Classification: Physical Science: Environmental Science
Wind farms change the ground-level climate
Alona Armstrong1, 2, Ralph R. Burton3, Susan E. Lee3+, Stephen Mobbs3, Nicholas Ostle1, 4,
Victoria Smith3, Susan Waldron2 and Jeanette Whitaker4
1
School of Geographical and Earth Sciences, University of Glasgow, Glasgow G12 8QQ, UK
2
Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ, UK
3
National Centre for Atmospheric Science, School of Earth and Environment, University of
Leeds, Leeds, LS2 9JT, UK
4
Centre for Ecology and Hydrology, Lancaster Environment Centre, Library Avenue,
Bailrigg, LA1 4AP, UK
+
Now at School of Civil Engineering, University of Birmingham, Edgbaston, Birmingham,
B15 2TT, UK
Corresponding author: a.armstrong@lancaster.ac.uk, +44 (0)1524 510243
Keywords: wind turbines, micro-climate, carbon cycling
1
Abstract
The global drive to produce low-carbon energy has resulted in an unprecedented deployment
of onshore wind turbines: there has been a 25 % increase in wind turbine production capacity
over the last decade which is predicted to grow annually by 6 % between 2011 and 2035 (1).
This represents a significant land use change for wind energy generation, from 75,200 km 2 in
2012 to 301,333 km2 in 2035, with uncertain consequences for local climatic conditions and
the regulation of ecosystem carbon (C) cycling. There is a paucity of data on the effects of
wind farms on soil and atmospheric climates (2-4), limiting our ability to determine the true
C balance of this renewable energy technology (5). Here, we present high-resolution data
from a wind farm collected during operational and idle periods that shows that wind farms
affect several measures of ground-level climate. Specifically, we discovered that operational
wind turbines raised air temperature (TA) by 0.22 °C and absolute humidity (AH) by 0.03 g
m-3 during the night, and increased the variability in air, surface and soil temperature
throughout the diurnal cycle. Further, the microclimatic influence of turbines on TA and AH
decreased logarithmically with distance from the nearest turbine. These effects on local
microclimate have uncertain implications for ecosystem C cycling and understanding needs
to be improved to determine the overall C balance of wind energy.
2
Significance statement (120 word lay person stating significance)
The nascent increase in wind turbine deployment across the world, promoted by the drive to
produce low-carbon energy, represents a significant land-use change, with uncertain
consequences for the hosting ecosystems. We present the first field data showing that wind
turbines can affect several measures of the ground-level microclimate known to influence
carbon cycling. We demonstrate warming and moistening at night-time, and increases in the
variability in air, surface and soil temperatures, and absolute humidity. Given the centrality of
climate, and especially temperature, as a driver of terrestrial C cycling, these findings
demonstrate the potential for the functioning of the host landscape to be affected by wind
farm operation. We estimate that carbon sequestration may be reduced by 4% per annum.
3
Introduction
The electricity generating capacity of wind turbines is predicted to increase more than any
other renewable, amounting to over 7 % of global production by 2035 (6). Deployment of
this magnitude will result in substantial land use change, an additional 226,133 km2, given
the relatively low energy density of wind power (7) and the anticipation that 80 % of wind
capacity will be onshore (6). Effects of this land use change on human populations, nonvolant wildlife, avian and bat communities have received consideration (8-11). However, the
effect on ground-level microclimates has not been resolved, despite implications for
processes and properties, including C cycling (5), of the hosting ecosystem.
Wind farms have been postulated to affect climatic conditions from the within-farm to global
scale through modification of the vertical distribution of energy and moisture within the
atmosphere and their exchange between the land surface and atmosphere (12). Previous
studies have modelled effects (12-16), measured air temperature differences upwind and
downwind of wind turbines (2-4), and used satellite data to examine temperature effects over
a 10,000 km2 area (17). To fully resolve the implications of such temperature differentials on
hosting ecosystems, high resolution spatially explicit field data are needed when turbines are
operational and idle.
We used TA and AH data during a meteorologically ‘normal’ period (see SI) from 101
locations across an area of peatland at Black Law Wind Farm, Scotland, plus surface and soil
temperature (TSU and TSO, respectively) at 36 locations, clustered at four sites, across the
whole farm (Fig. S1) to assess whether there was wind turbine-induced changes to the
microclimate. We compared the data during periods when the wind farm was operational
(ON) and idle (OFF), and sites downwind (D) or not downwind (ND) of wind turbines (Fig.
