4/30/13, Boise State University, Dept of Geosciences, Hydrologic Sciences, Jim McNamara Research Director (1999-present), Pam Aishlin site maintenance and data management. Data provided herein is for the Lower Gage (LG) stream measurement site of Dry Creek Experimental Watershed, Boise, Idaho. DATA STREAM: 1. Measurement site was installed by Boise State University in 1999 with data available since 2/5/1999. Instrumentation installed varies over time and may include sensors for stream stage, stream temperature and electrical conductivity measurement. Site maintenance and data collection is provided by Boise State University Hydrologic Sciences. 2. Data files are provided by Campbell Scientific dataloggers and/or self logging instruments, and retrieved either via site visit or telemetry. Files are archived at Boise State University, Hydrologic Sciences, Dept of Geosciences. These raw data files occur at varied data collection intervals, e.g. 10 minute, 30 minute or hourly. ***NOTE: Raw data files are level 0, raw data (CUAHSI HIS standards). (CUAHSI HIS standards, 0 = raw data, 1 = Quality controlled data that have passed quality assurance procedures such as routine estimation of timing and sensor calibration or visual inspection and removal of obvious errors, 2 = Derived products that require scientific and technical interpretation and may include multiple-sensor data. An example is basin average precipitation derived from rain gages using an interpolation procedure.) 3. Select data files are quality checked and post processed for gapfilling, noise filtering and/or corrected according to established instrument calibrations. For stage data, random noise and negative values occuring over short time intervals are replaced with proximate average stage values. During the winter months ice may affect data accuracy, for obvious ice-induced noise (> 5cfs change per time increment occurring in conjunction with freezing temperatures), values are replaced with time range before/after average values. For electrical conductivity, in some instances, negative values are reported when stream is dry, wires are disconnected or other failure occurs. These noise values are replaced with -6999, no data values. Stage-discharge rating curves are applied to continuous stream stage data for output of continuous discharge time series. Where feasible, gaps occuring in the discharge timeseries due to sensor failure/error may be filled via interpolation/regression from same site supplementary instruments or other sites within the same watershed. Raw data timestamps may vary from final output timestamps, in which case raw data will be interpolated for output on the hour. 4. Data as collected and provided is, by protocol, in mountain standard time, year-round. SITE NOTES: Within DCEW, Dry Creek fluvial physical characteristics and process vary between plane-bed, step-pool and cascade reach types (Montgomery-Buffington Classification for Pacific NW mountain fluvial environments). Channel slope is approximately 0.03 at the LG stream reach. Riparian vegetation is abundant. This Lower Gage site serves as the outlet measurement point for DCEW (26.9 km2). At this location, the channel is confined, set within a steep v-shaped section of the watershed that includes exposed bedrock. While bank full stage is approximately 3 ft (1m) at this site, record flood stage is indicated at 4 ft (January 17, 2011 and spring runoff 2006, 100-126 cfs, 2.8-3.6 m3/s). In some years, flow ceases at LG late summer and resumes in September as daily average solar input decreases. Instrumentation at this site is installed within stilling wells at stream channel bottom within a natural pool from which water drains freely over bedrock. Limited sand accumulation occurs within this pool, while bank configuration remains relatively stable. On-site measurement of established staff gage values are regressed against on-site discharge measurement to determine stage-discharge rating curves for the site, annually and subannually. On-site discharge measurement has been conducted with velocity-area, dilution and slope-area survey based methods. DATA/INSTRUMENT NOTES: Capacitance logger - Capacitance loggers used prior to 2007 were Trutrack (manufactured in New