CSU ROSS HydroShare ReadMe

This folder contains high-frequency (15-minute) water quality sensor data collected by the ROSSyndicate group at Colorado State University in the Cache la Poudre River Watershed in Northern Colorado, United States. Data collection began in Fall 2019 and data collection is currently ongoing. This publication currently includes data collected between 9/2019 - 11/2025.

Publication Structure

This publication is updated annually once data has passed quality assurance procedures and has been both automatically and manually verified by project technicians. As such, we maintain a folder for each year's data and a general metadata folder containing site locations, delineated natural watersheds, and sensor procedures.

Generally, each year's folder contains a raw (CSU_ROSS_*year*_raw.csv) and cleaned (CSU_ROSS_*year*_cleaned.csv) dataset. Both datasets are provided in long format. Copies are also available in .parquet format. The following parameters were collected:

• Water Depth (m)
• Water Temperature (C)
• Dissolved Oxygen (mg/L)
• pH
• Turbidity (NTU)
• Specific Conductivity (µS/cm)
• Chl-a Fluorescence (RFU)
• FDOM Fluorescence (RFU)
• ORP (V)

Note: Not all parameters were collected at all sites or during all years. Each year's dataset may vary slightly based on available resources, context, and field notes used to verify or modify raw data. Where available, technician field notes will be uploaded with each year's dataset. Specific details for each year are provided below.

2025 Data

The data collected in 2025 was processed using the following script: poudre_sonde_network/data_organization/Process_2025.Rmd

Columns in cleaned dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_final: Value for the parameter after passing through QAQC procedures. Erroneous data was removed and, where relevant, raw data was modified to correct for calibration or biofouling errors.
• user_qc_flag: Quality control flag describing why datum was removed or if a data point is flagged/suspect
• modifications: Describes the quality assurance procedures applied to a specific data point, which include:
• calibration correction: Datum corrected due to a calibration error or calibration drift.
  • See Calibration Correction Tool for details
• drift correction: Sensor biofouling on Turbidity or FDOM sensors resulted in drift, which was corrected using linear or exponential functions.
  • See poudre_sonde_network/data_organization/drift_correction_2025.Rmd for details
• manual data exclusion: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
  • See poudre_sonde_network/manual_verification_tool for verification tool workflow
• NA : No modification of datum.

Columns in raw dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_post_cal_corr: Value after applying calibration corrections using the Calibration Correction Tool
• value_post_ver: Value after removing data identified by manual verification via the Manual Verification Tool. Also documented in user_qc_flag and qc_status.
• value_post_drift_corr: Value after correcting for sensor drift due to biofouling. For parameters where drift was not corrected, this matches value_post_ver.
• auto_qc_flag: Flags generated by the auto-qaqc system in the {ross.wq.tools} package. These serve as the basis for comparison with human-verified outcomes.
• qc_status: Final status of a data point:
• OMIT: Datum is considered erroneous and should not be used.
• FLAGGED: Datum may have minor issues; usable for most analyses but should be excluded if a high degree of accuracy is required.
• PASS: Datum is considered valid and accurate based on human verification.
• user_qc_flag: Quality control flag describing why a qc_status was applied.
• modifications: Summary of procedures (calibration correction, drift correction, or manual exclusion) applied to the point.

Field Notes (CSU_ROSS_sensor_field_notes_2025.csv)

This file contains field notes collected by technicians throughout the year.

• site: Site identification code.
• visit_comments: Notes on site conditions and work completed.
• DateTime_UTC: Visit start time in UTC.
• sonde_employed: Sensor status following the visit:
• 1: Sensor was removed following the visit.
• 0: Sensor was deployed during the visit.
• NA: No change (sensor remained deployed).
• pre_clean, post_clean, and post_cal values: Values recorded during visits to quantify the effects of cleanings or calibrations.
• cals_performed: List of calibrations performed during the visit.

2024 Data

The data collected in 2024 was processed using the following script: poudre_sonde_network/data_organization/Process_2024.Rmd

Note: Due to limited ports on some AquaTROLL models, the following site-parameter combinations were measured on a second sonde affixed to the primary sonde: archery-Turbidity, archery-FDOM, riverbend-FDOM, riverbend-chl-a, cottonwood-FDOM, cottonwood-chl-a. There were minor differences (\< 5 cm) in deployment depth between sondes.