4
S1), where appropriate. Given the trend for stable boundary layers at night (warm air above
cold air) and neutral (well-mixed) or unstable (cold air above warm air) boundary layers
during the day we examined the effects during the day and night, and diurnal patterns.
Results and discussion
Integrated effects of distance to turbine affects TA and AH. Firstly, to determine the
integrated effect of the whole wind farm across the measurement period, we analysed the
mean day and night-time differences in TA and AH departure for each site from the site-wide
mean between ON and OFF periods, using all data (i.e. data were not categorised as D or
ND) (see Fig. S1). To assess the spatial extent of effects we related the departures to distance
from the nearest turbine.
During the night, air closer to a wind turbine was warmer and moister, with TA departures
reaching 0.25 °C and AH departures 0.1 g m-3 and positive departures evident up to 200 m
away (Fig. 1a). The night-time warming and moistening was caused by downward mixing of
warmer moister air by the turbines during stable conditions (18). Analysis of midnight
soundings from the two nearest upper-air stations reveals that during the most stable
conditions the lapse rate is positive for both temperature and moisture (see SI) (13). During
the day, air closer to a wind turbine was cooler, with departures up to 0.05 °C, but AH was
not influenced (Fig. 1b). This weaker day-time effect is attributable to a convectively-driven,
well mixed boundary layer (18). Importantly, the trends observed for TA during the day and
night and AH at night can be approximated by a logarithmic function (Fig. 1), demonstrating
the effect of wind turbines can be quantified and thus represented in models of Earth surface
energy balance (e.g. numerical weather prediction models).
5
Temporal-spatial variation in TA, TSU, TSO, and AH. To quantify the temporal-spatial
effects of wind turbine operation we examined the variation in TA, TSU, TSO, and AH
throughout the diurnal cycle (irrespective of wind direction) using the coefficient of variation
of each of the measurements across the whole site for each measurement interval averaged
for both the ON and OFF periods as a function of time of day.
The spatial variations in the TA, TSU, TSO and AH data were significantly greater during the
ON period compared with the OFF period (Fig. 2), suggesting turbines increased vertical
mixing and turbulence. This has not been evident in other studies (2-4, 12, 13, 17, 18). The
differences in temperature variation between ON and OFF periods were greatest for TSU data
(Fig. 2b) and smallest for TA data (Fig. 2a), reflecting relatively well-mixed air at 2 m and
increased variability at the surface arising from peatland micro-topography and vegetation
shading.
The difference in TA and AH variability between the ON and OFF periods was greater during
night than day (Figs 2a, d), whereas it was approximately equal for TSU and TSO, with smaller
differences during transition periods around sunrise and sunset (Figs 2b, c). This
demonstrates that night-time TA and AH are most sensitive to turbines, due to downward
mixing of warm air. Further, the mixing down of warm air during the ON period appears to
have affected the diurnal trend in TSO and AH: the peaks in variability in TSO and AH were
later in the day during the OFF period compared with the ON period (Figs 2c, d). This
suggests that the night-time background gradient of TSO and AH are eroded by turbulence
earlier in the day during the ON period.
Directional effects of turbines on TA and AH. Our final analysis examined the magnitude
6
of effect of wind turbines on TA and AH during the night when the relationship between
departure and distance from the turbine was stronger (Fig. 1, see SI for day-time results). We
excluded all data from one hour before sunrise to one hour after sunset and filtered for wind
directions between 220° and 240° (aligned along the main axis of the wind farm) (Fig. S1).
Temperature departures for all the data that fit these criteria were calculated for each
sampling location, (as for Fig. 1), thus allowing the relative temperature and AH of sites
downwind and not downwind of turbines to be compared.
We found that temperatures were significantly warmer (Fig. 3a) in areas downwind of
turbines during the ON period, with an average relative warming of 0.22 °C. Although
sensors recorded lower relative humidity (RH) close to the turbines (Fig. S5), the AH of air
downwind of the turbines was, on average, 0.03 g m-3 moister (Fig. 3b). This was due to the
exponential variation of saturated water vapour pressure with temperature. Analysis of the
data shows that the relative increase in saturated water vapour pressure had more effect on
AH than the combined increase in temperature, and lowering of RH (see SI), consistent with
turbine-induced mixing of warmer and moister air downwards (12). During the OFF period,
temperature and AH departures were variable with no statistically significant difference
between D and ND sites (Figs 3c, d).
