in sediment eventually it will be unable to suck in more water In cases like this the SUNA will usually read 0 even though the YSI is working properly We have experienced that the PVC end can become dislodged during storms In one instance the pump was still in the water and data were still being collected so unless a site visit was performed after a storm we would not have known that the PVC was missing 4 3 7 SUNA calibration Over time the lamp in the SUNA may drift slightly Every 4 to 6 months the SUNA should be calibrated which can be accomplished using SUNACom The method is described in the SUNACom section of the SUNA Owner s Manual and is summarized here The flow cell should be removed and the optical path should be cleaned thoroughly with a Q tip and isopropyl alcohol Plastic food wrap should be tightly wrapped around the SUNA near the slit for the optical path The SUNA should be placed on a flat surface with the optical path slit facing up A hole is then poked in the plastic wrap and the slit filled
48. he data for each variable The web service definition language WSDL address for the nitrate database is located at http his09 umbc edu BaltNitrate cuahsi 1 1 asmx VVSDL The web service is called by a SOAP client developed using PHP The client uses PHP SOAP and Simple XML to call the vveb service methods and parse the SOAP response The data is then displayed through a web interface developed using PHP and javascript as shown in Figure 6 19 UMBC Sensor Project DR 5 Choose the site Specific Conductance mS cm Nitrate mg L Temperature C Discharge cfs DRS EEN 307 4 0 View data beginning with 2 00 October 18 2011 and 0 8 i Coa 9 so 25 ending with october 21 2011 Update Graph 160 0 6 5 20 Z 140 2 a 120 ay 3 Le 4 1 00 2710 2 w 2 s 8 00 to 2 9 5 Edo 8 E oso 8 5 G 0 2 o 3 0 20 i 8 0 4 0 00 Date and Time EST SPecific Conductance Date and Time EST Date and Time EST Temperature mS cm C October 18 2011 00 02 00 0 91 October 18 2011 00 02 00 October 18 2011 00 02 00 13 70 October 18 2011 00 32 00 0 91 October 18 2011 00 32 00 October 18 2011 00 32 00 13 62 October 18 2011 01 02 00 0 91 October 18 2011 01 02 00 October 18 2011 01 02 00 13 52 October 18 2011 01 32 00 0 91 October 18 2011 01 32 00 October
49. he need for manual priming manually forcing water into the suction port A self priming bilge pump is appropriate for our application since unattended operation is required 2 2 2 Pump sizing For the water to reach the sensors the pumps must overcome the elevation head and head losses due to friction The elevation head remains constant at a site and is determined by measuring the required lift at a site The frictional head losses are a function of fluid velocity tubing diameter and length and fluid physical properties which depend on temperature Additional losses due to tubing bends are a minor component of overall head loss 2 2 2 1 Effect of battery voltage on flow rate Since most pumps are overefficient or underefficient with respect to their rated flows it is prudent to test a pump before deployment and observe its actual flow rate this flow rate is used in calculating the friction losses and for determining whether the pump can overcome those losses 4 5 and 6 amp Rule bilge pumps were chosen for testing The pumps are designed to be powered by 12 V batteries but rated 12 V batteries can have a voltage potential up to 13 7 V when fully charged 10 8 V is the recommended lower limit for a 12 V battery before being recharged Flow rates at voltages ranging from 13 7 V to 12 V and at 10 8 V were therefore measured in the lab Tests were performed using tap water at 19 5 C The relationship between input power and measured fl
50. i e it is open to the atmosphere The flow cell fills with water through an inflow port near the bottom the outflow port is near the top of the flow cell If the flow rate exceeds the capacity of outflow port the water will leak out of the top of the flow cell YSI states that the maximum flow rate that the flow cell can accommodate is 5 gallons per minute Since the flow line is split in half before reaching the YSI this means that a pump flow rate greater than 10 gal min should not be used 2 2 2 4 Pump service life Bilge pumps have a finite life expectancy Over time the internal impeller becomes worn This causes the system to cavitate and allows bubbles in the water line which can lead to errors with the SUNA nitrate measurement In addition the brushes on the internal motor become worn down more quickly as the frequency of cycling the pumps on and off increases The Rule pump has a plastic impeller The Rule pump manufacturer specifications state the pump has an expected service life of approximately 100 hours or about 60 days for our application i e with the pump running for roughly 2 minutes during every half hour sampling interval However we found that the Rule pumps were reliable for only about 40 days in most cases but did last up to 60 days at some sites Indications that the pumps are beginning to break down are that 1 they fail to self prime and 2 they pump with air bubbles in the water line Wale makes a 3
51. ies This also prevented larger sediment particles from entering the pump housing chamber We chose woven copper mesh cloth as the netting material because in addition to working as a sediment filter the copper acts as a biocide to deter periphyton growth Periphyton is a mixture of algae bacteria and detritus and appears as brown specks clinging to the pipe The periphyton growth can be fast and thick and actually can impede water flow into the perforated holes if the employed netting does not have biocidal properties Copper mesh cloth is flexible and can be easily wrapped around the end of the 2 PVC that protects the pump housing We have found that a fine mesh of 100 x 100 0 0045 wide diameter and 0 006 width opening has been effective in keeping sediment out of the PVC and does not restrict water flow As a precaution to protect the pumps from 41 sediment or other small debris we also wrapped copper mesh around the pump intake The copper mesh needs to be replaced approximately every 6 months as fine holes can be punched into it from passing storm debris Even after placement of netting over the perforated holes we still had problems with decaying leaves and trash mainly plastic bags getting sucked tightly against the intake holes by the pump causing the water intake to drop and the sensors to fail to record a reading We were able to rectify this problem by using a combination of two methods 1 building rock barriers in
52. ignal is difficult to discern when interference is present but clearly observable when the interference has been removed This problem can be confirmed by evaluating the SUNA spectra using the SUNACom software with the sensor connected to a PC Figure 19 If the spectra output show noise in the signal as in Figure 19a this is an indication that there is debris in the optical path that is not being flushed out during runs one likely cause of this behavior is a broken O ring If the spectra output is good the data should plot as a smoothly varying curve as a function of time as shown in Figure 19b Figure 19a shows that when there is a blockage of the optical path the light counts are on the order of 1000 to 2000 When light is passing through the water sample there should be a peak near or over 40 000 light counts as shown in Figure 19b Anything below 10 000 counts means that adequate light is not passing through the water This could also be caused by high turbidity 4 2 2 Pump under over sizing Pumps are purchased with a specified nominal flow rate e g 8 gal min As discussed in Section 2 2 2 2 and shown in Table 2 the observed flow rate is less than specified and can vary with battery charge Also the rated head loss that can be overcome with the pump will depart from manufacturer s specifications if a tubing size smaller than the pump outlet orifice is used As shown in Appendix A and Figure 3 swings in temperature will affect vi