Zealand) water level sensors that also provided water temperature and air temperature. Capacitance loggers used since 2007 at our stream sites have been purchased from Odyssey, also based in New Zealand. These sensors provide waterlevel data only, resolution 0.001 ft. 2016-2017 recession lag observed in data. Fall 2017 additional perforations added to stilling well installation to reduce possible water level lag. Solinst - not used at LG. Solinst self-logging water level and barologgers are acquired from the Canada-based Solinst company. Pressure transducer - Pressures transducers employed at our stream sites 1999-2013 include Global and Druck water level sensors. These sensors include atmospheric pressure compensation via vent tubing encased within the sensor cable to which a vented dessicant capsule is attached at cable end (within the datalogger box). The sensors are wired and programmed using Campbell Scientific dataloggers, CR10x prior to 2013. LG 2013, CR10x. Sensor data resolution is 0.01 to 0.001 ft. Models employed at our sites are selected for expected maximum water depth <10 ft. A lower precision, higher range Druck pressure tranducer was employed at LG late 2003 till mid 2006. Druck models employed include PDCR 940, 1230 and 1830 series. Druck pressure transducer installed summer 2006 is 1 PX DC PDCR 940 SN651580. 10 V. 2.5psig. Sep 2014 a new pt, Keller, is installed, co-located with the existing PT for comparison. Results were comparable. The new Keller PT is located across the stream from the existing site, deployed vertically in a pvc stilling well (moved to horizontal pvc installation summer 2017). Both PT sensors are located in natural pool upstream of a cascade drop. 2017 installation improvements: Reinstalled Druck in screened pvc w/ additional perforations made in stilling well pipe. Goal is to promote rapid water level adjustments in pvc pipe. Keller is reinstalled in horizontal position in horizontal-vertical "L" joined pvc, also well perforated. Base caps are used at both installations. Both sensor's stage data are offset to correspond with pool located staff gage. program notes: Druck uses a differential measurement. The Keller uses single ended voltage measurement. 'Keller Acculevel sensor 0-5 vdc output for 0-5 psi ' supply is 8-28 vdc, therefore use SW12v SW12 (1) Delay (0,60,Sec) VoltSe (stage,1,mV5000,5,1,250,250,0.0023067,0.77)'offset was -.155 090917 changed'using slight delay after sw12 '2.3067 ft water/psi and div by 1000 to convert to volts; 0.145833 ft for 1.75 inches offset for PT SW12 (0) Ultrasonic depth sensor - Constructed at Boise State University using thermisters and a Maxbotix sonar sensor, this sensor was installed in 2012 at LG for testing as a viable option for measurement of water depth. Thus far, the sensor has performed well, although data provided includes noise indicative of a rough water surface. NOTE ON WATER LEVEL SENSORS COMPARISONS: Typically one or more water level sensors are colocated at this site to provide backup data, but also to provide comparison of sensor accuracy and performance under various conditions. We find that pressure related sensors, such as the solinst (which must be corrected with barometric data) and pressure transducers, which are typically vented to atmosphere tend to have diurnal noise, possibly related to diurnal atmospheric pressure signals not matched by pressure signals beneath the water surface. With the ultrasonic depth sensors, diurnal noise seems to occur due to incomplete capture of air column temperature and/or sensor housing-electronic board temperature effects. Due to sonic readings off the water surface, the ultrasonic sensor contains includes noise as capture of the turbulent water surface (when flow is particularly turbulent) The capacitance rod is has a greated depth from bottom physical offset, thus is less accurate near no-flow conditions. However, the capacitance rod seems to occur overall less diurnal noise. Keep in mind, however, that diurnal water level fluctuation does actually occur, esp during warm weather growing season months. Conductivity-Temperature sensor - Campbell Scientific conductivity sensors employed at our sites include CS547A probes which require the Campbell A547 datalogger interface device. Older 247 models were employed prior to 2004. Data resolution is assessed as 0.001 ms/cm. Sep 2014 a new CS cond-T sensor is installed, co-located with the existing sensor for comparison. Both sensors are located in natural pool upstream of a cascade drop. The existing sensor is verified as outputting Temp data 3.7 to 5 degrees high, as suspeced. Cond data in existing sensor is muted and less reliable compared with new sensor, as suspected. 