Conclusion
Together, these findings provide the first field evidence that operational wind turbines can
have a measureable effect on ground-level climate: increased TA and AH during the night and
greater variability in TA, TSU, TSO and AH throughout the diurnal cycle. Although the effects
on TA and AH were statistically significant, the observed differences were smaller than the
background variation recorded across the site (Fig. 3), attributable to differences in
7
topography, hydrology, soil properties and vegetation type. Importantly, we demonstrate that
the effects on both TA and AH can be described by a logarithmic function of distance from
nearest turbine, a generic approach showing for the first time how the integrated effect of a
wind farm may be estimated.
Given the centrality of climate, and especially temperature, as a driver of terrestrial C
cycling, these findings demonstrate the potential for the biogeochemical functioning of the
host landscape to be affected by wind farm operation. Night-time warming was most
influenced, suggesting that microclimate effects on night-time biological processes that
govern C cycling, i.e. decomposition and plant-soil respiration, may be important. However,
effects on respiration could also be offset by plant physiological responses to warming (19).
Our results indicate that wind farm operation had greater effects on spatial and diurnal
variability in TSU and TSO than TA. TSU and TSO are recognised as stronger regulators of
plant-soil C dynamics (20, 21), consequently, the effects on the net C balance of the hosting
ecosystem may be stronger than potentially inferred from previous studies which only
measured TA (2-4).
Although, the scale of the observed effects on TA and AH was small relative to typical
seasonal and diurnal variances, changes of a similar magnitude have been postulated to
impact on ecosystem C gains and losses at the global scale given the temperature sensitivity
of productivity and decomposition processes (22). Our results, calibrated against C flux data
from a peatland in northern England (23), suggest an estimated reduction in C sink capacity
of 4 % attributable to night-time warming caused by wind turbines. Together, these findings
advance our scientific understanding of the effect of operational wind turbines on groundlevel climate. Given the importance of temperature controls on plant-soil C cycling, improved
8
estimates of wind turbine-induced microclimatic effects on soil C stocks and greenhouse gas
emissions is needed.
Materials and methods
Site: Black Law Wind Farm, Scotland (55°46′01″N 03°44′20″W, altitude 250-320 m),
comprises 54 turbines over 18.6 km2. The turbine hub heights are approximately 70 m and
the total capacity is 124 MW.
Measurements: Air temperature (TA) and relative humidity (RH) were measured every 15
minutes using HOBO U23 loggers (Onset, USA) at 2 m above the land surface at 101
locations across a 2.6 by 1.4 km area of Black Law Wind Farm (Fig. S1). Absolute humidity
(AH) was derived from RH and TA. Surface and soil (-5 cm) temperatures (TSU and TSO,
respectively) were recorded every 30 minutes using Hobo Pendant data loggers (Onset, USA)
at 36 locations, clustered at four sites (Fig. S1). At the same sites soil moisture (-10 cm) was
measured every 30 minutes using Campbell CS625 water content reflectometers (using a site
specific calibration) connected to Campbell CR200 loggers (Campbell Scientific Limited,
UK). Wind direction was measured with a Gill 2D sonic anemometer (Gill Instruments, UK)
at 2 m above the land surface every 10 seconds and averaged over 10 minute intervals (Fig.
S1).
The wind farm was operational (hereafter referred to as ON) from 24th May 2012 to 7th June
2012, idle (hereafter OFF) from 12th June 2012 to 25th July 2012 and operational (ON) from
28th July to 15th November 2012. The switching on and off of the turbines was a phased
operation lasting several days; data from this period were excluded.
9
Data processing: ‘Day’ was classified as one hour after sunrise to one hour before sunset, as
defined by the National Oceanic and Atmospheric Administration algorithm (24), and ‘night’
as one hour after sunset to one hour before sunrise to avoid transition periods. Fraction of
day, between 0 and 1, was calculated, with sunrise as 0.0 (or 1.0), sunset as 0.5 and time
linearly scaled between given sunrise and sunset are the meteorologically relevant temporal
controls on boundary layer development (25) (using time-of-day is potentially misleading
since the sun rose and set at significantly different times during the period of measurement at
Black Law). The data were binned into 24 pseudo-hourly bins for analysis.