53. in 1 2in 3 8in 12V 10 8 V 12V 10 8 V 12V 10 8 V 12V 10 8 V Station h ft H ft h ft H ft h ft H ft h ft H ft h ft H ft H ft h ft H ft h ft H ft DRKR 1 3 10 0 62 9 6 3 0 12 1 5 10 8 6 18 4 2 13 34 43 17 26 DR1 1 1 11 0 53 10 2 6 12 1 3 11 7 4 17 3 6 13 29 39 14 24 DR2 2 7 11 1 3 9 3 6 4 14 3 1 11 18 26 9 1 17 72 80 36 44 DR3 1 1 8 4 0 56 7 9 2 7 10 1 3 8 6 7 8 15 3 8 11 31 38 15 22 DR4 2 2 9 6 1 1 8 5 5 1 13 2 5 10 15 22 7 3 15 58 65 29 36 DR5 1 6 11 0 77 10 3 7 13 1 8 11 11 20 5 3 14 42 51 21 30 53 0 C 4 5 amp pump D 1 2in 3 8 in 12V 10 8V 12V 10 8 V Station h ft H ft h ft H ft h ft H ft h ft H ft DRKR 5 5 15 3 1 12 22 31 12 21 DR1 4 7 15 2 7 12 19 28 11 20 DR2 12 20 6 7 15 46 54 26 34 DR3 5 0 12 2 8 10 20 27 11 18 DR4 9 5 17 5 4 13 37 45 21 28 DR5 6 8 16 3 9 13 27 36 15 24 30 C 4 5 amp pump D 1 2in 3 8 in 12V 10 8 V 12V 10 8 V Station h ft H ft h ft H ft h ft H ft h ft H ft DRKR 4 5 14 2 6 12 18 27 10 19 DR1 3 9 14 2 2 12 15 25 9 18 DR2 9 7 18 5 5 13 38 46 22 30 DR3 4 1 11 2 3 9 6 16 23 9 16 DR4 7 7 15 4 4 12 30 38 17 25 DR5 5 6 15 3 2 12 22 31 12 22 54 Appendix B Parts list Item Sensors SUNA SUNA flow cell SUNA power telemetry cable 8 pin pigtail SUNA carry
54. ing this action can short out the relay and leave it stuck on whatever is connected to the relay will remain on An 8 gage wire was used for the earth ground The wire was run to the copper grounding rod that was set near the enclosures 4 1 6 Grounding rod One half inch diameter 8 ft copper rod can be used as a ground The rod can be hammered into the ground using a post driver until about 2 ft of rod remains exposed followed by a sledge hammer to finish the job Approximately 2 inches of rod should be left exposed The 8 gage ground wire is attached the near the top of the exposed rod with a copper grounding rod clamp It is important to put the clamp on the rod before it is set in the ground because the tip of the malleable rod will be rendered flatter and slightly larger in diameter from hits by the post driver and sledge hammer rendering it infeasible to slip on the tightly fitting clamp afterwards The rod should be installed near the enclosure and can be covered with small rocks for camouflage 4 2 Troubleshooting sensors After our sensor stations were deployed we ran into a number of problems that took a bit of time to identify and remedy We hope that by providing some lessons learned in what can go wrong others can more quickly resolve the same problems 4 2 1 Blockage of SUNA optical path At our first deployment DR Franklintown we recorded a highly fluctuating nitrate signal We had assumed that this was a n
55. ing case YSI 600LS BCR C T SV YSI 600 flow cell Vented cable 25 ft Adapter DB 9 MS 8 Power Adapter Power Flying Lead Adapter Field Flying Lead Conductivity calibration solution Power 30 Watt lightweight unbreakable solar panel 10 Watt lightweight unbreakable solar panel Mounting bracket for 30 Watt panel Mounting bracket for 10 Watt panel Solar charge controller 12V 6 amp 12 V 40 amp hr gel battery Battery charger Communication Raven XT Raven antenna Dual band cellular PCS antenna magnetic mount Mounting bracket for Raven XT Monthly phone service Cable for modem connection CS I O to 9 pin RS 232 DCE interface Brand and Model Satlantic Satlantic Satlantic Satlantic YSI YSI YSI YSI YSI YSI YSI SunSaver MK Professional series 12 V 10 A 3 stage charger GPRS Modem G2263 CD Sierra Wireless part 100 170 1013 Campbell Scientific Part number 600 12 696 6191 6095 6100 6096 60907 BSP3012 LSS BSP1012 LSS 30LTHPM 10LTHPM SS6 12V M40SLDG INTPS121 0 364188 AP85 18 M S2 324142 SC932A 55 Vendor Satlantic Satlantic Satlantic Satlantic YSI YSI YSI YSI YSI YSI YSI batterystuff com batterystuff com batterystuff com PovverUp batterystuff com batterystuff com batterystuff com Tessco AntennaPlus Tessco AT amp T Campbell Scientific Cost 18 000 00 350 00 225 00
56. ing diameter becomes the main design variable affecting calculation of total head to be overcome through friction loss This is illustrated for DR2 in Figure 3 100 90 80 70 60 50 40 30 Head Loss Due to Friction ft 035 04 045 O05 055 06 O65 OF O75 Tubing Diameter in Figure 3 Effect of tubing diameter and temperature on friction loss at DR2 for the 6 amp Rule pump We initially chose 3 8 in ID tubing for the 4 5 amp pump because this is the tubing size that fits the intake ports of the sensor flow cells However we subsequently increased the tubing diameter to 0 5 in ID in order to reduce the head loss due to friction using this pump The tubing can be reduced to 3 8 in ID tubing using barb fittings close to the flow cell intake ports The calculations in Appendix A and the DR2 example shown in Figure 3 also illustrate the effect of fluid viscosity due to changes in temperature on head loss the effect becomes apparent for smaller tubing This range of temperatures 0 C 30 C is expected over the range of seasons when the sensors will be deployed 2 2 2 3 Considerations of maximum flow rate Although a powerful enough pump must be used to overcome head lift and friction losses there is a limit on the maximum flow rate that can be used in our system owing to the design of the YSI flow cell The YSI flow cell is not directly attached to the sensor The sensor sits inside the flow cell which is not sealed
57. ission cable SUNA Pig tail cable e Purple data C2 on CR10X e Green data C3 on CR10X e Red power 12V on CR10X Black power 1 on SUNA relay e Black ground G on CR10X e White ground G on CR10X RavenXT wireless modem Red wire 12V on CR10X e Black wire G on CR10X Relay for SUNA and CR10X Relay for bilge pump 1 SUNA pig tail black power 1 Positive on bilge pump 2 12V Power In on CR10X 2 6 load positive on SunSaver6 3 C4on CR10X e 3 C5 on CR10X 4 CR10X 4 G on CR10X Battery for sensors and CR10X e Positive 2 battery positive on SunSaver 6 for sensors e Negative 1 battery negative on SunSaver 6 for sensors Battery for bilge pump e Positive 2 battery positive on SunSaver 6 for bilge pump e Negative 1 battery negative on SunSaver 6 for bilge pump Solar Panel for sensors and CR10X e Positive 4 solar panel positive on SunSaver 6 for sensors e Negative 3 solar panel negative on SunSaver 6 for sensors Solar panel for bilge pump e Positive 4 solar panel positive on SunSaver 6 for bilge pump e Negative 5 solar panel negative on SunSaver 6 for bilge pump 23 SunSaver 6 for sensors and CR10X e 1 Battery for sensors and CR10X negative e 2 Battery for sensors and CR10X positive 3 Solar panel for sensors and CR10X negative 4 Solar panel for Sensors and CR10X positive e 5 G Power In on CR1
58. l RavenXTG Sierra Wireless Cellular Modem Groffman P M N L Law K T Belt L E Band and G T Fisher 2004 Nitrogen fluxes and retention in urban watershed ecosystems Ecosystems 7 393 403 Kaushal S S P M Groffman L E Band C A Shields R P Morgan M A Palmer K T Belt C M Swan S E G Findlay and G T Fisher 2008a Interaction between urbanization and climate variability amplifies watershed nitrate export in Maryland Environmental Science amp Technology 42 5872 5878 Pellerin BA BD Downing C Kendall RA Dahlgren TEC Kraus J F Saranceno RGM Spencer and BA Bergamaschi 2009 Assessing the sources and magnitude of diurnal nitrate variability in the San Joaquin River California with an in situ optical nitrate sensor and dual nitrate isotopes Freshwater Biology 54 376 387 doi 10 1111 j 1365 2427 2008 02111 x Pickett S T A M L Cadenasso J M Grove P M Groffman L E Band C G Boone W R Burch C S B Grimmond J Hom J C Jenkins N L Law C H Nilon R V Pouyat K Szlavecz P S Warren and M A Wilson 2008 Beyond urban legends An emerging framework of urban ecology as illustrated by the Baltimore Ecosystem Study BioScience 58 139 150 Shields C A L E Band N Law P M Groffman S S Kaushal K Savvas G T Fisher and K T Belt 2008 Streamflow distribution of non point source nitrogen export from urban rural catchments in the Chesapeake Bay watersh