Columns in cleaned dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_final: Value for the parameter after passing through QAQC procedures. Erroneous data was removed and, where relevant, raw data was modified to correct for calibration or biofouling errors.
• user_qc_flag: Quality control flag describing why datum was removed or if a data point is flagged/suspect
• modifications: Describes the quality assurance procedures applied to a specific data point, which include:
• calibration correction: Datum corrected due to a calibration error or calibration drift.
  • See Calibration Correction Tool for details
• drift correction: Sensor biofouling on Turbidity or FDOM sensors resulted in drift, which was corrected using linear or exponential functions.
  • See poudre_sonde_network/data_organization/drift_correction_2024.Rmd for details
• manual data exclusion: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
  • See poudre_sonde_network/manual_verification_tool for verification tool workflow
• non-resolved drift: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
• NA : No modification of datum.

Columns in raw dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_post_cal_corr: Value after applying calibration corrections using the Calibration Correction Tool
• value_post_ver: Value after removing data identified by manual verification via the Manual Verification Tool. Also documented in user_qc_flag and qc_status.
• value_post_drift_corr: Value after correcting for sensor drift due to biofouling. For parameters where drift was not corrected, this matches value_post_ver.
• auto_qc_flag: Flags generated by the auto-qaqc system in the {ross.wq.tools} package. These serve as the basis for comparison with human-verified outcomes.
• qc_status: Final status of a data point:
• OMIT: Datum is considered erroneous and should not be used.
• FLAGGED: Datum may have minor issues; usable for most analyses but should be excluded if a high degree of accuracy is required.
• PASS: Datum is considered valid and accurate based on human verification.
• user_qc_flag: Quality control flag describing why a qc_status was applied.
• modifications: Summary of procedures (calibration correction, drift correction, non-resolved drift, or manual exclusion) applied to the point.

Field Notes (CSU_ROSS_sensor_field_notes_2024.csv)

This file contains field notes collected by technicians throughout the year.

• site: Site identification code.
• visit_comments: Notes on site conditions and work completed.
• DateTime_UTC: Visit start time in UTC.
• sonde_employed: Sensor status following the visit:
• 1: Sensor was removed following the visit.
• 0: Sensor was deployed during the visit.
• NA: No change (sensor remained deployed).
• pre_clean, post_clean, and post_cal values: Values recorded during visits to quantify the effects of cleanings or calibrations.
• cals_performed: List of calibrations performed during the visit.

2023 Data

The data collected in 2023 was processed using the following script: poudre_sonde_network/data_organization/Process_2023.Rmd

Note: Due to limited ports on some AquaTROLL models, the following site-parameter combinations were measured on a second sonde affixed to the primary sonde: archery-Turbidity, archery-FDOM, riverbend-FDOM, riverbend-chl-a, cottonwood-FDOM, cottonwood-chl-a. There were minor differences (\< 5 cm) in deployment depth between sondes.

Columns in cleaned dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_final: Value for the parameter after passing through QAQC procedures. Erroneous data was removed and, where relevant, raw data was modified to correct for calibration or biofouling errors.
• user_qc_flag: Quality control flag describing why datum was removed or if a data point is flagged/suspect
• modifications: Describes the quality assurance procedures applied to a specific data point, which include:
• calibration correction: Datum corrected due to a calibration error or calibration drift.
  • See Calibration Correction Tool for details
• drift correction: Sensor biofouling on Turbidity or FDOM sensors resulted in drift, which was corrected using linear or exponential functions.
  • See poudre_sonde_network/drift_correction_tool for details
• manual data exclusion: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
  • See poudre_sonde_network/manual_verification_tool for verification tool workflow
• non-resolved drift: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
• NA : No modification of datum.

Columns in raw dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_post_cal_corr: Value after applying calibration corrections using the Calibration Correction Tool
• value_post_ver: Value after removing data identified by manual verification via the Manual Verification Tool. Also documented in user_qc_flag and qc_status.
• value_post_drift_corr: Value after correcting for sensor drift due to biofouling via the Drift Correction Tool. For parameters where drift was not corrected, this matches value_post_ver.ver`.
• auto_qc_flag: Flags generated by the auto-qaqc system in the {ross.wq.tools} package. These serve as the basis for comparison with human-verified outcomes.
• qc_status: Final status of a data point:
• OMIT: Datum is considered erroneous and should not be used.
• FLAGGED: Datum may have minor issues; usable for most analyses but should be excluded if a high degree of accuracy is required.
• PASS: Datum is considered valid and accurate based on human verification.
• user_qc_flag: Quality control flag describing why a qc_status was applied.
• modifications: Summary of procedures (calibration correction, drift correction, non-resolved drift, or manual exclusion) applied to the point.

