The Fenwick Journal of Field Science publishes peer-reviewed research, reviews, brief reports, and correspondence on methods and findings in observational field science. It appears quarterly in print and continuously online.
Editorial board
Editor-in-Chief
Josiah Carberry, Fenwick Institute for Field Methods
Deputy Editor
Adaeze Okafor, University of Harrowgate
Associate Editors
Rafael Quintero, Port Amsel Climate Office
Hiroshi Tanaka, Lindqvist Bay Institute of Technology
Fixture notice: this is a fictional journal issue written for Folio's JOURNAL-ISSUE-01 fixture. The journal, people (other than the ORCID example identity Josiah Carberry), institutions, identifiers, data, and references are invented. DOIs use the Crossref test prefix 10.5555 and do not resolve.
Contents
Volume 14 · Issue 4 · December 2026. Page numbers are authored values for this print edition; they were set from the deterministic Chromium render and must be regenerated if the composition changes.
Cities do not switch off at dusk. Streets cool, or fail to; lights come on and change what lives beneath them; traffic thins and the soundscape that remains shapes how people sleep. Yet field science has long been a daytime discipline. Our instruments, our working hours, and our satellite overpasses have all favoured the afternoon.
This special issue gathers work that takes the night seriously as a field setting. The papers share a concern with method: how to observe something faint, intermittent, or easily confounded when the observer cannot see well and the usual reference conditions do not hold.
Abara and colleagues use a before–after control–impact design to ask what happens to nocturnal insects when residential streets switch from sodium to white light-emitting diode lamps. Their design is a model of how to separate an intervention from the seasonal and weather noise that dominates insect counts. Hartono and Mbeki review a quarter-century of methods for measuring night-time noise exposure and find that the field still disagrees about what should be measured, let alone how. Volkov's brief report is a reminder that equipment fails in predictable ways in the cold, and that predictable failures should be reported. Laine's letter asks authors to state the cloud screening criteria they used to define an observation night — a small reform with a large payoff for comparability.
Three themes recur. The first is transparency about observation windows: what counts as a night, which nights were excluded, and why. The second is the relationship between what is measured outdoors and what people experience indoors, where most night-time exposure occurs. The third is equity. In each paper, the places with the least favourable night-time conditions are disproportionately the places where people with the fewest resources live.
We are grateful to the guest editors, Mei-Ling Hartono and Rafael Quintero, to the reviewers who read these papers carefully and quickly, and to the authors who responded to sometimes demanding requests for additional analyses. We hope this issue encourages more field scientists to stay out after dark.
Declarations
Competing interests
J.C. is a co-author of a study cited by one article in this issue and did not handle its review.
Research Articles
Original studies reporting new field observations, with full methods and data availability statements.
Fenwick Journal of Field Science14(4) · December 2026Research Article · e0501
Research Article · Open Access
Street lighting retrofits and nocturnal insect activity on residential streets
Eighteen streets were monitored for two summers before and after a lamp retrofit.
Flying insect counts at lamps fell by about a quarter on retrofitted streets relative to controls.
The reduction was concentrated in moths; beetle and fly counts changed little.
Abstract
Municipalities are replacing sodium street lamps with white light-emitting diode (LED) lamps to save energy. Because many nocturnal insects are attracted to short-wavelength light, the retrofit could increase insect mortality at lamps. We monitored insect activity at lamps on nine retrofitted and nine control residential streets for one summer before and one summer after the retrofit. Relative to controls, counts on retrofitted streets fell by 24% (95% CI 11% to 35%), driven by moths. The lamps installed had a warm colour temperature and full cut-off optics, which may explain why attraction did not increase.
Keywords
artificial light at night
moths
street lighting
BACI design
1 Introduction
Artificial light at night alters the behaviour of many nocturnal insects, drawing them to lamps where they circle until exhausted or are taken by predators (Arden, 2022; Tanaka, 2023). The spectrum of the light matters: lamps that emit more short-wavelength light generally attract more insects (Whitcombe and Laine, 2021).