59. ler December 2011 Preliminary Assessment of Real Time Sensor Deployment in Baltimore Urban VVatersheds CUERE Technical Report 2011 001 UMBC Center for Urban Environmental Research and Education Baltimore MD This report supercedes a previous version dated February 2011 ON THE COVER Sensor station at DR4 Table of Contents Abstract 1 Introduction 1 1 Objectives 1 2 Choice of sensors 1 3 Site description 1 4 Timeframe of initial testing 2 Design 2 1 Overview 2 2 Hydraulics 2 2 1 Pump type 2 2 2 Pump sizing 2 2 2 1 Effect of battery voltage on flow rate 2 2 2 2 Calculation of friction losses 2 2 2 3 Considerations of maximum flow rate 2 2 2 4 Pump service life 2 2 3 Y configuration and location of sensors 2 2 4 Filtering 2 2 5 Determination of on off time for pump and sensors 2 3 Power 2 3 1 Battery sizing 2 3 2 Solar panel sizing 2 3 3 Use of relays 2 4 Data logging 2 5 Communication 2 5 1 Raven set up and programming 2 5 2 Base station communication with datalogger 2 5 3 Data archiving and serving 2 6 Wiring 2 7 Parts chart costs and itemization 3 Laboratory testing 3 1 Chemical QA QC 3 2 Testing hydraulics wiring datalogger programming 3 2 1 Sensors 3 2 2 Hydraulics 3 2 3 Wiring and datalogger 4 Field deployment 4 1 Construction Page viii O UT UT U1 UT A a NNP xx 1 00 BUY SUN 24 24 29 29 29 30 30
60. m voltage 12 V Backup days of power required 1 Battery amp rating 40 amp hours Sunlight hours per day 8 Select panel size 30 watts Output Number of panels needed calculated 1 2 3 3 Use of relays Relays are essentially switches used to connect the datalogger to equipment when on and off modes are desired We used relays for the pump and the SUNA otherwise they would be in the continuous on mode once powered up This enabled us to 1 conserve power 2 keep the batteries and solar panels needed to reasonable sizes and 3 conserve the SUNA lamp which has a finite specified life Various types of relays are available for this application solid state relays were chosen Solid state relays are ideal for our application because they allow a low input signal 15 from the datalogger to activate a much higher voltage potential used for the bilge pump and sensors Most solid state relays work in the same way using four ports connected to the input signal device datalogger and device to be used and the device s power supply For our relays Port 1 connects to the power input of the device to be turned on and off the bilge pump and SUNA Port 2 connects to the power supply for that device which is the battery positive or the load positive on the SunSaver if a solar panel is used Port 3 is the input excitation port this is connected to the datalogger or any device that triggers the system to turn on
61. measured every 30 minutes under base flovv conditions and during a storm 50 Ongoing work is being undertaken to tie observations into a physically based mathematical model to incorporate other spatial data vegetation impervious cover slopes and to deploy additional sensors e g oxygen at the sites in an effort to elucidate function and stream metabolism in these nested urban watersheds We will also be examining system behavior as a function of season and capturing the finer temporal scale behavior of storms as we change the frequency of data collection 6 Summary and recommendations We deployed Satlantic SUNA nitrate and YSI conductivity temperature sensors in a nested watershed design for an initial test period October December 2010 and March December 2011 in Baltimore MD to obtain high frequency stream water quality data Because our stations are located in an urban area we chose to secure the sensors in locked boxes out of the stream and to pump water to them as opposed to placing the sensors directly in the stream The sensors worked well in terms of continuous function and we were pleased with the service provided by the vendors to help us in troubleshooting Addressing the issue of the need to carry out in line filtering to remove turbidity to obtain robust nitrate measurements comparable to laboratory IC readings is an ongoing effort that we hope to resolve in the future The biggest learning curve for our group was d
62. n light is unable to get through the optical path of the water sample and reach the detector the second generation black casing SUNA generates an alpha string of NAN which the CR10X datalogger records as a zero integer The reasons that this error could occur include 1 the water may be excessively turbid 2 an air bubble could be stuck in the flow cell or 3 or the SUNA lamp may not be not working The zero recorded by the datalogger does not mean a zero value of nitrate in fact the SUNA cannot produce a zero integer value for a nitrate measurement CR10X recording of 6999 for SUNA nitrate value For the first generation SUNA stainless steel casing this error as discussed above is represented as repeating digits of 9 999999 The CR10X datalogger can be set to high or low resolution This refers to the highest value the datalogger can store In low resolution the datalogger can store a value up to 6999 If a value of 6999 or larger is sent to the datalogger and it is in low resolution it will record 6999 A 99999 string is larger than 6999 and therefore it will be recorded as 6999 A value of 6999 will also be recorded if more than 4 digits are sent as a data string while the datalogger is in low resolution CR10X recording of 99999 for SUNA nitrate value If the CR10X datalogger is set to high resolution and this error occurs SUNA sending a 99999 a value of 99999 will be recorded Low magnitude or negative nitr
63. nal influence of the biotic portion of specific conductance on the recorded signal Figure 22 shows observations recorded by the sensors during base flow and storms during our initial deployment at Franklintown DRKR The nitrate signal is erratic because these observations were recorded before we discovered and fixed the problem with the broken O rings blocking the path of the optical sensor Diurnal variation in 48 stream temperature is apparent as are slight increases in temperature during initial storm runoff even as seasonal daily mean temperatures have started to drop as fall season progresses This plot also demonstrates the classic behavior of base flow constituent concentrations being diluted from storm runoff A two end member mixing model can be applied to the specific conductance data to calculate percent old water of the total stormflow For these data sets such a calculation yielded old water as being 21 and 46 of the total storm flow for the two storms respectively Alternative conceptual models for mixing can be applied to the data as desired DR3 2 0 30 25 1 5 20 Nitrate mg L Specific Conductance mS cm Temperature C 0 0 0 11 17 10 11 18 10 11 19 10 11 20 10 11 21 10 11 22 10 11 23 10 11 24 10 11 25 10 Date Figure 21 Nitrate specific conductance and temperature data recorded every 30 minutes at DR3 Figure 23 illustrates a number of features of the nitrate signal behavior