Field Notes (CSU_ROSS_sensor_field_notes_2023.csv)

This file contains field notes collected by technicians throughout the year.

• site: Site identification code.
• visit_comments: Notes on site conditions and work completed.
• DateTime_UTC: Visit start time in UTC.
• sonde_employed: Sensor status following the visit:
• 1: Sensor was removed following the visit.
• 0: Sensor was deployed during the visit.
• NA: No change (sensor remained deployed).
• pre_clean, post_clean, and post_cal values: Values recorded during visits to quantify the effects of cleanings or calibrations.
• cals_performed: List of calibrations performed during the visit.

2022 Data

The data collected in 2022 was processed using the following script: poudre_sonde_network/data_organization/Process_2022.Rmd

Columns in cleaned dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_final: Value for the parameter after passing through QAQC procedures. Erroneous data was removed and, where relevant, raw data was modified to correct for calibration or biofouling errors.
• user_qc_flag: Quality control flag describing why datum was removed or if a data point is flagged/suspect
• modifications: Describes the quality assurance procedures applied to a specific data point, which include:
• calibration correction: Datum corrected due to a calibration error or calibration drift.
  • See Calibration Correction Tool for details
• drift correction: Sensor biofouling on Turbidity or FDOM sensors resulted in drift, which was corrected using linear or exponential functions.
  • See poudre_sonde_network/drift_correction_tool for details
• manual data exclusion: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
  • See poudre_sonde_network/manual_verification_tool for verification tool workflow
• non-resolved drift: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
• NA : No modification of datum.

Columns in raw dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_post_cal_corr: Value after applying calibration corrections using the Calibration Correction Tool
• value_post_ver: Value after removing data identified by manual verification via the Manual Verification Tool. Also documented in user_qc_flag and qc_status.
• value_post_drift_corr: Value after correcting for sensor drift due to biofouling via the Drift Correction Tool. For parameters where drift was not corrected, this matches value_post_ver.
• auto_qc_flag: Flags generated by the auto-qaqc system in the {ross.wq.tools} package. These serve as the basis for comparison with human-verified outcomes.
• qc_status: Final status of a data point:
• OMIT: Datum is considered erroneous and should not be used.
• FLAGGED: Datum may have minor issues; usable for most analyses but should be excluded if a high degree of accuracy is required.
• PASS: Datum is considered valid and accurate based on human verification.
• user_qc_flag: Quality control flag describing why a qc_status was applied.
• modifications: Summary of procedures (calibration correction, drift correction, non-resolved drift, or manual exclusion) applied to the point.

Field Notes (CSU_ROSS_sensor_field_notes_2022.csv)

This file contains field notes collected by technicians throughout the year.

• site: Site identification code.
• visit_comments: Notes on site conditions and work completed.
• DateTime_UTC: Visit start time in UTC.
• sonde_employed: Sensor status following the visit:
• 1: Sensor was removed following the visit.
• 0: Sensor was deployed during the visit.
• NA: No change (sensor remained deployed).
• pre_clean, post_clean, and post_cal values: Values recorded during visits to quantify the effects of cleanings or calibrations.
• cals_performed: List of calibrations performed during the visit.

2019-2021 Data

This data was collected prior to long-term funding from the City of Fort Collins and before rigorous SOPs were developed for data management. As such, no calibration data was collected by technicians, limited field notes were available and data quality is more variable over time. There are frequent large data gaps in this dataset due to sensor malfunctions, biofouling, and other issues.

USE CAUTION when using this dataset for analysis. In particular, optical variables (Turbidity, Chl-a) are highly susceptible to fouling/error and should be used with caution.

