Search for: All records

Modeling Irregularly-sampled and Multivariate Time Series (IMTS) is crucial across a variety of applications where different sets of variates may be missing at different time-steps due to sensor malfunctions or high data acquisition costs. Existing approaches for IMTS either consider a two-stage impute-then-model framework or involve specialized architectures specific to a particular model and task. We perform a series of experiments to derive insights about the performance of IMTS methods on a variety of semi-synthetic and real-world datasets for both classification and forecasting. We also introduce Missing Feature-aware Time Series Modeling (MissTSM) or MissTSM, a simple model-agnostic and imputation-free approach for IMTS modeling. We show that MissTSM shows competitive performance compared to other IMTS approaches, especially when the amount of missing values is large and the data lacks simplistic periodic structures–conditions common to real-world IMTS applications. 

Abstract Lakes and reservoirs play a key role in the global carbon cycle, representing important carbon sinks and sources within the terrestrial landscape under different environmental conditions. Changes in climate and land use have led to increased air and surface water temperatures; increased occurrence and duration of hypolimnetic anoxia; and altered hydrology and nutrient loading, which have the potential to affect how these freshwater ecosystems receive and process carbon. To assess how interacting environmental drivers influence carbon cycling in lakes and reservoirs, we used a 5-year whole-ecosystem experiment to investigate the effects of variable catchment, meteorology, and in-lake drivers on epilimnetic and hypolimnetic dissolved organic carbon in a small reservoir. Using a combination of whole-ecosystem models and time-series analyses, we found that primary production and other internal sources contributed a mean of 29% (range: 7–49%) of the dissolved organic carbon in the reservoir’s epilimnion over the 5-year period. We also found that sinking epilimnetic primary production, dissolved organic carbon from the sediments, and other factors were likely important sources of hypolimnetic dissolved organic carbon, especially during periods of anoxia. Both the epilimnion and hypolimnion were found to be intermittent sinks, yet net sources, of dissolved organic carbon. Overall, water temperature was identified as the most important environmental predictor for water-column dissolved organic carbon, with higher concentrations observed under seasonally elevated temperatures during the late summer and early fall. Our results suggest that lakes and reservoirs may become larger sources of dissolved organic carbon to downstream ecosystems in a warmer, more anoxic future. 

Crustacean and rotifer density and biomass were measured from 2014 to 2025 in five drinking water reservoirs in southwestern Virginia, USA. These reservoirs are: Beaverdam Reservoir (Vinton, Virginia), Falling Creek Reservoir (Vinton, Virginia), Carvins Cove Reservoir (Roanoke, Virginia), Gatewood Reservoir (Pulaski, Virginia), and Spring Hollow Reservoir (Salem, Virginia). Beaverdam, Falling Creek, Carvins Cove, and Spring Hollow Reservoirs are owned and operated by the Western Virginia Water Authority as primary or secondary drinking water sources for Roanoke, Virginia, and Gatewood Reservoir is a drinking water source for the Town of Pulaski, Virginia. The dataset consists of integrated vertical tow samples from the whole water column, just the epilimnion, and just the hypolimnion (as the difference between the full water column and epilimnion tows), as well as discrete depth measurements collected with a Schindler trap. Most samples were collected at the deepest site of each reservoir adjacent to the dam. Sampling frequency and duration varied among reservoirs and years and included weekly to monthly routine monitoring as well as intensive 24-hour sampling campaigns. In 2014-2016, zooplankton samples were collected approximately fortnightly in the spring, summer, and autumn months at Beaverdam Reservoir, Carvins Cove Reservoir, and Gatewood Reservoirs. Falling Creek Reservoir samples were collected weekly to monthly in spring and summer 2014, and Spring Hollow Reservoir samples were collected approximately fortnightly in the spring, summer, and autumn months of 2015 and 2016. In 2019, zooplankton samples were collected approximately weekly to monthly from April to November at Beaverdam Reservoir and April to September at Falling Creek Reservoir. In 2020, zooplankton samples were collected approximately weekly to monthly from May to December at Beaverdam Reservoir and June to September at Falling Creek Reservoir. In 2021, zooplankton were collected monthly from March to December in Beaverdam Reservoir. In 2022, zooplankton were collected monthly from January to May at Beaverdam Reservoir. In 2023-2025, zooplankton were collected approximately monthly from March or April to December in Beaverdam Reservoir. Falling Creek Reservoir zooplankton samples were sparsely collected during 2021 to 2025. During the 24-hour sampling campaigns conducted in Beaverdam Reservoir from 2019-2022, samples were collected from both the deepest pelagic site and a shallow littoral site. 