64. not reach the flow cell A pump that appears to be working well during warm weather may fail if the temperature drops significantly and smaller tubing is being used because the head loss due to friction increases significantly On the other hand a pump must not be oversized owing to the potential for leaks out of the top of the YSI flow cell as discussed in Section 2 2 2 If water is not reaching the flow cells due to problems with either pumping water in or draining water out the flow cells will be attempting to record an air measurement The values being recorded will drop suddenly The SUNA will show a low positive or negative value order 0 001 mg L the YSI will show a low positive value order 0 001 mS cm 4 2 3 Tubing kinks Kinking of the tubing can cause pressure buildup in the lines This pressure can force water out the top of the YSI flow cell or stress the tubing at the inflow or outflow ports on the flow cell To guard against the latter problem all tubing should be secured to ports with hose clamps An additional consideration of the tubing set up is that the water must be free to drain away from the flow cells Therefore the exit end of the plastic tubing must be kept free of debris soil 4 2 4 Leaf trash and periphyton clogs Based on our experience with leaf debris and dead crayfish clogging the inch open holes in the perforated pipe we covered the pipe holes with netting material held in place using wire t
65. nstalling SIM cards programming the modems establishing a connection performing data download and troubleshooting can be obtained from the Campbell Scientific Inc Instruction Manual RavenXTG Sierra Wireless Cellular Modem Campbell Scientific Inc 2006 Key steps are summarized here Once SIM cards have been purchased installed in the modem and activated the modems can be programmed using AceManager software AceManager can be downloaded from the Sierra Wireless Airlink Solutions website http www sierrawireless com support A template file Raven GPRS EDGE Template 9600 which configures the modem to be compatible with the Campbell Scientific CR10X datalogger can be downloaded from the Campbell Scientific website http www campbellsci com downloads Programming the wireless modem requires that it be connected to a PC using a RS 232 cable The modem must be powered with a 12 V battery and connected to a cellular antenna When everything is set up correctly and the modem is being powered AceManager can be opened and the modem can be accessed by clicking on connect and selecting PPP and the COM port that the modem is connected to After a connection has been established the template file can be loaded and write must be clicked on to upload the file to the modem When the template has been loaded the Device Port value under the Misc group should be 3001 3001 becomes the assigned port number as referred in the next
66. o organize the wires 16 gage wire was used for the miscellaneous wiring needed for the system Spade terminals were also used to connect wires to the relays and SunSavers and terminal An earth ground must be connected to the datalogger which is the labeled slot on the bottom right for the CR10X Typically 8 to 10 gage wire is used The wire is run toa copper grounding rod hammered into the soil near the instrument enclosure The wire is connected to the grounding rod with a copper grounding clamp The ground wires of the various components can be connected to any channel labeled G on the CR10X such that they all connect to the common ground However the Load ground Port 5 on the SunSaver used for the CR10X should be connected to the ground channel associated with adjacent to the 12V power in on the top right of the datalogger A summary of the wiring connections is provided in Table 7 22 Table 7 Summary of wiring connections for CR10X and sensor set up CR10X e 12V Power In 6 load positive on SunSaver6 for sensors and CR10X 12V Power In 2 on relay for sensors and CR10X Gon Power n 5 load negative on SunSaver6 for sensors and CR10X e 2 Purple wire on YSI data transmission cable e C3 Green wire on SUNA pig tail cable e C4 23 on relay for sensors and CR10X e C5 3 on relay for bilge pump 12V Red wire for RavenXT wireless modem e 12V Red wire on YSI data transmission cable YSI Data transm
67. ober 6 December 6 7 2010 The sensors were brought inside on December 6 7 due to concerns for pumping freezing water through the flow cells and potentially damaging them The sensors were re deployed in March 2011 and tested through December 2011 so as to be able to evaluate data over a period of four seasons 2 Design 2 1 Overview The design of our system is dictated in part by physical constraints of the field sites as well as the need to utilize components already deployed at chosen sites At five of our six sites DR1 DRS Figure 1 painted steel junction boxes 30 x 30 x 12 have been deployed by USGS under a joint project with UMBC for the purpose of housing Accububble pressure transducers and data logging communication equipment we had access to these boxes for housing additional sensors At one site DR5 two raingages are also part of the station At DR Franklintown a similar setup was not available to us and therefore we deployed the nitrate and conductivity sensors in a separate enclosure a 55 gallon drum located at this site for other BES experiments At all sites we chose to install a bilge pump in the stream and pump water to the sensors This configuration required sizing the pump and tubing to overcome head losses trenching PVC pipe to enclose pump tubing and wiring between the stream and the sensor enclosure and securing protecting the pump in the stream These steps are not needed for sites where the sens
68. of the BES main nested watersheds plus the forested reference watershed in year 2 of the project Figure 1 Table 1 USGS streamflow gaging stations are located at all of these sites This report addresses initial deployment in Dead Run and its subwatersheds A Riparian Zone Piezometers A Bedrock Wells Stream Flow Gages A Biogeochemical Study Plots e Stream Gage Baisman Run DeadiRun at 4 Franklintown 2 b 22 Ray X k JN Sa f Figure 1 Study watersheds for sensor project a Dead Run Falls 171 km and nested subwatersheds i 0 e 1 i 1 x r Table 1 Characteristics of study vvatersheds Name Area Percent Nested km Impervious Within Dead Run Sites DR Franklintown DRKR 14 1 45 0 DR3 5 0 48 2 DRKR DR4 6 2 49 8 DRKR DR1 1 3 51 1 DR3 DR2 1 9 44 7 DR3 DR5 1 53 44 9 DR4 Gwynns Falls Longitudinal Sites Gwynns Falls at Carroll Park 171 30 3 Villa Nova 84 5 21 1 Carroll Park Delight 10 6 18 6 Villa Nova Glyndon 0 7 21 1 Delight Pond Branch 0 4 0 Baisman Run Forested reference watershed 1 4 Timeframe of initial testing Laboratory QA QC chemical testing was carried out in spring of 2010 laboratory testing of wiring hydraulics and datalogger set up was carried out in summer of 2010 Initial field deployments of 6 sensor stations were carried out in Dead Run from Oct