Across Northmere, councils are replacing low-pressure and high-pressure sodium lamps with white LED lamps. The retrofit is often presented as an environmental gain because it saves energy. Whether it is also a gain or a loss for insects is unclear, since the outcome depends on lamp colour temperature, optics, and mounting height as well as on spectrum (Oyelaran and Abara, 2024).
We took advantage of a staged retrofit in Lindqvist Bay to compare insect activity at lamps before and after replacement, using streets scheduled for later replacement as controls.
2 Methods
2.1 Design and sites
The city's retrofit schedule allocated residential streets to replacement phases by maintenance district, not by ecological characteristics. We selected nine streets from the first phase (impact) and nine from the third phase (control), matched on lamp spacing, mounting height, and adjacent garden cover. Impact streets received 3000 K full cut-off LED lamps between the two survey summers; control streets kept sodium lamps throughout.
2.2 Insect sampling
At three lamps per street, we recorded flying insects within one metre of the lamp head with a low-light camera for ten minutes beginning one hour after sunset, on twelve nights per summer. Nights with rain, wind above 4 m s⁻¹, or air temperature below 10 °C were excluded (Laine, 2026). Insects were classified to order from video by two observers who did not know the street's treatment.
2.3 Analysis
We estimated the retrofit effect as the change in log counts on impact streets minus the change on control streets, as in Equation (1), where Y is the mean log count:
(1)
We fitted negative binomial mixed models with random intercepts for street and night and report the exponentiated interaction as a ratio of ratios with a bootstrap interval (Bexley, 2021).
3 Results
We analysed 1,248 ten-minute recordings and classified 18,402 insects. Moths made up 46% of insects before the retrofit, flies 31%, and beetles 12%. Counts on control streets were similar in both summers.
Figure 1. Estimated retrofit effect by insect order with 95% bootstrap intervals. Values below 1.0 indicate fewer insects at retrofitted lamps relative to controls.
Relative to controls, total counts at retrofitted lamps fell by 24% (ratio of ratios 0.76; 95% CI 0.65 to 0.89). The reduction was largest for moths (0.62; 0.53 to 0.75) and negligible for flies and beetles (Figure 1; Table 1).
Table 1. Mean insects per ten-minute recording.
Order
Impact before
Impact after
Control before
Control after
Moths
7.1
4.6
6.8
7.2
Flies
4.6
4.4
4.8
4.9
Beetles
1.8
1.7
1.9
1.9
Other
1.6
1.5
1.5
1.6
4 Discussion
Retrofitting with warm-white, full cut-off LED lamps was followed by fewer insects at lamps, not more. This contrasts with experiments using cool-white lamps, which have generally increased attraction (Whitcombe and Laine, 2021). We think two features of the installed lamps explain the difference: their low short-wavelength output and their optics, which direct light downward and reduce the visible glare that draws insects from a distance (Tanaka, 2023).
Counts at lamps measure attraction, not population size. Fewer insects at lamps could reflect reduced attraction or a smaller local population, though the stable control counts argue against a general decline. We did not measure insects away from lamps, and two summers cannot capture longer-term population effects.
For practice, our results suggest that the specification of an LED retrofit matters more than the switch to LED itself. Councils that choose warm colour temperatures and full cut-off optics may avoid increasing harm to nocturnal insects and could reduce it.
Declarations
Funding
Lindqvist Bay City Council environmental monitoring fund (fictional award LBCC-EM-24-03).
Competing interests
None declared.
Author contributions
I.A. designed the study and analysed data; T.L.-R. managed field logistics; F.O. classified video and drafted the discussion. All authors approved the final version.
Arden, P. (2022) Field methods in urban climatology: a practical handbook. 2nd edn. Port Amsel: Fenwick Academic Press (fictional).
Bexley, N. (2021) ‘A gentle introduction to crossed random effects’, Statistics for Field Science (fictional), 15, pp. 1–24. Available at: https://doi.org/10.5555/sfs.2021.015001.
Laine, R. (2026) ‘Report the cloud screen: a plea for explicit observation-window criteria’, Fenwick Journal of Field Science, 14(4), e0504. Available at: this issue.