Absolute humidity (AH) was derived from relative humidity (RH) by first calculating the
water vapour saturation pressure (SWVP - temperature-dependant), and then calculating the
vapour pressure (temperature- and RH-dependant) then finally the absolute humidity based
upon temperature and vapour pressure. The SWVP was found to be the dominant control on
AH as the former varies exponentially with temperature, as explained more fully in the SI.
Departures were used in the analysis of the TA, AH and RH data to remove diurnal and
seasonal signals (the greatest controls) from the data, (Figs 1, 3, S5). Essentially, the sitewide instantaneous mean was calculated for all measurement locations, and subtracted from
each individual measurement and then averaged over the ON and OFF periods and for D and
ND groups where appropriate. Consider first the vector of N hobo temperatures at time t:
𝑇(𝑡) = [T1(t), T2(t), …., TN(t)] for t = t1, t = t2, …, t = tM (i.e. M times in the sample set)
then calculating the site-wide scalar mean temperature at time t = tj gives
1
TAV (𝑡𝑗 ) = 𝑁 ∑𝑁
𝑖=1 𝑇𝑖 (𝑡𝑗 )
The vector of departures from the mean at time 𝑡𝑗 is
10
𝑇′(𝑡𝑗 ) = [T1(𝑡𝑗 ) - TAV(𝑡𝑗 ), T2(𝑡𝑗 ) - TAV(𝑡𝑗 ), ..., TN(𝑡𝑗 ) - TAV(𝑡𝑗 )]
allowing the time-averaged mean departure vector to be constructed:
̅ =[ 1 ∑𝑀 𝑇 ′ (𝑡𝑗 ) , 1 ∑𝑀 𝑇 ′ (𝑡𝑗 ) , … , 1 ∑𝑀 𝑇 ′ (𝑡𝑗 )]
𝑇′
𝑀 𝑗=1 1
𝑀 𝑗=1 2
𝑀 𝑗=1 𝑁
This vector is the one used in the paper, and when departures are mentioned they refer to this
vector, or elements of it.
Variation in TA, TSU, TSO, and AH was assessed by calculating the coefficient of variation
across the site for each measurement interval and averaging in pseudo hourly bins (based on
fraction of the day) for ON and OFF periods. The coefficient of variation cv was calculated
using the same steps as above for TA, AH and RH departures: the site-wide mean cv was
calculated and the time-averaged cv vector constructed (the vector of departures step was
omitted). Given the cv of a population is defined as the standard deviation divided by the
mean of the population: cv = σ / μ, the temperatures were first converted to degrees Kelvin. It
is inappropriate to use degrees Celsius to calculate cv, as negative or zero values of μ would
then be allowed, giving meaningless values of cv .
To examine the magnitude of effect of the turbines on TA, RH and AH, only data when the
wind direction was from between 220° and 240° (i.e. aligned along the main axis of the wind
farm) was considered (Fig. S1). Firstly, any cumulative effect of the turning turbines is likely
to be greatest for this directional spread as the wind direction is approximately aligned with
the main axis of the wind farm. Secondly, this directional filtering allows the measurement
instruments to be divided into three groups by location: (i) those that are always downwind of
the turbines – the downwind group, D; (ii) those that are never downwind of the turbines –
the not downwind group, ND; (iii) those that may or may not be downwind of the turbines,
the exclusion group, which were excluded from this directional analysis (Fig. S1). There was
11
sufficient data satisfying this filtering to test for statistical significance. Also, for the data to
be included the wind had to be originating from between 220° and 240° at the measurement
time and the previous 30 minutes, and also satisfy the day-night criteria for the entire 30
minutes period. This allowed the development of any underlying signal to be captured, while
still providing a significant amount of data to be retained. Departures for all the data that fit
these criteria were calculated for each sampling location, (as for Fig. 1, see above), thus
allowing the relative temperature and AH of sites downwind and not downwind of turbines to
be compared.
Statistics: Differences in coefficient of variation between ON and OFF periods for each
pseudo hour (Fig. 2) and temperature departures between D and ND groups were tested (Fig.
3) using a t-test with unequal variances using Stata13 (StataCorp, Texas).
