Fenwick Journal of Field Science Volume 14 · Issue 3 · September 2026 Article e0412 · pp. 201–208

Research Article · Open Access

Canopy cover and nocturnal surface cooling in mid-latitude residential blocks

Evidence from a three-city low-cost sensor network, 2023–2025

  1. Adaeze Okafor1,2,*
  2. Tomás Lindgren-Reyes1
  3. Josiah Carberry3 ORCID 0000-0002-1825-0097
  4. Mei-Ling Hartono2
  5. Rafael Quintero2,4
  1. 1 Department of Geography, University of Harrowgate, Harrowgate, Northmere
  2. 2 Centre for Urban Microclimate, Lindqvist Bay Institute of Technology, Lindqvist Bay, Northmere
  3. 3 Fenwick Institute for Field Methods, Port Amsel, Northmere
  4. 4 Port Amsel Climate Office, Port Amsel, Northmere

* Corresponding author: Adaeze Okafor, a.okafor@example.org, Department of Geography, University of Harrowgate, 14 Moor Lane, Harrowgate HG2 7QT.

DOI
https://doi.org/10.5555/fjfs.2026.1403.e0412
Canonical URL
https://example.org/fjfs/14/3/e0412
Article ID
e0412
Running title
Canopy cover and night cooling
History
Received · Revised · Accepted · Published
Cite as
Okafor A, Lindgren-Reyes T, Carberry J, Hartono M-L, Quintero R. Canopy cover and nocturnal surface cooling in mid-latitude residential blocks. Fenwick J Field Sci. 2026;14(3):e0412.
Licence
© 2026 The Authors. Published under the Creative Commons Attribution 4.0 International licence (CC BY 4.0).

Abstract

Background

Night-time heat drives much of the health burden of urban heat, yet most street-scale studies report afternoon surface temperatures. Whether residential tree canopy speeds or slows nocturnal surface cooling remains contested, because canopy both shades surfaces by day and traps long-wave radiation by night.

Methods

We deployed 312 shielded low-cost loggers across 48 residential blocks in three mid-latitude cities and recorded surface and air temperature at ten-minute resolution for three summers. We modelled block-level cooling rate between sunset and 04:00 with crossed random effects for block and night, adjusting for impervious cover, sky view factor, wind speed, and cloud cover.

Results

Across 3,512 clear, calm block-nights, each ten-percentage-point increase in canopy cover was associated with a 0.061 °C h⁻¹ faster cooling rate (95% CI 0.048 to 0.074). The association weakened above 45% canopy and reversed under canopy with low sky view factor.

Conclusions

Moderate residential canopy accelerates night cooling in these cities; dense, enclosed canopy can retain heat. Planting guidance should consider sky view as well as canopy fraction.

Keywords
  • urban heat
  • nocturnal cooling
  • tree canopy
  • sky view factor
  • low-cost sensors
  • mixed-effects models

1 Introduction

Heat-related illness in temperate cities is increasingly a night-time problem. When minimum temperatures stay high, bodies cannot recover from daytime exposure, and the risk of hospital admission rises sharply among older residents and people with chronic disease [5], [27]. Yet the evidence base for neighbourhood-scale heat adaptation is dominated by afternoon measurements, typically surface temperature from satellite overpasses or single walking transects [1].

Street trees are the most widely promoted heat adaptation for residential areas. By day their benefits are well established: shade lowers surface temperature directly and transpiration cools the surrounding air [6], [22]. By night the picture is less clear. Canopy that shades pavement by day also obstructs the sky at night, slowing long-wave radiative loss from the surfaces below [9]. Dense canopy may also reduce wind speed near the ground and suppress the mixing that removes warm air from street canyons [20].

Previous field studies of this trade-off disagree. Studies conducted in open suburban settings have tended to report faster cooling under moderate canopy, while studies in compact neighbourhoods have reported slower cooling under dense canopy [1]. Most were short campaigns of a few nights in one city, with too few sites to separate canopy from the correlated effects of building density, impervious cover, and sky viewa.

Low-cost temperature loggers now make it practical to observe many blocks for many nights [2], [3]. Their known weaknesses — calibration drift, radiation errors from poor shielding, and inconsistent siting — can be controlled with intercalibration, standard shields, and documented siting rules [13], [19].

Here we use a three-summer, three-city network to ask two questions. First, is residential canopy cover associated with faster or slower nocturnal surface cooling once impervious cover and weather are accounted for? Second, does the association depend on how enclosed the canopy is, as measured by sky view factor? Our analysis plan was pre-registered before the second summer of data collection [26].