Oyelaran, F. and Abara, I. (2024) ‘Specification, not technology: what determines the ecological footprint of street lighting retrofits in residential neighbourhoods with mixed garden and hard-standing frontages?’, Urban Ecology Notes (fictional), 9, pp. 55–73. Available at: https://doi.org/10.5555/uen.2024.009055.
Tanaka, H. (2023) ‘Glare, spectrum, and the distance from which moths are drawn to lamps’, Remote Sensing of Cities (fictional), 5(2), pp. 101–118.
To describe how night-time noise exposure has been measured in urban field studies and to identify sources of incomparability.
Scope
Field studies published 2000–2025 that measured outdoor or indoor sound levels at night in residential settings.
Findings
Studies used at least nine distinct indicators, four definitions of night, and three classes of instrument. Fewer than one in five reported both indoor and outdoor levels.
Recommendations
Report the night window, the indicator, the instrument class, and microphone placement explicitly; pair outdoor with indoor measurement where possible.
1 Introduction
Night-time noise disturbs sleep, and disturbed sleep is linked to cardiovascular and mental health outcomes.1 Regulators set night-time limits in terms of outdoor indicators, but people sleep indoors, behind walls and windows whose insulation varies widely.2
Field studies of night-time noise have multiplied as sound level meters have become cheaper, but their methods vary so much that results are hard to compare. This review describes that variation and proposes a minimum reporting set.
2 Approach
We searched three bibliographic databases for field studies published between January 2000 and June 2025 that measured sound at night in residential settings. Two reviewers screened titles, abstracts, and full texts and extracted indicators, night definitions, instruments, and microphone placement.3 We included 87 studies.
3 What was measured
The equivalent continuous sound level over the night period was the most common indicator, but studies also reported maximum levels, the number of events above a threshold, percentile levels, and intermittency ratios. Table 1 summarises the indicator families.
Table 1. Night-time noise indicator families in 87 field studies, with typical night definitions and instrument classes.
Indicator family
Examples
Studies
Typical night window
Instrument class
Main limitation
Equivalent level
LAeq,night, Lnight
71
23:00–07:00
Class 1 or 2
Hides short loud events
Maximum level
LAFmax
38
23:00–07:00
Class 1
Sensitive to single events
Event counts
Number above 60 dB
22
22:00–06:00
Class 1
Threshold choice varies
Percentile levels
LA90, LA10
19
Varies
Class 2 or low-cost
Background versus peaks
Intermittency
Intermittency ratio
9
23:00–07:00
Class 1
Few validated thresholds
Indoor levels
Bedroom LAeq
15
Occupant bedtime
Class 2 or low-cost
Window position unknown
Spectral content
Low-frequency share
6
23:00–07:00
Class 1
Rarely comparable
Note: studies could report more than one indicator family.
Definitions of night varied as much as indicators. Most studies used a fixed clock window of eight hours, but some used occupants' reported bedtimes, and a few used civil twilight.4 Clock windows are convenient for regulation; bedtime windows match exposure better but cannot be compared across households without adjustment.
4 Where and how it was measured
Outdoor microphones were mounted at façades, at free-field positions, or on street furniture, and placement was often reported incompletely. Façade measurements are several decibels higher than free-field measurements at the same location, so unreported placement alone can explain apparent differences between studies.5
Only 15 studies measured inside bedrooms, and only 12 measured inside and outside at the same dwelling. Those paired studies found indoor-outdoor differences ranging from 10 to 32 dB, depending on window position and glazing.6
Low-cost sensors appeared after 2015 and now feature in a quarter of studies. Their accuracy at the low levels typical of quiet nights is limited, and few studies reported calibration against a reference instrument.
5 Recommendations
We propose that every night-time noise field study report, at minimum: the night window and how it was defined; the indicator or indicators; the instrument class and calibration; microphone placement relative to façades; and whether windows were open or closed for indoor measurements. Paired indoor and outdoor measurement should be the default wherever access allows.
Declarations
Funding
None.
Competing interests
M.-L.H. was guest editor of this special issue; the review was handled independently by the Deputy Editor.
Author contributions
Both authors screened and extracted studies and wrote the review.