'CONDUCTIVITY PROBE PROGRAM INSTRUCTIONS: BrFull(ecRS, 1, mV2500, 1, VX1, 1, 2500, True, True, 0, 250, -0.001, 1) ecRS = 1*(ecRS/(1.0-ecRS)) ANNUAL ERRORS/GAPS/UPDATES: 1999- instruments include pressure transducer and conductivity-temperature sensor. Stage-discharge avail from 2/5/99. ec and stream temp not avail until 9/26/99 (provided upon request). Data range 329 days, not accounting for included minor gaps in data. Offset added in post processing following Nov data gap-PA. 2000- instruments include pressure transducer and/or capacitance water level sensor. Campbell Scientific 10x datalogger. Data range 366 days, data gaps, 2000: 4/27-5/18, 6/11-7/9, 7/31-9/28, 10/6-10-27, 11/17-11/30 2001- instruments include pressure transducer and/or capacitance water level sensor. Data range 319, data ends 11/15/2002. 2002- instruments include pressure transducer and/or capacitance water level sensor. 2003- instruments include pressure transducer and conductivity-temperature sensor. Data range 209 days. Conductivity-temperature sensor is provided upon request. Capacitance data avail from 10/27/03. 2004- instruments include pressure transducer (Druck higher range, lower precision model), capacitance water level sensor and conductivity-temperature sensor. New instruments, PT and cond-T, installed 2/13/04 with stage data avail 3/27/04. T and ec avail consistently from 7/2/04. Data range is 365 days for capacitance sensor. Capacitance data considered more reliable for stage in 2004. suggested rc is a=4.1296 b=3.6175 ***2004-present: Timestamp of raw data may vary from final output timestamp wherein raw data is interpolated for output on the hour. Random noise and negative values are removed via replacement with proximate average stage values. During the winter months ice may affect data accuracy, for obvious ice induced noise (> 5cfs change per 30 min in conjunction with freezing temperatures), values are replaced with time range before/after values for average values. For ec, in some instances negativevalues are reported when stream is dry, wires are disconnected or other failure occurs, these noise values are replaced with -6999. 2005- instruments include pressure transducer (Druck higher range, lower precision model), capacitance water level sensor and conductivity-temperature sensor. Capacitance data considered more reliable for stage. Data range 365 days. Gaps 1/15/05-1/30/05. 9/11/05 7:49, 9/24/05-9/25/05 9:49-10:49. suggested rc is a=4.1296 b=3.6175 2006- instruments include Druck pressure transducer (1 PX DC PDCR 940 SN651580. 10 V. 2.5psig. Installed summer 2006), capacitance water level sensor and conductivity-temperature sensor.Capacitance logger (internal battery) failed 1/6/2006. Data range is 350 days. Data gaps: 2/7/06-2/9/06, 3/11/06 500-1100 and 1400-1800,3/14-3/15/06, April 4 through May 8 LG data unreliable due to sensor dislodge during peak streamflow conditions. If a gapfilled data set is desired for LG inclusive of 4/4/2006-5/8/2006, Confluence 2 Main discharge can be considered as an approximation of discharge at LG during this time period. gaps 8/31/06 - 9/22/06. 10/14-10/19/06, 10/27 1000-1300, 11/1-15, 12/9-16 2006. 12/16-12/31/06. Prior to 4/4/06 peak discharge rc used power relationship a=4.1256 b = 3.6175, PT unreliable 4/4/06 -5/8/06; 5/8/06-8/31/06 rc a= 0.0131 b=8.3391. new rating curve as of 9/22/06 a=2.5465 b=4.5316. 2007- instruments include pressure transducer and conductivity-temperature sensor. Data range 365 days. Gaps: 12/16/06 -1/22/07. (data on 1/12,13/2007 is ice effected), 1/27 300, 1/27-2/6/07,2/26 1600, 3/11-121500, 3/26-27 1100,7/28-8/26, 10/5-6 1100, 10/14 1300-1700, 11/27-28 1400. new rating curve as of 1/22/07 a=2.5723 b=3.91. ., alternate suggested rc is a=2.4518 b=4.2 for 2007-2008. 