2 Methods

2.1 Study cities and block selection

Harrowgate, Lindqvist Bay, and Port Amsel are mid-sized cities in the maritime-temperate climate zone of Northmere, with summer mean daily maxima between 21 and 24 °C [16]. In each city we drew sixteen residential blocks from the municipal parcel register, stratified by canopy cover tertile and building height class, so that canopy cover varied as independently of building density as the housing stock allowed.

A block was eligible if it contained at least twenty dwellings, had continuous public footway on both sides, and had not undergone major construction in the previous two years. Blocks adjacent to parks larger than one hectare, water bodies, or major roads were excluded to limit advective effects. Figure 1 shows the sampling design.

Sampling design Three city panels, each containing a four by four grid of blocks. Rows are canopy tertiles from low to high; columns are building height classes. Every cell holds one block with a sensor marker. Harrowgate Lindqvist Bay Port Amsel Columns: building height class 1–4 Rows: canopy tertile, low (top) to high (bottom) Markers: four example blocks per city
Figure 1. Stratified block sampling design in the three study cities. Each panel crosses building height class with canopy cover tertile; one block was drawn per cell. Source: authors; block boundaries from municipal parcel registers [10].

2.2 Instruments and measurement protocol

Each block received six or seven loggers: four surface-contact sensors on footway paving and two or three air-temperature sensors at 2.5 m on lamp columns, in louvred radiation shields [13]. All loggers were intercalibrated in a climate chamber before each summer and checked against a reference thermometer at the end of the season; per-sensor offsets and slopes are reported in Table A1 [3].

Canopy cover was calculated from a 0.5 m lidar canopy height model as the share of block area under vegetation taller than 2 m [11]. For block B with pixels p and canopy height hp, canopy fraction is given by Equation (1):

fc = 1|B| ∑p∈B 1 [hp>2 m]
(1)

Sky view factor was estimated from hemispherical images taken at each surface sensor at midsummer [4]. Impervious cover was taken from the municipal land-cover maps, and hourly wind speed and cloud cover from the nearest synoptic station.

2.3 Outcome and statistical model

For each block i and night t, the cooling rate Cit is the mean decline in surface temperature per hour between local sunset and 04:00, averaged over the block's surface sensors (Equation 2). Nights were retained when cloud cover stayed below three oktas and mean wind speed below 3 m s⁻¹, conditions under which radiative cooling dominates [14].

Cit = T‾i(tsunset)−T‾i(t04:00) t04:00−tsunset
(2)

We fitted linear mixed models with crossed random intercepts for block and night [8], [30], as specified in Equation (3), where I is impervious fraction, S is sky view factor, W is wind speed, and u, v, and ε are block, night, and residual errors:

Cit = β0+ β1fc,i+ β2Ii+ β3Si + β4fc,iSi+ β5Wt+ ui+vt+εit
(3)

Inference used cluster-robust standard errors by block [21]. We checked residual spatial autocorrelation with Moran's I on block centroids [17]. All processing code is open [19]; the model specification in Listing 1 is reproduced from the pre-registration [26].

Listing 1. Pre-registered model specification.
cooling ~ canopy * svf
  + impervious + wind
  + (1 | block) + (1 | night)
weights: n_sensors
nights: cloud < 3 okta,
        wind < 3 m/s
Cooling rate against canopy cover by sky view class Line chart. Cooling rate rises from about 0.22 to 0.50 degrees per hour as canopy increases from 5 to 45 percent in open blocks, and flattens or falls above 45 percent in enclosed blocks. Values are listed in Table 1. 0.10.20.30.40.5 10305070 Canopy cover (%) Cooling rate (°C h⁻¹) Open (sky view ≥ 0.6) Enclosed (sky view < 0.6)
Figure 2. Modelled nocturnal cooling rate against canopy cover for open and enclosed blocks, holding impervious cover and wind at their means. Line style and marker shape, not colour, distinguish the series. Note: points are model predictions at canopy bin midpoints; numerical values and confidence intervals are given in Table 1.

3 Results

3.1 Network performance and data retention

Of 312 loggers deployed, 297 returned complete records for all three summers; fifteen failed through battery faults, vandalism, or water ingress and were replaced within a median of six days. After screening for cloud and wind, 3,512 of 4,140 block-nights (84.8%) were retained. Retention did not differ by canopy tertile (χ² = 1.9, two degrees of freedom).

