Water Quality at Sawyer Mill Reservoir on Bellamy River, Dover, NH

NOTE: This data was posted to meet the requirements of the EPA Low Cost Nutrient Sensor Challenge (Phase 2). The data set includes the results of a preliminary test of one type of low cost nutrient sensor and should not be used without consultation with the authors.
Dams and their reservoirs are increasingly being removed from the landscape, often because they are aging and would need costly repairs, have no significant utility and/or to improve anadromous fish passage and connectivity with spawning areas. Understanding the role of reservoirs in nitrate removal will inform ongoing decisions regarding dam removal.Because reservoirs created by dams are potentially effective at removing nitrogen (Gold et al. 2016), such dam removals come with tradeoffs, including reduced nitrate removal. Yet we have a poor understanding of the effectiveness of the reservoirs on smaller rivers that are common in much of New England and elsewhere, as well as how their effectiveness varies during different parts of the growing season and during storm events within season.

Our overarching approach used high frequency nitrate sensors to characterize nitrate concentration patterns and fluxes in different kinds of streams and rivers that drain into and out of reservoirs to understand variability in water quality. From these measurement we can also quantify the effectiveness of reservoirs to retain nitrate across a range of flow conditions. In order to help interpret these nitrate results, we also deployed ancillary high frequency sensors that measure specific conductance and water stage/discharge.

This dataset contains quality controlled level (Level 1) data for all of the variables measured for the EPA Nutrient Sensor Action Challenge. Individual file contain specific variables from data Collected in 2018. We implemented a simple workflow to develop usable datasets (Levels 0 and Level 1, Table 2), . Data was processed using custom Matlab code (Mathworks Inc., Natick, MA), and MS Excel. Unprocessed raw data (Level 0), consisting of multiple data streams at native measurement resolution, were compiled on an ongoing basis.

Low-cost high frequency nitrate sensor outputs data every 6 seconds. While, the other high frequency nitrate sensor (SUNA) output 16 frames of data every 15 minutes. Other sensors ( stage height, conductivity, temperature) provided data at 15-minute intervals. Grab sample (nutrients and chloride), and hand-held sensor measurements ( conductance, temperature, and dissolved oxygen) data was collected weekly or biweekly, in addition to periodic flow measurement data. Level 0 data consists raw sensor files, without an processing performed. For the next level of processing, outliers (Level 1), and bad data points were identified and removed based on existing or historic data, stage height was transformed to discharge by applying site-specific rating curve equation, and temporal aggregation performed and each site’s data was compiled into one CSV file.

Each file header contains site location and an explanation of variable names.

注：本数据集发布旨在满足美国环境保护署（EPA）低成本营养素传感器挑战赛（第二阶段）的要求。该数据集包含了一种低成本营养素传感器初步测试的结果，未经原作者咨询，不得使用。随着景观中水坝及其水库的逐渐拆除，这通常是因为它们老化且需要昂贵的维修费用，没有显著的功效，或为了改善溯河洄游鱼类通道与产卵区的连通性。了解水库在硝酸盐去除中的作用将有助于对水坝拆除的持续决策。由于由水坝形成的水库在去除氮素方面可能具有潜在的有效性（Gold等人，2016年），此类水坝拆除伴随着权衡，包括硝酸盐去除能力的降低。然而，我们对水库在常见于新英格兰大部分地区及其他地区的较小河流中的有效性，以及它们在生长季节不同阶段和季节内的风暴事件中的有效性变化，理解甚微。

我们的总体方法利用高频硝酸盐传感器来表征不同类型溪流和河流中硝酸盐浓度模式和通量，这些溪流和河流流入和流出水库，以了解水质变异性。从这些测量中，我们还可以量化水库在多种流量条件下保留硝酸盐的有效性。为了帮助解释这些硝酸盐结果，我们还部署了辅助的高频传感器，以测量特定电导率和水位/流量。

本数据集包含EPA营养素传感器行动挑战赛所测变量的质量控制级别（Level 1）数据。每个文件包含2018年收集的具体变量。我们实施了一个简单的流程来开发可用的数据集（Level 0和Level 1，表2）。数据使用定制Matlab代码（Mathworks Inc.，Natick，MA）和MS Excel进行处理。未经处理的原始数据（Level 0），包括在原生测量分辨率下的多个数据流，持续汇编。

低成本高频硝酸盐传感器的数据输出间隔为每6秒一次。而另一种高频硝酸盐传感器（SUNA）每15分钟输出16帧数据。其他传感器（水位高度、电导率、温度）的数据输出间隔为15分钟。每周或每两周收集抓取样本（营养素和氯）和手持式传感器测量数据（电导率、温度和溶解氧），此外还收集定期的流量测量数据。Level 0数据由原始传感器文件组成，未进行任何处理。在下一级处理中，基于现有或历史数据，识别并移除了异常值（Level 1）和坏数据点，通过应用特定地点的比率曲线方程将水位高度转换为流量，并进行了时间聚合，每个地点的数据被汇编成一个CSV文件。

每个文件标题包含地点位置和变量名称的解释。