Description: | Daily water temperature (°C) was estimated for each of the three locations (Yolo Bypass, and Sacramento River at Sherwood Harbor and Rio Vista) from a variety of sources. Data integration code can be found at https://github.com/Delta-Stewardship-Council/ybfmp_temperature. The f_get functions bring in the raw data from the data_raw folder and relevant EDI publications. The f_get_hourly and f_gert_WQ_w_fish scripts incorporate Pien et al., 2020 (hourly QC’ed California Data Exchange Center data for Lisbon Weir, LIS, and Rio Vista, RVB and RIV) and Pien and Kwan 2022 (water quality data collected during the Yolo Bypass Fish Monitoring Program’s fish collection). The f_get_RSTR (Yolo Bypass at the rotary screw trap) and f_get_SHWharbor (Sacramento River at Sherwood Harbor) organize and QC the Yolo Bypass Fish Monitoring Program’s (YBFMP) logger data, which are then made into daily summaries in f_make_daily. The imputation and linear regression estimation to fill missing dates was conducted in the f_make_(site name) scripts.
The data in the data_raw folder is continuous water temperature collected by the YBFMP between 1998 and 2019. Data were collected in the Yolo Bypass at the Screw Trap in the Toe Drain (STTD) and at the Sacramento River at Sherwood Harbor (SHR) using a HOBO loggers (Onset Computer Corp "Optic StowAway Temp" loggers). Logger data ranged from daily means (1998) to a fifteen-minute collection interval (2013-2017 for the Yolo Bypass, 2009-2019 for Sherwood Harbor). Missing data occurred regularly due to malfunctions in equipment, shortages of staff, and inability to access sites (see frequency table).
A daily mean, maximum, minimum, standard deviation and coefficient of variation in water temperature was produced as well as columns for sample size (n, number of measurements per day), method (data collection or estimation), category, length (number of consecutive missing dates) and site. Method and category refer to the source or type of estimate used to produce the daily values (see methods table). Within our period of interest (1998-2019) consecutive missing days of data ranged from 693 days (2018/01/31- 2019/12/25 in the Yolo Bypass) to one day. In the Yolo Bypass, when possible missing data was substituted with water temperature measured during fish sampling (method = WQ_w_fish, category = data). Remaining gaps of missing data that were less than or equal to seven consecutive days were imputed in R using the package “imputeTS”, an exponential weighted moving average (method = imputeTS, category = 7&under). When seven or more consecutive days were missing in the Yolo Bypass, first a linear model with the Lisbon Weir station data (Adjusted R-squared = 0.7967) and then a linear model with Rio Vista station data (Adjusted R-squared = 0.8106) was used to estimate water temperature (method = lm_lis and lm_rv, respectively). At Sherwood Harbor a linear model with Rio Vista station data (Adjusted R-squared = 0.9745) was used to estimate Sacramento River water temperature in cases with greater than seven days of consecutive missing dates (method = lm_rv, category = Over7). Despite our best efforts, missing days of data remain for the Yolo Bypass (n = 188) and Sherwood Harbor (n = 310) locations (NA).
==================== Data Sources =========================
DWR-6-SOP-016_v1.1_BeachSeineSampling (https://github.com/AEU-DISE/publish_fish/tree/main/metadata/methods_references), DWR-6-SOP-017_v1.1_FykeTrapSampling (https://github.com/AEU-DISE/publish_fish/tree/main/metadata/methods_references), DWR-6-SOP-018_v1.1_RotaryScrewTrapSampling (https://github.com/AEU-DISE/publish_fish/tree/main/metadata/methods_references), Fish_Publication_QAQC_Workflow_v1.0 (https://github.com/AEU-DISE/publish_fish/tree/main/metadata/methods_references)
=========================================================== |