Intercalibration offsets were small: the median absolute offset against the chamber reference was 0.09 °C and no sensor exceeded 0.25 °C (Table A1). End-of-season checks showed drift below 0.05 °C for all sensors retained in analysis.

3.2 Canopy cover and cooling rate

Blocks cooled at a mean of 0.33 °C h⁻¹ between sunset and 04:00 (standard deviation 0.09). Cooling was fastest in Lindqvist Bay, the most exposed of the three cities, and slowest in Port Amsel, where older terraces enclose many streets.

In the adjusted model, canopy cover was associated with faster cooling: each ten-percentage-point increase corresponded to 0.061 °C h⁻¹ faster cooling (95% CI 0.048 to 0.074). Impervious cover slowed cooling (−0.012 °C h⁻¹ per ten points; 95% CI −0.019 to −0.005), and wind speed had the expected positive association. Table 1 reports predicted cooling by canopy bin and sky view class.

Table 1. Predicted cooling rate (°C h⁻¹) by canopy cover and sky view class.
Canopy (%)OpenEnclosed
0–100.250.21
10–200.310.26
20–300.380.32
30–400.430.34
40–500.450.33
50–600.460.30
≥ 600.460.27

Note: open blocks have mean sky view factor ≥ 0.6. Predictions hold impervious cover at 52% and wind speed at 1.4 m s⁻¹. Bootstrap 95% intervals are ±0.02 to ±0.04 °C h⁻¹ for all cells.

The interaction between canopy and sky view was substantial (β₄ = 0.29; 95% CI 0.18 to 0.40). Among open blocks the canopy association was monotonic and flattened above about 45% canopy. Among enclosed blocks it peaked between 30% and 40% canopy and declined thereafter, so that the most heavily treed enclosed blocks cooled no faster than sparsely treed ones (Figure 2).

3.3 Block-level estimates

Block-level means, intervals, and covariates are listed in Table 2. The table runs across pages in full page width; its header row repeats on each page where the renderer supports repeated table headers.

Table 2. Block characteristics and mean nocturnal cooling rate, all 48 study blocks.
CityBlockCanopy (%)Impervious (%)Sky viewNightsCooling (°C h⁻¹)95% CI
HarrowgateHG-0111800.80830.170.14 to 0.21
HarrowgateHG-0230590.69730.310.25 to 0.36
HarrowgateHG-0316790.77630.200.14 to 0.24
HarrowgateHG-0456350.51770.520.47 to 0.57
HarrowgateHG-0543450.54750.400.37 to 0.45
HarrowgateHG-0644430.60780.460.42 to 0.49
HarrowgateHG-0730600.69630.300.25 to 0.35
HarrowgateHG-0827630.70690.330.29 to 0.38
HarrowgateHG-0946440.57700.460.42 to 0.51
HarrowgateHG-1030650.69620.350.31 to 0.40
HarrowgateHG-1152310.54870.520.47 to 0.57
HarrowgateHG-1238430.60660.410.37 to 0.47
HarrowgateHG-1322590.74620.250.21 to 0.30
HarrowgateHG-1432590.63820.340.28 to 0.39
HarrowgateHG-1533510.63630.340.30 to 0.37
HarrowgateHG-1612800.78700.220.17 to 0.26
Lindqvist BayLB-0113760.81770.230.19 to 0.27
Lindqvist BayLB-0236440.59860.400.36 to 0.44
Lindqvist BayLB-0348450.55860.460.41 to 0.52
Lindqvist BayLB-0427520.72710.340.29 to 0.39
Lindqvist BayLB-0529570.68750.350.32 to 0.40
Lindqvist BayLB-0629590.64650.310.26 to 0.36
Lindqvist BayLB-075860.83760.180.14 to 0.22
Lindqvist BayLB-0811710.84840.220.18 to 0.26
Lindqvist BayLB-0929550.70620.360.31 to 0.41
Lindqvist BayLB-1048290.52670.450.39 to 0.48
Lindqvist BayLB-1127660.72670.310.26 to 0.36
Lindqvist BayLB-1229460.65840.340.30 to 0.39
Lindqvist BayLB-1328490.70730.300.25 to 0.35
Lindqvist BayLB-1427580.68690.320.27 to 0.38
Lindqvist BayLB-1524700.70880.290.25 to 0.33
Lindqvist BayLB-1628620.69780.300.24 to 0.34
Port AmselPA-0149340.49620.480.43 to 0.52
Port AmselPA-0256310.46800.540.50 to 0.58
Port AmselPA-0311720.83660.230.19 to 0.28
Port AmselPA-0425540.72750.320.28 to 0.36
Port AmselPA-0528650.68740.340.28 to 0.39
Port AmselPA-0642450.60640.420.37 to 0.46
Port AmselPA-0714730.80650.200.14 to 0.25
Port AmselPA-0842460.58680.450.40 to 0.51
Port AmselPA-0928540.69810.330.29 to 0.39
Port AmselPA-1054320.50880.480.44 to 0.51
Port AmselPA-1121690.77680.250.20 to 0.30
Port AmselPA-1219760.77670.230.18 to 0.27
Port AmselPA-1311670.82870.180.13 to 0.23
Port AmselPA-1425570.74880.280.24 to 0.33
Port AmselPA-1537500.59760.390.34 to 0.43
Port AmselPA-1626520.66760.290.24 to 0.32