69. ogram is needed to turn on the pump and the SUNA sensor and record sensor measurements of nitrate specific conductance and temperature Since we have Accububble pressure transducers co located at most stations plus two precipitation gages at one station our program has options for recording these data as well 16 The datalogger program written for and utilized in this project is given in Appendix C and annotated to aid the reader in interpretation The user manual providing the syntax for CR10X programming can be downloaded from http www campbellsci com documents manuals cr10x pdf 2 5 Communication 2 5 1 Raven set up and programming Wireless cellular modems RavenXTG Sierra Wireless Cellular Modems were chosen as the telemetry for the sensor system In order to remotely access data via the wireless modem a SIM card with an assigned phone number and an assigned static IP address was purchased installed in the wireless modem and activated After speaking with an account representative from AT amp T we determined that 3072 KB of data usage would be a sufficient amount for transmitting the sensor data This estimate includes transmission of stage data at these sites where stage data are collected every 5 minutes SIM cards can be purchased and delivered without being activated Activation at a later date such as the day before deployment is recommended to avoid paying for the service before it is being used Detailed instructions for i
70. ors can be installed directly in the stream An additional consideration in many of the design components is the frequency at which measurements are to be made We did not want to consume energy by taking measurements too frequently during base flow when changes vary slowly over time however a main objective was to capture the behavior of storms where measurements every 5 minutes are typical for urban systems For the purpose of initial testing we set the sensors to record measurements every 30 minutes A feature of our system is that the dataloggers can be called via cellular modem to change this to a shorter period as desired This option to change the recording interval on the fly will be implemented in future deployments alternatively this feature can be implemented by triggering sensor measurements with water level changes in stations where the sensors are co deployed with pressure transducers Discussed in this section are considerations for system hydraulics pump sizing filtering power data logging communication wiring and data serving that were utilized in our 4 sensor deployments 2 2 Hydraulics 2 2 1 Pump type Bilge submersible pumps are commonly used in boating and are readily available inexpensive compact and robust They are designed to be cycled on and off often and to last without corrosion in sea water Most bilge pumps are self priming if they are placed in water and switched on they will pump water without t
71. osure Once the desired path is determined trenching can begin It is important to trench before cutting or connecting the piping because the path may turn out to be obstructed by large roots or concrete foundations After the trenching is completed the pipe segments can be fit together The first step is to attach the adapters and reducer as shown in Figure 16 to build the pump housing Then the bilge pump can be connected to the plastic tubing secured with a hose clamp and placed inside of the PVC housing The pipe can then be assembled starting from the 2 to 1 reducer and ending at the sensor enclosure The plastic tubing and pump wires can be fed through each piece of PVC as work progresses All PVC connections should be cleaned with PVC primer and glued with PVC cement A hacksaw can be used to cut all pieces to desired lengths and elbows can be used to conform to the topography as needed The pumps do have a finite lifetime so one connection in the PVC should remain unglued to access the wires It is helpful to have a splice in the wires near this unglued connection for ease of changing the pumps As shown in Figure 17 a cap can be fit over the end of the PVC bilge pump housing This 34 cap should not be glued so that it can be easily removed A fine copper mesh can be wrapped around the perforated 2 PVC to act as a biocide and filter A cage can be built in front of the PVC intake as a barrier against leaves and other
72. ow rate for the two pumps evaluated is shown in Figure 2 This figure illustrates that as the voltage drops from around 12 5 V to 10 8 V the flow rate decreases by approximately one third The slope of the fitted line for the 4 5 amp pump is half that of the 6 amp pump this shows that the flow rate of the 4 5 amp pump is less affected by a drop in voltage than the 6 amp pump These ranges in flow rates need to be considered when deciding on which pump to use The lower flow rate of the chosen pump should be used to conservatively determine the purge time needed between runs as discussed in Section 2 2 4 Table 2 lists the average flow rate observed for fully charged and low voltage conditions 5 0 4 5 4 0 3 5 sd y 0 069x 1 2 Flow Rate gal min R 0 98 HAZA 5 Amp Pum 2 5 o 20 6 1 5 1 0 45 50 55 60 65 70 75 80 Power Watts Figure 2 Pump flow rate dependence on battery voltage power watts voltage volts current amps Table 2 Observed average pump flow rates with tap water at 19 5 C Fully Charged Low Voltage Pum Manufacturer s Battery Battery P Specifications 13 7 V 12V 10 8 V 6 amp 8 gal min 4 2 gal min 2 8 gal min 4 5 amp 4 gal min 2 9 gal min 2 1 gal min 2 2 2 2 Calculation of friction losses The Darcy Weisbach equation can be used to determine frictional head losses LV 22 where h is the head losses due to friction L fis an empirical friction f
73. ow this can adversely affect the battery s ability to subsequently charge 2 3 2 Solar panel sizing Solar panels are available for recharging 12V batteries There are a number of factors involved in sizing a solar panel for this purpose including wattage required by the equipment number of hours of daylight available battery size and rate of battery drawdown A useful tutorial and an interactive calculator for determining number and sizes of solar panels required are available at batterystuff com 14 http www batterystuff com solar calculator html http www batterystuff com tutorial solar calculator html This software allows calculation of numbers and sizes of solar panels for a given application For the example of our 6 amp pump and 12V 40 amp hour battery the batterystuff com materials were used to determine that one 30 watt solar panel would be adequate for running the pump for 1 minute 48 times per day The details used in the solar calculator are provided in Table 6 This choice allows for our anticipated increase in sampling frequency during storms which would increase the number of hours per day of pump usage For our other sensors powered by the second 12V battery we determined that a 10 watt panel was adequate Table 6 Sizing of solar panel using http www batterystuff com solar calculator html Input data DC Pump wattage 12V 6 amps 72 watts Usage per day 0 8 hours 1 minute 48 times per day Syste
74. own 28 3 2 Testing hydraulics wiring datalogger programming Before deployment the entire system should be set up in the lab and tested This enables the user to check that the sensors are configured with the correct settings proper power is being supplied to the system and that datalogger is programmed correctly to turn the pump and sensors on off to take sensor readings at correct times 3 2 1 Sensors The SUNA should be set to the SDI 12 mode via a PC and Satlantic SUNACom software so it can communicate with the datalogger The YSI should be set up to record specific conductance and temperature only If any other measurements are set up to be recorded they will overwrite the specific conductance and temperature data strings in the datalogger If the system is run with deionized water the YSI should report a specific conductance measurement near 0 mS cm and the SUNA should report a nitrate concentration slightly above 0 Even if there is no nitrate in a sample the output will never be 0 integer it will always be a small real number regardless of the units being used The YSI should be calibrated for specific conductance This can be done using the EcoWatch software with specific conductance solutions supplied by YSI The SUNA does not require initial calibration as this is done in the factory Update calibrations must be performed with deionized water as described in Section 4 3 7 3 2 2 Hydraulics Testing in the lab will al