Crustacean and rotifer density and biomass were measured from 2014 to 2022 in five drinking water reservoirs in southwestern Virginia, USA. These reservoirs are: Beaverdam Reservoir (Vinton, Virginia), Falling Creek Reservoir (Vinton, Virginia), Carvins Cove Reservoir (Roanoke, Virginia), Gatewood Reservoir (Pulaski, Virginia), and Spring Hollow Reservoir (Salem, Virginia). Beaverdam, Falling Creek, Carvins Cove, and Spring Hollow Reservoirs are owned and operated by the Western Virginia Water Authority as primary or secondary drinking water sources for Roanoke, Virginia, and Gatewood Reservoir is a drinking water source for the Town of Pulaski, Virginia. The dataset consists of integrated vertical tow samples from the whole water column, just the epilimnion, and just the hypolimnion (as the difference between the full water column and epilimnion tows), as well as discrete depth measurements collected with a Schindler trap. Most samples were collected at the deepest site of each reservoir adjacent to the dam. Sampling frequency and duration varied among reservoirs and years and included weekly to monthly routine monitoring as well as intensive 24-hour sampling campaigns. In 2014-2016, zooplankton samples were collected approximately fortnightly in the spring, summer, and autumn months at Beaverdam Reservoir, Carvins Cove Reservoir, and Gatewood Reservoirs. Falling Creek Reservoir samples were collected weekly to monthly in spring and summer 2014, and Spring Hollow Reservoir samples were collected approximately fortnightly in the spring, summer, and autumn months of 2015 and 2016. In 2019, zooplankton samples were collected approximately weekly to monthly from April to November at Beaverdam Reservoir and April to September at Falling Creek Reservoir. In 2020, zooplankton samples were collected approximately weekly to monthly from May to December at Beaverdam Reservoir and June to September at Falling Creek Reservoir. In 2021 and 2022, zooplankton were collected monthly from March to December in 2021 and January to May in 2022 at Beaverdam Reservoir. Falling Creek Reservoir zooplankton samples in 2021 and 2022 were sparsely collected. During the 24-hour sampling campaigns conducted in Beaverdam Reservoir from 2019-2022, samples were collected from both the deepest pelagic site and a shallow littoral site. 

Metal and nutrient loads were calculated from 2019-2024 from the inflow stream to Falling Creek Reservoir (FCR), a drinking water reservoir located in Vinton, Virginia, USA. The reservoir is owned and operated by the Western Virginia Water Authority and is managed as a secondary drinking-water source for the city of Roanoke, VA. Only Fe, Mn, and nutrients (TN and TP) were analyzed and calculated in 2019. The full suite of metals (Li, Na, Mg, Al, K, Ca, Fe, Mn, Cu, Sr, Ba) and nutrients were analyzed from 2020-2024. The loads that were collected using an ISCO automated sampler located at the main inflow tributary to FCR. Sampling frequency was approximately fortnightly from spring to fall (March - November). Load calculations were performed using the calculated cumulative flow over the sampling period from the ISCO and the analyzed total metal and nutrient concentrations. Please note we are publishing this data package before the nutrient samples have been analyzed, but will be included in later versions. 