The data collected in 2019-2021 was processed using the following script: poudre_sonde_network/data_organization/Process_2019_2021.Rmd

Columns in cleaned dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_final: Value for the parameter after passing through QAQC procedures. Erroneous data was removed and, where relevant, raw data was modified to correct for calibration or biofouling errors.
• user_qc_flag: Quality control flag describing why datum was removed or if a data point is flagged/suspect
• modifications: Describes the quality assurance procedures applied to a specific data point, which include:
• drift correction: Sensor biofouling on Turbidity or FDOM sensors resulted in drift, which was corrected using linear or exponential functions.
  • See poudre_sonde_network/drift_correction_tool for details
• manual data exclusion: Datum identified as erroneous by manual technician verification that could not be resolved by the previous two methods. This datum should not be used in analysis.
  • See poudre_sonde_network/manual_verification_tool for verification tool workflow
• NA : No modification of datum.

Columns in raw dataset:

• DateTime_UTC: Date-time in UTC.
• site: Site ID used to identify the site.
• parameter: Parameter name.
• units: Units for the parameter.
• value_raw: Raw measured value.
• value_post_ver: Value after removing data identified by manual verification via the Manual Verification Tool. Also documented in user_qc_flag and qc_status.
• value_post_drift_corr: Value after correcting for sensor drift due to biofouling via the Drift Correction Tool. For parameters where drift was not corrected, this matches value_post_ver.
• auto_qc_flag: Flags generated by the auto-qaqc system in the {ross.wq.tools} package. These serve as the basis for comparison with human-verified outcomes.
• qc_status: Final status of a data point:
• OMIT: Datum is considered erroneous and should not be used.
• FLAGGED: Datum may have minor issues; usable for most analyses but should be excluded if a high degree of accuracy is required.
• PASS: Datum is considered valid and accurate based on human verification.
• user_qc_flag: Quality control flag describing why a qc_status was applied.
• modifications: Summary of procedures (drift correction, or manual exclusion) applied to the point.

Field Notes (CSU_ROSS_sensor_field_notes_2021.csv)

This file contains field notes collected by technicians throughout the year. NOTE: Field notes were not available in 2019 or 2020.

• site: Site identification code.
• visit_comments: Notes on site conditions and work completed.
• DateTime_UTC: Visit start time in UTC.
• sonde_employed: Sensor status following the visit:
• 1: Sensor was removed following the visit.
• 0: Sensor was deployed during the visit.
• NA: No change (sensor remained deployed).
• pre_clean, post_clean, and post_cal values: Values recorded during visits to quantify the effects of cleanings or calibrations.

Metadata

• CSU_ROSS_location_metadata.csv: Site coordinates, naming conventions, and periods of record. Includes relevant flow gages with discharge records.
• The column grab_sample_site links to the relevant ID in our published grab sample dataset on Zenodo.
• CSU_ROSS_delineated_basins.gpkg: Delineated natural watersheds for each site. These basins were delineated using the USGS NHDPlusV2 dataset and Stream Stats.
• Note: ALL basins have either transbasin diversions, ditches, canals, reservoirs, or urban stormwater systems upstream. Most of these diversions are not captured in the delineated basins. Please use caution when using these basins for analysis.
• Note: For sites which have had housings move (elc, bellvue), the delineated basin is for the most recent location.
• CSU_ROSS_sensor_methods_metadata.csv: Explains the general calibration methods and sensors used for each parameter.
• CSU_ROSS_user_qc_flags.csv: Descriptions of the user_qc_flag flag types used in the raw and cleaned dataset, provides descriptions of the flag, and potential correction pathways.

csu_ross_hydroshare_import.R:

• This script shows a method to import a .zip file downloaded from HydroShare using BagIt. This script organizes the downloaded .zip file and shows users how to access sensor data and spatial metadata.

Additional Resources

• Poudre Water Quality Network: Data collection for long-term monitoring along the Cache la Poudre River through the City of Fort Collins and the primary repository for data cleaning. GitHub | Project Page
• Upper CLP DSS: ROSS work supporting real-time water treatment decision support in the Cache la Poudre River. GitHub
• ROSS CPF Reservoir Study: Investigation of post-fire water quality impacts from the Cameron Peak Fire. GitHub | Project Page
• ross.wq.tools: R package containing functions to support water quality data processing and analysis. GitHub

Most code will not run without proper credentials. Please contact the authors (Sam Struthers, Kathryn Willi, or Juan De La Torre) if you require additional information.