Note: nights counts retained clear, calm nights across three summers. Confidence intervals are block-cluster bootstrap intervals with 2,000 resamples. Source: street-scale night cooling dataset [25].

3.4 Sensitivity analyses

Results were robust to excluding the first two weeks of each summer, when leaf area was still increasing [29]; to using air rather than surface temperature as the outcome (attenuated by about one third); and to defining the night window from sunset to sunrise. Residual Moran's I was 0.04 (p = 0.31), indicating no meaningful spatial autocorrelation after adjustment.

A model with city-specific canopy slopes did not improve fit (likelihood-ratio test p = 0.42). We therefore report the pooled estimate, while noting that Port Amsel contributed most of the enclosed, heavily treed blocks.

4 Discussion

In three maritime-temperate cities, residential blocks with more tree canopy cooled faster at night, except where dense canopy combined with low sky view. This pattern reconciles the apparently contradictory findings of earlier, smaller studies: open suburban settings sample the rising limb of the relationship, while compact neighbourhoods sample its falling limb [1], [6].

The mechanism we favour is the balance between daytime shading and night-time sky obstruction. Moderate canopy reduces the heat stored in paving and walls during the day [18], so there is less to release at night; as canopy closes over the street, it also obstructs the cold night sky, slowing radiative loss [9]. Where the street is already enclosed by buildings, the extra obstruction from trees tips the balance.

Box 1. What this study cannot show

  • Cooling rate is not exposure. Indoor temperature depends on building fabric and ventilation behaviour, which we did not measure.
  • The design is observational. Canopy was not assigned, and unmeasured differences between blocks may remain.
  • Three cities in one climate zone cannot establish how the relationship behaves in hotter or drier climates.

These findings have practical implications. Planting programmes in Northmere commonly set canopy targets of 30% to 40% for residential streets [10]. Our results support such targets in open streets but suggest that, in narrow terraced streets, species with lighter crowns or wider spacing may deliver more night-time benefit than maximum canopy closureb. Equity matters too: the enclosed, older streets where dense canopy retained heat are also where residents with the highest heat vulnerability tend to live [12].

The study has strengths beyond its size. Pre-registration constrained analytic flexibility [23], [26]; intercalibration and standard shields limited instrument error [3]; and the crossed design separated block from weather variation. Its limitations are summarised in Box 1.

5 Conclusions

Residential tree canopy is associated with faster night-time surface cooling in mid-latitude maritime cities, up to a point. Beyond roughly 40% canopy in streets that are already enclosed, additional canopy is associated with slower cooling. Guidance that sets canopy targets without regard to sky view risks under-delivering night-time heat benefits where they are most needed.

Declarations

Acknowledgements

We thank the residents of the 48 study blocks for permitting sensor installation, and the municipal forestry offices of all three cities for access to inventory data.

Funding

This work was supported by the Northmere Research Council (fictional grant NRC-UC-2022-117) and by the Lindqvist Bay Institute of Technology seed fund. The funders had no role in design, analysis, or the decision to publish.

Competing interests

R.Q. is employed by the Port Amsel Climate Office, which provided wind and cloud data. The other authors declare no competing interests.

Author contributions

Conceptualisation: A.O., R.Q. Methodology: A.O., T.L.-R., J.C. Instrumentation: M.-L.H. Formal analysis: A.O., J.C. Writing, original draft: A.O. Writing, review and editing: all authors.

Data and code availability

Block-night data and processing code are deposited at https://doi.org/10.5555/fdr.2026.ncool [25]; sensor locations are generalised to block centroids to protect residents' privacy.