Discrete depth profiles of water temperature, dissolved oxygen, oxidation-reduction potential, conductivity, specific conductance, and pH were collected with multiple handheld water quality probes and discrete depth profiles of photosynthetically active radiation (PAR) were collected with a LI-COR underwater light meter from 2013 to 2024 in five drinking water reservoirs in southwestern Virginia, USA. These reservoirs are: Beaverdam Reservoir (Vinton, Virginia), Carvins Cove Reservoir (Roanoke, Virginia), Falling Creek Reservoir (Vinton, Virginia), Gatewood Reservoir (Pulaski, Virginia), and Spring Hollow Reservoir (Salem, Virginia). Beaverdam, Carvins Cove, Falling Creek, and Spring Hollow Reservoirs are owned and operated by the Western Virginia Water Authority as primary or secondary drinking water sources for Roanoke, Virginia, and Gatewood Reservoir is a drinking water source for the Town of Pulaski, Virginia. All discrete depth profiles were collected on approximately 1-meter intervals. The data package consists of two datasets: 1) Secchi depth data; and 2) discrete depth profiles of multiple water quality variables measured by handheld sensors. The Secchi data and discrete depth profiles were measured at the deepest site of each reservoir adjacent to the dam, as well as other in-reservoir sites. Handheld sensor measurements were also collected at a gauged weir on the primary inflow tributary, other inflows, and outflows at Falling Creek Reservoir; inflows and outflows at Beaverdam Reservoir; and inflows at Carvins Cove Reservoir. In 2021, YSI handheld data were also collected from a littoral site in Beaverdam Reservoir. Data were collected approximately fortnightly in the spring months (March - May), weekly in the summer and early autumn (June - September), and monthly in the late autumn and winter (October - February) in Falling Creek and Beaverdam Reservoirs; data coverage in the other three reservoirs varies among years. Note there are some YSI depth profiles and Secchi observations that were measured at night during overnight sampling. All of these observations have the correct time associated with them. There was a major revision of this dataset from its previous version, which included correcting times of observations, including values of negative ORP, and adding observations from an outflow at Falling Creek Reservoir. 

Discrete depth profiles of water temperature, dissolved oxygen, oxidation-reduction potential, conductivity, specific conductance, and pH were collected with various handheld YSI water quality probes and discrete depth profiles of photosynthetically active radiation (PAR) were collected with a LI-COR underwater light meter from 2013 to 2023 in five drinking water reservoirs in southwestern Virginia, USA. These reservoirs are: Beaverdam Reservoir (Vinton, Virginia), Carvins Cove Reservoir (Roanoke, Virginia), Falling Creek Reservoir (Vinton, Virginia), Gatewood Reservoir (Pulaski, Virginia), and Spring Hollow Reservoir (Salem, Virginia). Beaverdam, Carvins Cove, Falling Creek, and Spring Hollow Reservoirs are owned and operated by the Western Virginia Water Authority as primary or secondary drinking water sources for Roanoke, Virginia, and Gatewood Reservoir is a drinking water source for the Town of Pulaski, Virginia. All YSI and PAR depth profiles were collected on approximately 1-meter intervals. The data package consists of two datasets: 1) Secchi depth data; and 2) YSI and PAR profiles. The Secchi depths and YSI and PAR depth profiles were measured at the deepest site of each reservoir adjacent to the dam and other in-reservoir transects. YSI measurements were also collected at a gauged weir on the primary inflow tributary, other inflows and outflows at Falling Creek, inflows and outflows at Beaverdam, and inflows at Carvins Cove. In 2021, YSI profiles were also collected from a littoral site in Beaverdam. Data were collected approximately fortnightly in the spring months (March - May), weekly in the summer and early autumn (June - September), and monthly in the late autumn and winter (October - February) in Falling Creek and Beaverdam Reservoirs; data coverage in the other three reservoirs varies among years. Note there are some YSI profiles and Secchi observations that were measured at night during overnight sampling. All of these observations have the correct time associated with them. There was a major revision of this dataset from version 11, which includes correcting times of observations, including negative ORP values, and adding observations from an outflow at Falling Creek Reservoir. 

Abstract Zooplankton play an integral role as indicators of water quality in freshwater ecosystems, but exhibit substantial variability in their density and community composition over space and time. This variability in zooplankton community structure may be driven by multiple factors, including taxon-specific migration behavior in response to environmental conditions. Many studies have highlighted substantial variability in zooplankton communities across spatial and temporal scales, but the relative importance of space vs. time in structuring zooplankton community dynamics is less understood. In this study, we quantified spatial (a littoral vs. a pelagic site) and temporal (hours to years) variability in zooplankton community structure in a eutrophic reservoir in southwestern Virginia, USA. We found that zooplankton community structure was more variable among sampling dates over 3 years than among sites or hours of the day, which was associated with differences in water temperature, chlorophyll a, and nutrient concentrations. Additionally, we observed high variability in zooplankton migration behavior, though a slightly greater magnitude of DHM vs. DVM during each sampling date, likely due to changing environmental conditions. Ultimately, our work underscores the need to continually integrate spatial and temporal monitoring to understand patterns of zooplankton community structure and behavior in freshwater ecosystems. 