2008- instruments include pressure transducer and conductivity-temperature sensor. Conductivity-T probe removed for calibration check 8/29/2008-9/29/2008, conductivity incurring sporadic error after replacement. Data range 365 days. rating curve applied for 2008 to 5/30 2100 is a=2.5418 b=4.2., after 5/30 2100 a=2.4522 b=4.2, alternate suggested rc is a=2.4518 b=4.2 for 2007-2008. 2009- instruments include pressure transducer and conductivity-temperature sensor. Data gaps: No discharge data 3/6 to 3/12/2009 due to instrument/program failure. Data range 365 days. gap 4/5/09 1500-1900. New rating curve for 2009 continues as slightly adjusted from that of 2008, replaced with a=2.4592 b=4.2593. For new rating curve applied following summer 2009 thunderstorm and sediment deposition 7/25/09 1100, a=2.8027 b=3.7152 . 2010- instruments include pressure transducer and conductivity-temperature sensor. Hourly data, data interval changed to 30 min on 6/14/2010 1230, changed to 15 min data. Data range 365 days. gap: 7/16/10 1100, 7/23/2010 to 7/28 due to low bat. 8/12 to 9/9 due to program changes (low flow). New rc for 2010, a=2.8646 b= 3.6876 for low q, high q a=10.493 b= 1.6587 with stage break at 1.95 ft where rc intersects with low q rating curve. From 9/9/10 changed to a=2.8475 b=3.607 (rating curve values are approximate, because rc used changes slightly as new measurements are conducted on site) 2011- instruments include ultrasonic water level sensor, Odyssey capacitance waterlevel sensor, Druck pressure transducer, Campbell Sci conductivity-temperature sensor. Data range 365 days. data is collected at 30 min interval. Data gaps: 7/19/11 16:00, 8/25-8/31/2011 wk at site. rating curve applied to stage data. low Q a=2.8475 b=3.607, high Q a=10.493 b=1.6587 break at stage 1.95 ft, rating curve valid as of 8/30/11. Jan 17 2011 rain on snow event, applied slope-area eqn to rating curve. 2012- instruments include ultrasonic water level sensor, Odyssey capacitance waterlevel sensor, Druck pressure transducer, Campbell Sci conductivity-temperature sensor. Data range 365 days. data is collected at 30 min interval. Data gaps: 1/12/12-1/17/12 filled with caprod data for Q, -6999 for ec and T. 9/23 to 10/1/12 due to low bat. discharge gapfilled via cap logger data, 11/2/12 1930 - 11/3 930 low bat. similar 11/3 - 11/4 /12, 11/4 - 11/6/12. gap 12/11-12/22,gapfilled via cap logger data. rating curve applied to stage data. new rating curve after 8/30/11 applied to 2012 data low Q a=2.9055, b=3.691 high Q a=10.065 b=1.6882 break at 1.86 ft. Rain on snow event (ROS) Jan 19 2012, complicated, snow,then rainfall. PT data presented herein for LG is assessed as reasonable. ALSO 2/22/12 ROS. 2013- instruments include ultrasonic water level sensor, Odyssey capacitance waterlevel sensor, Druck pressure transducer, Campbell Sci conductivity-temperature sensor. NOTES IN PROGRESS: Gaps filled for PT data and ice effects removed for stream discharge Jan 1-8, 2013, using secondary instrument (caplogger data). Additional suspect noise occurs mid through late February that is not corrected; it is suggested that noise removal due to ice effects upon this streamflow data be considered in conjunction with freezing conditions at the stream site, precipitation occurring as snow vs. rainfall and watershed air temperatures that would affect timing of snowmelt. Noise has been removed Feb 10-11 and March 23-24, using avg of previous 24 hrs, due to freezing conditionsconcurrent with no precipitation. No flow conditions occurred mid July through late Sep. Data gap for stage-discharge 10/22-23 also filled with collocated cap rod data. 2014- Ice corrections made for 1/4 0000-1900 and 1/19 - 1/22. gap 5/18/ - 5/19. Site instrumentation and procedures remains the same as above for 2013. However a new measurement system was added in August 2014. This system is co-located with the existing station Aug-Dec 2014 for collocated sensor performance comparisons. Data for 2014 (to 11/1/2014) remains sourced from the original measurement system, cr10x w/ pt. A period of no flow occurred late summer 2014. Data from 11/1/2014 is sourced from the new measurement system. See sensor notes above. No caprod data 10/7/14 to 12/9/14. Comparison of new and existing sensors verified that old cond-T sensor provides T data that is offset 3.7 degrees celsius erroneously high. T data has been corrected accordingly for our final data file 2014. Fall discharge at LG is verified by site visits as unusually low, compared to upstream sites and prior years. This may indicate environmental change. Note stream observed as barely flowing 7/21/2014, no flow 8/29/2014. 