75. pump and wires to be pulled until the existing splice is visible and out of the water Once the existing splice is exposed and out of the water the old pump can be disconnected from the wires leading to the battery and the new pump can be spliced to these wires Before closing the PVC connection it is good practice to check to make sure that the new pump runs correctly with the newly spliced wires 4 3 4 Cleaning solar panels As the solar panels become dirty their photocells collect fewer photons and their efficiency becomes reduced Solar panels should be cleaned with water or glass cleaner every few months and more often when they are more susceptible to becoming dirty such as in late spring when pollen counts are high 4 3 5 Cleaning sensors The YSI and SUNA sensors and flow cells should be cleaned routinely Since the YSI flow 46 cell retains some water between runs a film will grow over time on the bottom of the flow cell This can be removed with a flexible brush Bathroom cleaner such as Scrubbing Bubbles can also be used in the YSI flow cell Matthew Longfield YSI personal communication December 6 2010 The SUNA is more delicate and the flow cell and optical paths should be cleaned with Q tips and isopropyl alcohol 4 3 6 Site check after storms We have noticed that the flow cell and optical path on the SUNA appear to be dirty after a storm regardless of the time elapsed since previous cleaning this is most likely due to hi
76. r deployment in the streams 1 3 Site description This project uses the Baltimore Ecosystem Study BES LTER as an observational platform The BES LTER founded in 1998 is composed of the Baltimore metropolitan area Anne Arundel Baltimore Carroll Harford and Howard Counties and Baltimore City Maryland The area contains an urban core as well as older urban and suburban residential areas rapidly suburbanizing areas and a suburban rural fringe Work to date has focused on delimiting the social economic and ecological patch structure of Baltimore and documenting how the patches interact and how the patch structure has changed over time Pickett et al 2008 Long term data sets quantify the fluxes of water carbon and nutrients through the urban water cycle and terrain Weekly stream chemistry data NO3 total N POa total P 5042 have been collected for over 13 years at 11 sites time series are available at http beslter org These data have been used to quantify urban biogeochemical budgets and cycles Groffman et al 2004 Kaushal et al 2008 Shields et al 2008 The USGS MD DE DC Water Science Center maintains a series of streamflow gaging stations at which the weekly BES chemistry samples are taken and at which the sensors for this project were to be deployed The work plan called for deployment at six nested 2 suburban locations in year 1 of the project Dead Run and along a longitudinal urban rural gradient
77. rge If a site visit shows a battery is near 11 5 V and the pumps have not been running it most likely means the low voltage disconnect has been tripped and the battery should be replaced To bypass this low voltage cutoff the wires that should be connected to the load on the SunSaver Port 2 on the relay and the ground wire can be connected directly to the battery Technical support documents for Morningstar SunSaver can be downloaded from http www morningstarcorp com en support product cfm Productld 4 If the batteries are charged and the solar controller seems to be in working order and the system is still not running the datalogger may need to be reset This can be accomplished by unplugging and re plugging in the power to the datalogger 4 2 6 Ice Ice can cause two problems with the sensors the way we have them deployed out of the stream and in housings 1 water freezing in the tubing between runs and 2 ice particles moving through the system during runs that could potentially damage the sensors by torquing the flow cell parts The water in the pump could also freeze thereby not allowing for any water to be pumped to the sensors Owing to the desire to avoid 43 these problems we removed the sensors from the field sites when air temperatures started to fall significantly below freezing i e in early December 4 2 7 Interpretation of error codes and erroneous data values CR10X recording of for SUNA nitrate value Whe
78. rom high frequency sensor deployment across a range of hydrologic conditions be used to fill in the gaps from our current weekly long term monitoring to explain interannual changes in residence times and flushing of solutes e How well can a physically based watershed flow and transport model represent solute transport behavior across a range of time scales This document focuses on our field deployment experiences in from October to December 2010 and March to December 2011 and includes sections on 1 overview and choice of sensors 2 design considerations for hydraulics power data logging wiring communications and data serving 3 laboratory testing prior to field deployment 4 field deployment steps including troubleshooting 5 discussion of example data sets and 6 summary and recommendations 1 2 Choice of sensors Our choice of sensor brands and models is based on recommendations from colleagues associated with CUAHSI Consortium of Universities for the Advancement of Hydrologic Science Inc http www cuahsi org employees of the U S Geological Survey and our prior experience with vendors We chose the Satlantic http www satlantic com SUNA sensor for nitrate measurement based on positive experience with in situ testing of the Satlantic ISUS sensor by University of Florida W Graham personal communication 2008 and U S 1 Geological Survey Pellerin et al 2009 the SUNA uses the same technology as th
79. ry 31 Sample P70 42817 r Reps Of the SUNA data collected only NO3 2 8 Loc YSI_SC mg L is stored 32 Sample P70 9486 1 1 Reps 2 12 Loc Sna_N _mgL 33 Do P86 1 20 Set Output Flag Low Flag 0 34 End P95 End Subroutine 35 Beginning of Subroutine P85 1 2 Subroutine 2 subroutine to measure the stage and YSI measure stage 65 36 SDI 12 Recorder P105 measure accububbler stage 1 0 SDI 12 Address 2 0 Start Measurement aM 321 Port 4 4 Loc Abb stage 5 1 0 Multiplier 6 0 0 Offset save stage 37 Do P86 1 10 Set Output Flag High Flag 0 Julia changed array ID from 100 to 300 here on 9 27 10 so the rain gauge data could use the array id 100 38 Set Active Storage Area P80 422962 1 1 Final Storage Area 1 2 300 Array ID 1 9 27 10 Julia changed time to include year 39 Real Time P77 426036 1 1110 Year Day Hour Minute midnight 0000 40 Sample P70 28326 1 1 Reps 2 4 Loc Abb stage 41 Sample P70 425135 1 1 Reps 2 3 Loc Batt_Volt 42 Do P86 1 20 Set Output Flag Low Flag 0 43 End P95 End Subroutine 44 Beginning of Subroutine P85 1 9 Subroutine 9 subroutine to load default values 45 Do P86 1 11 Set Flag 1 High 46 Z F x 10 n P30 default station ID 66 1 9000 F 2 00 Exponent of 10 3 1 Z Loc Sta ID Default Suna warmup time is number of seconds in Sna_secON the number of seconds in the pump warm up time