Abstract Ecosystems around the globe are experiencing changes in both the magnitude and fluctuations of environmental conditions due to land use and climate change. In response, ecologists are increasingly using near‐term, iterative ecological forecasts to predict how ecosystems will change in the future. To date, many near‐term, iterative forecasting systems have been developed using high temporal frequency (minute to hourly resolution) data streams for assimilation. However, this approach may be cost‐prohibitive or impossible for forecasting ecological variables that lack high‐frequency sensors or have high data latency (i.e., a delay before data are available for modeling after collection). To explore the effects of data assimilation frequency on forecast skill, we developed water temperature forecasts for a eutrophic drinking water reservoir and conducted data assimilation experiments by selectively withholding observations to examine the effect of data availability on forecast accuracy. We used in situ sensors, manually collected data, and a calibrated water quality ecosystem model driven by forecasted weather data to generate future water temperature forecasts using Forecasting Lake and Reservoir Ecosystems (FLARE), an open source water quality forecasting system. We tested the effect of daily, weekly, fortnightly, and monthly data assimilation on the skill of 1‐ to 35‐day‐ahead water temperature forecasts. We found that forecast skill varied depending on the season, forecast horizon, depth, and data assimilation frequency, but overall forecast performance was high, with a mean 1‐day‐ahead forecast root mean square error (RMSE) of 0.81°C, mean 7‐day RMSE of 1.15°C, and mean 35‐day RMSE of 1.94°C. Aggregated across the year, daily data assimilation yielded the most skillful forecasts at 1‐ to 7‐day‐ahead horizons, but weekly data assimilation resulted in the most skillful forecasts at 8‐ to 35‐day‐ahead horizons. Within a year, forecasts with weekly data assimilation consistently outperformed forecasts with daily data assimilation after the 8‐day forecast horizon during mixed spring/autumn periods and 5‐ to 14‐day‐ahead horizons during the summer‐stratified period, depending on depth. Our results suggest that lower frequency data (i.e., weekly) may be adequate for developing accurate forecasts in some applications, further enabling the development of forecasts broadly across ecosystems and ecological variables without high‐frequency sensor data. 

We monitored water level and water quality in Beaverdam Reservoir (Vinton, Virginia, USA, 37.31288, -79.8159) with visual observations and high-frequency (10-minute and 15-minute) sensors in 2009-2023. All variables were measured at the deepest site of the reservoir adjacent to the dam. Beaverdam Reservoir is owned and managed by the Western Virginia Water Authority as a secondary drinking water source for Roanoke, Virginia. This data package is comprised of three datasets: 1) BVR_WaterLevel_2009_2023.csv, 2) BVRSensorString_2016_2020.csv, and 3) BVRPlatform_2020_2023.csv. 1) BVR_WaterLevel_2009_2023.csv contains water level observations of the staff gauge by both the Western Virginia Water Authority and the Virginia Tech Reservoir Group LTREB field crew. This dataset spans 2009 to 2023, with data collection still ongoing. 2) BVRSensorString_2016_2020.csv consists of a water temperature profile at ~1-meter intervals from the surface of the reservoir to 10.5 m below the water, complemented by a dissolved oxygen logger at 5 m or 10 m, depending on the time of year. A sonde measuring water temperature, conductivity, specific conductance, chlorophyll a, phycocyanin, total dissolved solids, dissolved oxygen, fluorescent dissolved organic matter, and turbidity was additionally deployed at ~1.5 m depth. This dataset spans 2016 to 2020, with no additional data collection beyond the last observation. The third dataset is BVRPlatform_2020_2023.csv, with data collection still ongoing. This dataset contains: a) a temperature string with 13 temperature sensors deployed ~1 m apart from the surface to 0.5 m above the sediments of the reservoir; b) two dissolved oxygen sensors, one in the middle of the string and one sensor above the sediments; and c) a pressure sensor just above the sediments. The same sonde from the first 2016-2020 dataset is also included in this 2020-2023 dataset, deployed at 1.5 m below the surface. The sensors on the temperature string (thermistors, dissolved oxygen sensors, and pressure sensor) are permanently fixed to the platform and do not change with the water level. In the methods, we describe how to add a depth measurement to each observation.