Calculating the change in ecosystem carbon loss from Scottish wind farms: We made a
crude estimation of the effects of the temperature changes observed in this study on the C
balance of wind farm hosting ecosystems in Scotland, to provide an indication of the
potential magnitude of impact. We used the respiration rates from a peatland warming
experiment in northern England to derive the potential change in the C balance. Given there
were no significant differences in temperature between the D and ND groups during the day
when the turbines were OFF (see SI), we assumed day time fluxes were as for the control
vegetation plot at the northern England site. We assumed the 0.22 °C temperature increase
we observed at night occurred during the growing and non-growing season. We used the
ambient and warmed respiration rates from the control vegetation plots in the northern
England study and to calculate the effect of a 0.22 °C temperature rise by linearly scaling. We
assumed an average day length of 12 hours 20 minutes and that the growing season occurred
12
from May to September (as per the northern England study(24)) to calculated the fluxes over
the entire year with and without wind turbine-induced warming and calculated the percentage
difference.
Acknowledgements
This study was supported by the UK Natural Environment Research Council
(NE/H01036X/1, NE/H010351/1, NE/H010335/1). AA acknowledges financial support from
an Energy Lancaster fellowship during which all data analysis and manuscript preparation
was done. We thank Scottish Power Renewables and the land owners for allowing site access.
We thank Martin Coleman, Hemanth Pasumarthi, Salvatore Peppe, Harriet Richardson,
Kenny Roberts, Fraser Russell, Gavin Thompson and Scott Wylie for assistance in the field,
and Barbara Brooks, James Groves, Salvatore Peppe and Felicity Perry for assistance
calibrating the loggers.
Author Contributions
A.A., R.R.B, S.E.L, S.M., N.O., S.W. and J.W. conceived the research. A.A., S.L., V.S. and
S.W. undertook the field research. A.A. and R.R.B. analysed the data. A.A. and R.R.B.
drafted the manuscript with feedback and contributions from S.E.L, S.M., N.O., S.W. and
J.W.
References
1.
2.
3.
IEA (2013) World Energy Outlook 2013 (International Energy Association, Paris, France).
Baidya Roy S & Traiteur JJ (2010) Impacts of wind farms on surface air temperatures.
Proceedings of the National Academy of Sciences 107(42):15679-15684.
Rajewski DA, et al. (2013) CWEX: Crop/Wind-energy EXperiment: Observations of surfacelayer, boundary-layer and mesoscale interactions with a wind farm. Bulletin of the American
Meteorological Society 94(5):655-672.
13
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Smith CM, Barthelmie RJ, & Pryor SC (2013) In situ observations of the influence of a large
onshore wind farm on near-surface temperature, turbulence intensity and wind speed
profiles. Environmental Research Letters 8(3):034006.
Armstrong A, Waldron S, Whitaker J, & Ostle NJ (in press) Wind farm and solar park effects
on plant-soil carbon cycling: uncertain impacts of changes in ground-level microclimate.
Global Change Biology.
IEA (2012) World Energy Outlook 2012: Renewable Energy Outlook (International Energy
Association, Paris, France).
Denholm P, Hand M, Jackson M, & Ong S (2009) Land-Use Requirements of Modern Wind
Power Plants in the United States. (U.S. Department of Energy, National Renewable Energy
Laboratory, Colorado).
Lovich JE & Ennen JR (2013) Assessing the state of knowledge of utility-scale wind energy
development and operation on non-volant terrestrial and marine wildlife. Applied Energy
103(0):52-60.
Northrup JM & Wittemyer G (2013) Characterising the impacts of emerging energy
development on wildlife, with an eye towards mitigation. Ecology Letters 16(1):112-125.
Pearce-Higgins JW, Stephen L, Douse A, & Langston RHW (2012) Greater impacts of wind
farms on bird populations during construction than subsequent operation: results of a multisite and multi-species analysis. Journal of Applied Ecology 49(2):386-394.
Knopper L & Ollson C (2011) Health effects and wind turbines: A review of the literature.
Environmental Health 10(1):78.
Baidya Roy S, Pacala SW, & Walko RL (2004) Can large wind farms affect local meteorology?
Journal of Geophyscial Research 109(D19):D19101.
Baidya Roy S (2011) Simulating impacts of wind farms on local hydrometeorology. Journal of
Wind Engineering and Industrial Aerodynamics 99:491-498.