80. scosity of the water which in turn increases the head loss as the temperature drops or decreases the head loss as the temperature increases This effect is significant for smaller diameter tubing 38 DR3 2 0 1 5 Nitrate mg L 0 50 0 0 11 8 10 11 9 10 11 10 10 11 11 10 11 12 10 11 13 10 11 14 10 11 15 10 11 16 10 Date DR3 2 0 1 5 m g 1 0 6 0 50 0 0 11 17 10 11 18 10 11 19 10 11 20 10 11 21 10 11 22 10 11 23 10 11 24 10 11 25 10 Date Figure 18 Example field data illustrating that a the optical path being blocked by broken O ring top figure b the sensor is working correctly bottom figure 39 Spectra Graph x Graph gt 16 Spectra 2 000 1 750 4 1 500 1 250 Light Data Range 1 000 750 500 2 000 1 500 1 000 4 500 a a a ee Dark Data Range 190 200 210 220 230 240 250 260 270 280 200 300 310 320 330 340 350 360 370 VVavelength nm Light Counts Dark Counts Spectra Graph x MOSSES KOEIE Spectra 500 ALN NLA a Dark Data Range 190 200 210 220 230 240 250 260 270 280 200 300 310 320 330 340 350 360 370 Wavelength nm Light Counts Dark Counts Figure 19 SUNA spectra illustrating that a the optical path is being blocked by a broken O ring top figure b the sensor is working correctly bottom figure 40 If the head losses are too great to be overcome by the pump water will
81. so reveal whether there are any leaks in the system Leaks in the flow cells could result from 1 manufacturing defect 2 excess pressure in the flow tubing due to unintended kinks or 3 a flow rate that is too high for the YSI flow cell If there is a kink in the tubing or if the flow rate is too high the YSI flow cell will leak from the top The flow rate should be tested with the system fully wired as shown in Figure 7 The flow rate is higher when the bilge pump is triggered from the relay than when it is connected directly to a battery This is because the relay is getting power from both the battery and the datalogger when it is active The lab testing can also provide a check as to whether the pump is sized correctly to overcome the calculated expected head losses We did this by looping the plastic tubing over a pipe near the lab ceiling to simulate the height of the lift and length required in the field As mentioned in Section 2 2 4 testing the recommended filtering system in the laboratory before going to the time and expense of field deployment led us to identify 29 problems hydraulics unrelated to turbidity removal 3 2 3 Wiring and datalogger These components should be set up exactly as intended for the field to ensure that the system is error free 4 Field deployment 4 1 Construction 4 1 1 Enclosures At 5 of our 6 stations 30 x 30 x 12 metal junction boxes by Hoffman Inc had previously been deployed by
82. t Home Depot Home Depot Home Depot Home Depot Backwoods Solar Electric Systems Inc Backwoods Solar Electric Systems Inc Global Water Supply Grainger Home Depot Mar Cor Purification Mar Cor Purification Mar Cor Purification Mar Cor Purification Grainger Grainger Home Depot Home Depot 33 00 86 00 4 97 5 97 6 99 4 69 4 69 4 69 1 97 1 95 1 95 2 78 98 00 72 00 68 00 123 00 4 15 198 37 0 00 9 77 49 00 3 56 23 60 7 00 9 70 Enclosure Hinge cover medium Hoffman box 30 x 30 x Type 3R 12 contractor Grounding rod copper 5 8 x 8 ft Grounding rod wire 8 stranded THHN green 10 ft Grounding rod clamp 1 2 bronze Galvanized pipe 2 EMT conduit x 10 Concrete Fast Set Sakrete 50 b Padlock High security Great Stuff Gaps amp Cracks 16 oz Construction materials PVC pipe material 1 1 2 x 10 PVC40 PE solidcore pipe Cap 2 slip Elbows 1 1 2 PVC 90 SXS Elbows 1 1 2 PVC 90 long sweep Elbows 1 1 2 PVC 22 5 HXH 2 x 1 1 2 DWV reducer increaser HXH 2 DWV female adapter HXFPT 2 DWV male adapter HXMPT 2 x 2 PVC PW DWV SCH40 pipe PVC primer 8 oz PVC cement 8 oz Zip ties 14 natural cable tie 10 pk 200 PC garden cable tie tube 12 x12 sheet of 100 x 100 copper mesh Copper screening 100 mesh Rebar 1 2 x 4ft Cage Plywood 15 32 or 1 2 2 ft x 2 ft Flat slotted galvanized metal xx ft
83. ta being recorded represent air temperature 4 2 5 Batteries dying solar controller and datalogger problems We had several instances where it appeared that a solar panel was not properly charging a battery at a site where the solar panel had been installed for several years We believe that this was a problem with the SunSaver controller shorting out The SunSaver is very unforgiving and if wires cross on the panel and short out the system the SunSaver will no longer be able to charge the battery We believe this may have happened and gone unnoticed There is a green LED light labeled charging on the SunSaver this light is on if a battery is connected to the SunSaver and light is reaching the solar panel However apparently this light also may be on when the battery may not actually be charging if the SunSaver has been shorted out We confirmed this using a SunSaver we knew had shorted out and was no longer able to charge a battery i e its LED green charging light was green even though it was not functioning Some SunSavers are made with a red LED light that indicates they have a low voltage disconnect This means that if the battery goes below a specific voltage 11 5 V with the SunSaver 6 the load is disconnected from the battery The system will engage the load again once the voltage reaches 12 6 V This system is intended to preserve battery life if a battery goes well below this 11 5 V cutoff it could lose its capacity to hold a cha
84. tamp to the database readable format YYYY MM DD HH MM SS The Python script runs hourly 15 minutes after the LoggerNet download The CUAHSI Streaming Data Loader SDL http his cuahsi org odmsdl html is used to automatically upload data into the ODM The SDL is a GUl based program that allows one to automatically load data into the ODM The SDL is configured by entering the appropriate site and parameter information and mapping the proper site to each data file and the parameter to a column within the data file The SDL is set to run asa scheduled task on the dedicated computer with an hourly time interval 15 minutes after the Python script The data is uploaded into an ODM database on the Hydrologic Information System HIS server at UMBC hisO9 umbc edu WaterML Data from LoggerNet Veb envies 5 PHP SOAP and CUAHSI HIS SimpleXML Observations Data CUAHSI Model SDL JPGraph Web Interface PHP and Javascript Figure 5 System architecture for serving nitrate sensor data The ODM database is configured with WaterML web services The web services expose a number of methods that allow for programmatic access to the data stored in the ODM using the Simple Object Access Protocol SOAP The methods include GetSites which returns the site information contained in the ODM Getsitelnfo which returns metadata about each site GetVariablelnfo which returns information about variables collected at a site and GetValues which returns t
85. tes a strong signal whereas a flashing light indicates a weaker signal In order to transmit data the signal must be greater than 90dBm The network connection and signal strength of the modem can be checked using AceManager either by directly connecting to the modem with a RS 232 cable or if an internet connection is available by connecting using the IP address under the UDP connection 5 Example data sets Despite the complexities of setting up securing and maintaining the sensor system we have obtained high frequency data that are not observable from weekly grab sampling Just from the initial application for two months in one season we were able to observe behavior of a number of phenomena some expected and some novel Figure 21 shows the benefit of deploying multiple sensors simultaneously The example plot is from DR3 following storm recovery This graph demonstrates the well known synchronization of high diurnal nitrate concentrations at low temperatures and vice versa The recovery of the strength of the diurnal nitrate signal following a storm is also apparent from this plot The data also illustrate the relationship between nitrate and specific conductance Specific conductance encompasses solutes that are both conservative biologically nonreactive as well as those that may be influenced by biotic activity such as nitrate The specific conductance data display a temporal pattern that mimics the nitrate data indicating the diur