Ethics

Sensor siting on private frontages was approved by the University of Harrowgate research ethics committee (fictional reference UH-REC-23-041).

Supplementary material

Supplementary Tables S1–S4 and the pre-registration are available with the online version at https://example.org/fjfs/14/3/e0412/supplementary.

Notes

  1. We use “sky view factor” for the fraction of the overhead hemisphere visible from a point, from 0 (fully obstructed) to 1 (open sky).

    ↩
  2. Species recommendations are outside the scope of this study; see the Harrowgate street tree inventory for crown-form data [10].

    ↩

References

  1. [1] Arden P, Whitcombe S, Laine R. Nocturnal heat retention in compact low-rise neighbourhoods: a review of field campaigns, 1995–2020. Fenwick J Field Sci. 2021;9(2):44–71. doi:10.5555/fjfs.2021.0902
  2. [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. [3] Quintero R, Abernathy-Svensson K, Mbeki T, Larkspur D, Oyelaran F, Tanaka H, Volkov I. A harmonised protocol for sensor intercalibration across municipal monitoring networks with heterogeneous hardware generations. Environ Monit Methods (fictional). 2024;31(4):512–540. doi:10.5555/emm.2024.3104512
  4. [4] Lindgren-Reyes T. Sky view factor estimation from street-level imagery. J Urban Geom (fictional). 2022;5:88–104. doi:10.5555/jug.2022.005088
  5. [5] Carberry J, Okafor A. Why nights matter: diurnal asymmetry in urban heat exposure. Fenwick J Field Sci. 2024;12(1):1–19. doi:10.5555/fjfs.2024.1201
  6. [6] Hartono M-L, Quintero R. Tree canopy, impervious cover, and the cooling rate of residential surfaces. Urban Microclim Lett (fictional). 2023;8:211–219. doi:10.5555/uml.2023.008211
  7. [7] Whitcombe S. Surface energy balance in courtyards. Port Amsel Univ Press; 2019.
  8. [8] Bexley N, Ostrowski J, Adeyemi K. Mixed-effects models for repeated environmental measurements nested in space and time. Stat Field Sci (fictional). 2020;14:300–327. doi:10.5555/sfs.2020.014300
  9. [9] Laine R, Arden P. Radiative cooling of pavements under clear skies. Fenwick J Field Sci. 2018;6(3):155–170. doi:10.5555/fjfs.2018.0603
  10. [10] Municipal Forestry Office of Harrowgate. Street tree inventory 2024: methods and data dictionary. Harrowgate (fictional): City Publications; 2024.
  11. [11] Tanaka H, Volkov I. Lidar-derived canopy height models for urban forestry: accuracy assessment at parcel scale. Remote Sens Cities (fictional). 2021;3(2):77–95. doi:10.5555/rsc.2021.0302077
  12. [12] Oyelaran F. Equity dimensions of night-time heat exposure. J Environ Justice Field Stud (fictional). 2022;2:9–30. doi:10.5555/jejfs.2022.002009
  13. [13] Abernathy-Svensson K. Shield design for passive temperature sensors. Instrum Notes (fictional). 2020;41:12–18. doi:10.5555/in.2020.041012
  14. [14] Mbeki T, Larkspur D. Cloud cover screening for nocturnal observation windows. Atmos Field Methods (fictional). 2023;19:401–417. doi:10.5555/afm.2023.019401
  15. [15] Arden P, Okafor A, Carberry J. Open data practices for citizen-assisted sensor networks. Fenwick J Field Sci. 2025;13(2):88–102. doi:10.5555/fjfs.2025.1302
  16. [16] Port Amsel Climate Office. Climatological normals 1991–2020: station summaries. Port Amsel (fictional); 2021. Available from: https://example.org/port-amsel/climate/normals/1991-2020/station-summaries-with-quality-control-flags.pdf.
  17. [17] Ostrowski J. Spatial autocorrelation diagnostics for block-level panels. Stat Field Sci (fictional). 2022;16:55–70. doi:10.5555/sfs.2022.016055
  18. [18] Hartono M-L. Thermal inertia of residential building stock. Build Phys Q (fictional). 2021;30:140–162. doi:10.5555/bpq.2021.030140
  19. [19] Volkov I, Tanaka H, Quintero R. Open-source processing pipeline for logger networks. J Open Field Softw (fictional). 2024;7:e112. doi:10.5555/jofs.2024.007112
  20. [20] Laine R. Wind sheltering by street trees. Fenwick J Field Sci. 2020;8(1):33–47. doi:10.5555/fjfs.2020.0801
  21. [21] Adeyemi K, Bexley N. Robust standard errors for clustered environmental panels. Stat Field Sci (fictional). 2019;13:210–231. doi:10.5555/sfs.2019.013210
  22. [22] Whitcombe S, Laine R. Evaporative cooling and irrigation regimes in private gardens. Urban Microclim Lett (fictional). 2022;7:100–109. doi:10.5555/uml.2022.007100
  23. [23] Carberry J. On the reporting of null results in field science. Fenwick J Field Sci. 2023;11(4):301–305. doi:10.5555/fjfs.2023.1104
  24. [24] Lindqvist Bay Institute of Technology. Research data management policy, version 3.2. Lindqvist Bay (fictional); 2024.
  25. [25] Okafor A, Lindgren-Reyes T, Hartono M-L. Street-scale night cooling dataset, three cities, 2023–2025 [dataset]. Fenwick Data Repository (fictional); 2026. doi:10.5555/fdr.2026.ncool
  26. [26] Quintero R, Okafor A. Pre-registration: canopy cover and nocturnal cooling. Open Field Registry (fictional); 2023. doi:10.5555/ofr.2023.0417
  27. [27] Larkspur D. Heat health warning thresholds and their night-time components. Public Health Field Rep (fictional). 2024;22:60–79. doi:10.5555/phfr.2024.022060
  28. [28] Arden P. Field methods in urban climatology: a practical handbook. 2nd ed. Fenwick Academic Press (fictional); 2022.
  29. [29] Tanaka H. Seasonal leaf-area dynamics of common street trees in maritime climates. Remote Sens Cities (fictional). 2023;5(1):12–29. doi:10.5555/rsc.2023.0501012
  30. [30] Bexley N. A gentle introduction to crossed random effects. Stat Field Sci (fictional). 2021;15:1–24. doi:10.5555/sfs.2021.015001

