Automated analysis of temperature variance to determine inundation state of wetlands

Abstract

Monitoring the inundation state (wet or dry) of wetlands is critical to understanding aquatic community structure but can be costly and laborintensive. We tested the ability of temperature data from cost-effective iButton data loggers to reflect the inundation state of wetlands in central Missouri, based on our hypothesis that dry ponds would show greater daily temperature variance than ponds that remained inundated with water. We evaluated this method with two experiments in large outdoor mesocosms, and in existing natural wetlands in which we had deployed iButtons. True inundation state from pond visits was compared to predicted inundation state over different temperature variance thresholds expected to delineate wet or dry ponds. We confirmed that the daily temperature variances of dry iButtons were higher than that of iButtons under water, as expected, but that variance was influenced by factors such as canopy cover. We also describe an automated procedure that can be used to determine whether a pond was wet or dry with greater than 80 % accuracy. Using this approach, changes in inundation state, the number of days wet and dry, and the number of drying and filling events can be calculated. Several caveats are also provided that should be considered prior to using this method to maximize the accuracy in assessing inundation state.

Temperature Loggers Capture Intraregional Variation of Inundation Timing for Intermittent Ponds

“…Upon fitting three-state HMMs to the pond-only data set, we found that a threshold between 2.9°C and 3.3°C minimized the number of false dry states for most ponds (Tables S2 and S3). This threshold is slightly lower than that proposed by Anderson et al (2015), who determined that using daily temperature variances cutoffs between 13 and 15 (corresponding to tSDs between 3.6°C and 3.9°C) for the wet state provided the most accurate predictions of pond inundation states in their field experiments. Within our pond-only data set, using a less conservative tSD threshold of 3.5°C decreased the accuracy leading to a false wet state prediction for pond T15U.…”

“…A rapid drop in daily temperature variance can reliably measure the precise timing of an inundation event (Anderson et al, 2015;Arismendi et al, 2017). For example, Anderson et al (2015) tested the ability of temperature sensors to accurately predict inundation states both in natural wetlands and in controlled mesocosms that varied in size and depth. The authors deployed temperature sensors for two 6-month periods in ponds over a 7,140 ha area.…”

Measuring microenvironments for global change: DIY environmental microcontroller units (EMUs)

“…Corrosion of untreated sensors and wire connections, over several weeks, increased failure rates during rain (Supporting Information Section 3.2). For all sensor types, moisture damage presents a problem, sometimes causing high failure rates (Anderson et al., ; Ashcroft & Gollan, ; Lebrija‐Trejos et al., ; Lewkowicz, ). The failure rate in our study (11.4%) was comparable to those of commercial sensors.…”

“…Incorporating measurement capabilities for additional environmental variables such as light or soil moisture increases costs to $500–1000/unit (Table ). In addition, commercial data loggers often fail permanently in field conditions, with reported failure rates of 7%–27% (Anderson et al., ; Ashcroft & Gollan, ; Lebrija‐Trejos, Pérez‐García, Meave, Poorter, & Bongers, ; Lewkowicz, ). Many do not have replaceable batteries or parts or cannot be repaired, thus limiting their life spans.…”

An Open-Source, Durable, and Low-Cost Alternative to Commercially Available Soil Temperature Data Loggers

“…Our field testing has allowed us to identify several areas in which our DIY soil temperature data logger could be improved in future iterations to increase reliability. Commercial data loggers experience failures in the field with reported failure rates of 7–27% per year [ 20 , 36 , 37 , 38 , 39 ]. The majority of our data logger failures were due to leaks.…”