Fiedler BH & Bukovsky MS (2011) The effect of a giant wind farm on precipitation in a
regional climate model. Environmental Research Letters 6:045101.
Keith DW, et al. (2004) The influence of large-scale wind power on global climate.
Proceedings of the National Academy of Sciences 101(46):16115-16120.
Wang C & Prinn RG (2010) Potential climatic impacts and reliability of very large-scale wind
farms. Atmospheric Chemistry and Physics 10:2053-2061.
Zhou L, et al. (2012) Impacts of wind farms on land surface temperature. Nature Clim.
Change 2(7):539-543.
Zhou L, Tian Y, Baidya Roy S, Dai Y, & Chen H (2013) Diurnal and seasonal variations of wind
farm impacts on land surface temperature over western Texas. Clim Dyn 41(2):307-326.
Peng S, et al. (2013) Asymmetric effects of daytime and night-time warming on Northern
Hemisphere vegetation. Nature 501(7465):88-92.
De Frenne P, et al. (2013) Microclimate moderates plant responses to macroclimate
warming. Proceedings of the National Academy of Sciences.
Graae BJ, et al. (2012) On the use of weather data in ecological studies along altitudinal and
latitudinal gradients. Oikos 121(1):3-19.
Xia J, et al. (2014) Terrestrial carbon cycle affected by non-uniform climate warming. Nature
Geosci 7(3):173-180.
Ward SE, et al. (2013) Warming effects on greenhouse gas fluxes in peatlands are modulated
by vegetation composition. Ecology Letters 16:1285-1293.
ESRL (Solar Calculation Details. (Earth System Research Laboratory ).
Stull RB (1991) An Introduction to Boundary Layer Meterology (Kluwer, Dordrecht).
14
15
Figure Captions
Figure 1. Turbine proximity influences observed effects of wind farm operation on TA and
AH. The effect of distance (x) from the nearest turbine on the temperature and AH departure
during the night (a) and day (b). Positive departures occur up to 200 m from the nearest
turbine. Blue dots represent the temperature departure difference for ON-OFF periods and red
triangles the AH departure. These data can be approximated by ΔT = 0.62 - 0.12 ln (x), r =
0.75, ΔAH = 0.15 - 0.03 ln (x), r = 0.61 during the night and ΔT = -0.11 + 0.02 ln (x), r =
0.53 during the day, with dotted blue lines representing temperature and solid red lines AH.
The daytime AH cannot be approximated by a logarithmic function (r < 0.2). Black dotted
lines represent the zero departure and 200 m-distance baselines.
Figure 2. Diurnal variations in TA, TSU and TSO and AH differ during wind farm operational
and idle periods. Data points represent the average coefficient of variation (CV) ± standard
error (SE) of the TA (a), TSU (b), TSO (c) and AH (d) during the ON (blue dashed lines) and
OFF (solid red lines) periods. Fraction of the day was calculated based on sunrise and sunset
times (see Methods), with 0 representing sunrise and 0.5 sunset and 0.04 approximately 1
hour, with sunset to sunrise shaded grey. There were significant differences in the variation
during ON and OFF periods for each hour (p<0.05 for pseudo 10:00 and 11:00 for the TSO
data and pseudo 07:00 for the AH data, p<0.01 for all other data and hour intervals).
Differences in soil moisture were also investigated but no significant effects found (see SI).
Figure 3. Night-time TA and AH departures during wind farm operational and idle periods.
Grey lines represent the Black Law Wind Farm road network, black squares turbines and
coloured circles the mean TA (a) and AH (b) departures for the ON period; and mean TA (c)
and AH (d) departures for the OFF period. During the ON period the TA was warmer and AH
16
moister in the D group compared with the ND group, indicated by the concentration of
yellow to red circles on plot (a) and darker blue circles on plot (b) compared with plots (c)
and (d) respectively. During the ON period, TA (a) and AH (b) departures in D and ND areas
were significantly different (p < 0.01 for TA and p < 0.05 for AH), whereas there were no
significant differences during the OFF period, plots (c) and (d) (p > 0.10). See Fig. S1 for
illustration of D and ND groups. Examination of day-time effects
17
Fig. 1a.
Fig.1b
18
Fig. 2a
Fig. 2b
Fig. 2c
19
Fig. 2d
Fig. 3a
20
Fig. 3b
21
Fig. 3c
22
Fig. 3d
23