Appendix A. Sensor intercalibration

Per-sensor calibration coefficients for surface sensors, by city. Offsets are degrees Celsius relative to the chamber reference at four set points.

Table A1. Surface sensor intercalibration offsets, slopes, and fit by sensor (five example sensors per city).

SensorCityModelShieldHeight (m)Offset 10 °COffset 20 °COffset 30 °COffset 40 °CSlopeR²Service notes
HG-S01HarrowgateTL-21bLouvred2.0-0.24+0.07-0.22+0.130.9790.988Battery fault, 6 nights
HG-S02HarrowgateTL-21bTube2.5-0.15-0.06-0.07-0.150.9930.998None
HG-S03HarrowgateTL-20Louvred2.0-0.09-0.09+0.02-0.231.0250.992None
HG-S04HarrowgateMk IITube3.0+0.01+0.10+0.07+0.241.0240.988None
HG-S05HarrowgateTL-20Tube2.0+0.23+0.07-0.03-0.061.0030.995None
LB-S01Lindqvist BayTL-20Louvred2.0-0.06-0.20+0.10+0.170.9930.998None
LB-S02Lindqvist BayTL-20Tube3.0+0.17+0.07+0.06-0.111.0200.983None
LB-S03Lindqvist BayMk IILouvred2.5-0.05+0.10+0.23+0.101.0110.994None
LB-S04Lindqvist BayMk IILouvred2.0+0.25-0.11-0.17-0.240.9850.987Replaced 2024-08
LB-S05Lindqvist BayTL-21bTube2.0-0.20-0.21-0.19+0.011.0100.994Battery fault, 6 nights
PA-S01Port AmselTL-20Louvred2.5+0.15+0.07-0.09-0.251.0010.995None
PA-S02Port AmselTL-20Tube2.0+0.24+0.03+0.04+0.080.9770.989None
PA-S03Port AmselTL-20Louvred3.0-0.22-0.07-0.24-0.231.0180.995None
PA-S04Port AmselTL-20Louvred2.0+0.19-0.16+0.01+0.040.9780.981Replaced 2024-08
PA-S05Port AmselTL-20Tube3.0+0.00-0.17+0.13+0.170.9950.988Replaced 2024-08

Fixture notice: this is a fictional article written for Folio's JOURNAL-ARTICLE-01 fixture. The journal, authors (other than the ORCID example identity Josiah Carberry), institutions, data, grant numbers, and references are invented. DOIs use the Crossref test prefix 10.5555 and do not resolve to real works.