10/7/2014 staff stage is only 0.58, at which flow is less than 1cfs. 2015- Rating curve Hi-Lo Q intercept 1.55 ft 19 cfs Rating curve power eqn = multiplier, exponent, Rsquared value. RCparametersLo = 3.5908 3.2688 , R2 0.9942 RCparametershigh = 8.567 1.806 , R2 0.987. Processed via matlab. PT and CAP RC are identical. Minimal flow above 1.55 ft occurred. Rain on snow early 2015 resulted in early peak stream flow. Minimal ice effects error, not corrected (e.g. apprx Dec 3, 2015). Stage data for discharge is provided by capacitance logger, except for low flow period (7/1 - 9/21/2015) for which stage data is provided by Keller PT due to capacitance installation lower limit. No flow is observed on site 7/4/15 and 8/28/15. The new Keller PT incurs drift repeatedly, corrected with offsets based on site visit staff gage readings. 9/22/15 Druck PT removed from cr10x and added to Cr1000 along with Keller PT and new cond probe. Druck data is less precise with cr1000 output. Old and new conductivity probes were colocated for testing during 2015. Temperature values on old cond probe were assessed as being a few degrees high. 2016 - Rating curve, Hi-Lo Q intercept 1.55 ft 19 cfs RCparametersLo = 3.5908 3.2688 , R2 0.9942 RCparametershigh = 8.567 1.806 , R2 0.987. Processed via matlab. PT and CAP RC are identical. Peak flow 1166 L/s, 41 cfs, early March 2016. Low flow - No flow 7/23 - 9/19. Note RC (rating curve) is less accurate near zero, discharge is site-visit-verified as zero vs. rating curve presentation as slightly above zero during no flow periods. Stage data for discharge is provided by capacitance logger, except for recession and low flow period (4/16 - 9/16/2016) for which stage data is provided by Keller due to capacitance installation lower limit. Continued drift occurs w/ Keller PT, corrected with offsets. Druck PT data corresponds w/ capacitance stage, though druck is less precise. 2017 - Rating curve, y=m*x^b. Hi-Lo Q intercept 1.541 ft 19 cfs RCparametersLo = 3.59 3.27 , R2 0.995 RCparametershigh = 8.728 1.799 , R2 0.985. Artifacts of the dual rating curve application may occur proximal to dicharge at 19 cfs, such as observed on 5/10/17 when discharge fluctuates abruptly from 19 cfs to 13 cfs. Stage data on this date verify gradual rise/fall in water level. RC (rating curve) is less accurate near zero, discharge is site-visit-verified as zero vs. rating curve presentation as slightly above zero during no flow periods. Collocated sensors: 2 sonar sensors, Druck PT, Keller PT, capacitance logger. Offsets were applied in matlab final processing for dates 3/31 during peak flow and 6/17 during recession. Otherwise, offsets applied in logger program at site visit/calibration, are effective. Druck PT data corresponds roughly w/ capacitance stage, though druck is less precise. Note RC (rating curve) is less accurate near zero, discharge is site-visit-verified as zero vs. rating curve presentation as slightly above zero during no flow periods. Observed no flow 9/9/2017 staff at 0.015 ft (corrected for site visit noise 9/9). Sensors removed and cleaned, reemplaced. New Echo sonar water level sensor emplaced 10/3/17. Keller continues to incur drift during 2017. New 9/9/2017 emplacement failed to solve the problem. Therefore, the Keller PT sensor is assessed as unreliable at this installation, particularly after 3/1/2017. Final published discharge 2017 is based on capacitance logger stage, except during recession and low-no flow conditions due to apparent drift and lower limits. The USD (SnowBoe,ultrasonic depth sensor) waterlevel sensor data is used for this recession-low/no flow period 4/29/2017 0600- 9/15/2017. Low-No flow is indicated intermittently from late Aug to mid Sep. The caplogger minimum discharge value, 2.38 L/s, is assumed from 9/15/2016 until 9/18/2016, when the capacitance logger minimum value engaged at 2.38 L/s and began to rise. Ice effects occur in stage data resulting in erroneous peak spikes during low flow freeze events, e.g. 2/20/2018. These are left to the user to assess via comparison with stream temperature and correct as needed, according to data use. Interpolation of flow through freezing events is a reasonable approach to correction. 