86. the effects of turbidity we prepared solutions encompassing a broad range of NTU values using titanium dioxide and measured the turbidity of the solutions using a YSI optical turbidity sensor These results were used to generate a calibration curve of titanium dioxide concentration vs turbidity We then chose four levels of turbidity and nine nitrate standards to make mixtures and measure with both the IC and the SUNA sensor The SUNA was able to measure nitrate concentrations in mixtures up to 250 NTU Above 250 NTU the SUNA gave a recording of NAN The percent error between the two sets of readings IC and SUNA for turbidity less than or equal to 250 NTU was significant especially for low nitrate concentrations 1 2 mg L which are relevant for our systems These findings suggest the need for in line field filtering to remove turbidity which is consistent with recommendations made by USGS Brian Pellerin personal communication 2010 Other researchers have evaluated the effects turbidity on the SUNA readings using methods that are a variation on the above Satlantic obtained SUNA readings up to 450 NTU of turbidity made from solutions of Arizona Road Dust Geoff Maclntyre personal communication 2011 USGS obtained similar results using a silt loam IHSS standard Brian Pellerin personal communication 2011 The next generation SUNA to be released in early 2012 is designed to operate in extremely turbid environments gt 1500 NTU
87. the pump used We observed bubbles were coming out of the pump after the pump shut off and water 45 back flowed from the sensors We identified that air was being pulled into the outflow of the SUNA and YSI outflow tubes from the backflow process To fix this problem the outflow tubes were extended into the stream downstream of the pump so that they were submerged and that water instead of air was pulled back into the plastic tubing 4 3 Maintenance 4 3 1 Power If solar panels are not installed at each location the batteries need to be replaced ona calculated schedule This can be done right before the programmed pump switch on time to make sure the system turns on when it is supposed to after installation of a new battery 4 3 2 Tubing At each trip to the site the tubes in the enclosure and the outflow tubes should be checked to make sure they are not crimped or becoming clogged Plastic tubing will biofoul over time and should be monitored and cleaned or replaced when needed 4 3 3 Pump replacement To replace a pump the splice that connects the power wires of the pump to the wires leading to the battery needs to be exposed and taken apart Since these wires are all inside the PVC the easiest way to splice in the new pump is to make the splice near an unglued connection in the PVC for example at the male and female threaded adaptors Figure 16 Further there also needs to be extra wire left near the pump battery to allow the
88. the relay for the bilge pump connects to positive load on one SunSaver The bilge pump ground connects to the negative port on the same SunSaver The power wire on the bilge pump connects to Port 2 on the bilge pump relay Port 3 on the pump relay is connected to channel C5 on the datalogger On the relay used for the SUNA Port 1 connects to the power black wire of the SUNA pigtail Port 2 on this relay connects to the 12V power in on the datalogger Port 3 connects to C4 on the datalogger For both relays Port 4 connects to ground There are two 12V channels on the bottom right of the datalogger The power on the YSI and Raven can connect to either of these channels and the ground for each can connect to any ground The data channel purple wire on the YSI connects to channel C2 on the datalogger The data channel green wire on the SUNA pigtail connects to channel C3 on the datalogger We used a wiring terminal to help support the SUNA wires The wires on the SUNA pigtail are a higher gage smaller wire than those used by the other components We connected the wires on the SUNA to a terminal as a junction and then ran wire from the terminal to the intended ports on the relay and datalogger This enabled us to position the terminal and SUNA wires so that there would not be as much physical stress placed on the small diameter wires These terminal strips are typically sold with 4 6 or 8 channels More of these terminals can be used if desired t
89. ts were set in the face of the bridge by drilling a 1 72 deep hole the same diameter as the bolts to be used and hammering to wedge the ends of the bolt into the holes Four bolts were used to mount two U brackets spaced vertically such that 2 inch galvanized pipe could be slid through them A horizontal bolt in the U brace was tightened to hold the galvanized pipe in place All solar panels were positioned to face south and to match the tilt of the solar panels already deployed at the sites by USGS 4 1 5 Wiring Before inserting the bilge pump into the PVC housing the existing power and ground wires on the pump need to be extended to the sensor enclosure After trenching 16 gage wire can be run from the pump to the enclosure and then cut leaving some excess This wire can then be then spliced to the pump wires by stripping the end of each wire and connecting them using a wire nut Since water may reach this spliced connection it should be sealed this can be done by wrapping the wire nut with electrical tape covering with heat shrink a material that shrinks and makes a tight seal when heated and wrapping once more in electrical tape to ensure the ends of the heat shrink wrapping are sealed Since the pumps have a finite lifetime it is advisable to make a splice in the wires somewhere near an easy access connection in the PVC The overall wiring diagram utilized in the enclosure is shown and discussed in Section 2 6 As mentioned in
90. ump on or off and the pump will remain on 4 2 9 Priming pumps If either the Rule or Wale pump sits in the PVC without water for more than one pumping cycle due to abnormally low base flow or some type of clogging of the PVC the pump must be re primed The Rule pump seems to re prime more easily For the Rule pumps while the pump is running water can be forced by hand into the perforated PVC and the pump will prime and begin to work When re priming the Wale pumps the 2 PVC must be removed to expose the pump and the plastic tubing must be removed from the pump s outlet port The pump can then be submerged and turned on Once water is forced into the inflow port and the pump is primed it can then be shut off and the plastic tubing can be reattached to the outflow port If the 2 inch PVC is removed for cleaning or troubleshooting and the Wale is removed from the water it typically needs to be re primed before it will run properly again This does not seem an issue with the Rule pumps as long as they are not turned on while they are out of the water 4 2 10 Relay shorting If the pump continues to run regardless of the time on the CR10 X the relay may have shorted out and need replacing This can happen if the wire connected to Port 1 on the relay is disconnected or reconnected while the system is supposed to be running 4 2 11 Backflow bubbles At one site our pumps would not self prime between runs regardless of

Technical!Report

Contents

Download Pdf Manuals

Related Search

Related Contents