Notes
Dorian Larkspur, “Heat Health Warning Thresholds and Their Night-Time Components,” Public Health Field Reports (fictional) 22 (2024): 60–79.
Winter deployments of low-cost loggers lose data when batteries fail on cold nights. In a climate-chamber test of 24 loggers with three battery chemistries, alkaline cells failed after a median of 9 nights at −10 °C, while lithium iron disulfide cells lasted the full 60-night test. Field networks operating below freezing should specify lithium cells and report battery chemistry.
1 Method
Twenty-four loggers of one model were fitted with alkaline, nickel-metal hydride, or lithium iron disulfide AA cells (eight each) and held in a climate chamber programmed with a nightly cycle from +5 °C by day to −10 °C by night, logging at ten-minute intervals [1]. A logger was counted as failed on the first night it recorded fewer than 90% of expected readings.
2 Results
All alkaline-powered loggers failed within 14 nights (median 9). Nickel-metal hydride loggers failed after a median of 23 nights. No lithium-powered logger failed during the 60-night test (Figure 1). Failures were preceded by two to three nights of missing readings concentrated between 03:00 and 06:00, the coldest part of the cycle, which would bias nightly minima upward if partial nights were retained [2].
Figure 1. Share of loggers still recording by battery chemistry, eight loggers per group. Line patterns and direct labels identify the series.
3 Implications
Networks deployed through winter should use lithium cells, and studies should report battery chemistry and the rule used to exclude partial nights [3].
Declarations
Funding and competing interests
Internal laboratory funds. None declared.
References
[1] Abernathy-Svensson K. Shield design for passive temperature sensors. Instrum Notes (fictional). 2020;41:12–18.
[2] Okafor A, Hartono M-L. Low-cost temperature loggers for street-scale microclimate monitoring: drift, shielding, and siting errors. Lindqvist Bay Tech Rep. 2023;17:1–38. doi:10.5555/lbit.tr.2023.17
[3] Laine R. Report the cloud screen: a plea for explicit observation-window criteria. Fenwick J Field Sci. 2026;14(4):e0504.
Fenwick Journal of Field Science14(4) · December 2026Letter · e0504
Letter
Report the cloud screen: a plea for explicit observation-window criteria
Sir — Night-time field studies routinely exclude cloudy or windy nights, for good reason: radiative processes dominate only under clear, calm skies. But the criteria used to define a qualifying night are often unstated, or stated so loosely (“clear nights were selected”) that they cannot be reproduced.
The choice matters. A threshold of two oktas rather than four can halve the number of qualifying nights in a maritime climate and shift the sample toward colder, drier air masses. Two studies of the same phenomenon in the same city can reach different conclusions simply because they screened nights differently.
I suggest that authors state, at minimum: the cloud and wind thresholds, the data source used to apply them, the time window over which they were applied, and the number of nights excluded. The papers by Abara and colleagues and by Volkov in this issue show how little space this takes.
Rosalind Laine
Information for authors
The journal considers research articles (up to 6,000 words), reviews (up to 8,000 words), brief reports (up to 1,500 words), and letters (up to 600 words). Manuscripts are submitted as semantic HTML or as a word-processor file with structured headings; the production team converts accepted manuscripts to the journal's HTML publication format.
Authors supply reference lists formatted in the journal's style. The journal does not validate bibliographic data; authors are responsible for the accuracy of every reference, identifier, and link.
Figures should be supplied as vector files where possible, with a text description for each. Colour must not be the only means of conveying information, because the print edition is grey-scale.
Every article includes funding, competing-interest, author-contribution, and data-availability statements written by the authors. The editors do not infer these statements from the manuscript.
Corrigendum
In Okafor et al., “Canopy cover and nocturnal surface cooling in mid-latitude residential blocks,” Fenwick J Field Sci 14(3): e0412, the confidence interval for the canopy–sky view interaction in Section 3.2 was printed as 0.08 to 0.40 in the online-first version; the correct interval, 0.18 to 0.40, appears in the version of record. This fictional notice exercises corrigendum composition in the fixture.
Northmere Field Methods Summer School 2027
Two weeks of practical training in sensor networks, field statistics, and open data. Applications open 1 February 2027.