2018 - Rating curve, Hi-Lo Q intercept 1.541 ft 19 cfs RCparametersLo = 3.548 3.287 , R2 0.995 RCparametershigh = 8.728 1.799 , R2 0.985. Discharge data for 2018 is provided from sensors as follows: Use caprod (Odyssey capacitance water level) 2018 to spring runoff 3/22/18. Use usd from 3/22/18 1300 when caprod nears partial failure and caprod = usd. Usd fails 4/29/18 1300, and caprod = usd, use caprod. Use druck from 7/13/18 10:00 following program correction to address low precision in data. Minor offsets applied throughout the year to maintain data alignment with onsite measured staff gage values. The druck remained the most consistent against data drift. All sensors are located such that physical drift is unlikely. Minor gap occurs 8/23-24/2018, druck program experiment. Jan and Feb peaks are indicated as ice-effects and are corrected herein, using -6999 for no data values. 2/19/18 - 3/10/18 usd indicates snow over ice and/or ice. USD (SnowBoe, ultrasonic depth sensor) temperature wire loose spring 2018, corrected with rewiring and program change accordingly, completed on 7/9/2018. Gaps at program changes 8/24 and 11/3. USD sensor failure/removal 11/3; end telemetry of usd data; Install echo sonic water level sensor. 2019 - Rating curve, Hi-Lo Q intercept 1.549 ft 19 cfs RCparametersLo = 3.548 3.287 , R2 0.995 RCparametershigh = 8.599 1.811 , R2 0.985. Minor offsets applied throughout the year to maintain data alignment with onsite measured staff gage values. Druck data used for most of 2019 discharge calculation. However, druck sensor began indicating problems via noisy data after May 2019. Therefore data for most of the remainder of 2019 is provided by capacitance rod sensors at the site. Exception is a gap fill by druck data 10/25 - 11/24 2019. 2020 - Rating curve RCparametersLo = 4.0907 3.3783 , R2 .9959 Druck water level data is exceptionally noisy for 2020, therefore Capacitance logger data is used for discharge. 2021 - Rating curve, same as for 2020, verified by 2021 gaging. Druck water level data is exceptionally noisy for 2020, therefore Capacitance logger data is used for discharge. Note that stage below 0.32 ft is assessed as out of range for the caplogger (below base of sensor as installed) and/or invalid due to ponding in natural pool at this stage. Thus -6999 no data values will be emplaced for this stage and lower. Discharge at 0.32 ft is 2.47 l/s. A new pressure transducer, CS451, is installed in addition to above sensors in 202. However, the CS451 result is similar to druck pressure transducer, also with a significant diurnal signal, range up to 2 ft per day with druck, 1.5 ft w/ CS451. The caplogger, in contrast, has a minimal diurnal signal at less than 0.1 ft. CAPLOGGER DATA is used for 2021, except for 8/27 11:00 to 10/12 13:00 due to data gap at caplogger. For this brief period, the CS451 is utilized. 2022 - Rating curve, same as for 2020. Due to minimal runoff this spring, the rating curve is not updated this year. Ice effects on stage/discharge may explain apparent increases in stage/discharge Jan 28-31, Feb 23- Mar 2, and March 10. This is indicated by freezing stream conditions and a lack of expected discharge induced change in conductivity. (e.g. not real increases in discharge). A new cs451 was installed Aug 2022. For 2022, to provide consistency, caplogger data is utilized for water level/discharge. Peak observed discharge is mid May at apprx 10 cfs, 563 L/s. A data gap occurs 5/20/22 - 6/21/22. Note that stage below 0.32 ft is assessed as out of range for the caplogger (below base of sensor as installed) and/or invalid due to ponding in natural pool at this stage. Thus -6999 no data values will be emplaced for this stage and lower. Discharge at 0.32 ft is 2.47 l/s. Due to low battery at the Campbell logger, a data gap occurs for conductivity and temperature data (with sporadic data available) 6/2/22 - 8/26/22.