As spatial scale increases, z energy increases following a quadratic
(scale2, β = -13.96, SE= 0.16, p <0.0001
| scale, β= 24.12, SE= 1.15, p<0.0001); see Figure 3.
Compared to the null scans, lapwing nest surrounds possessed higher 3D
variation across all spatial scales (nest, β = 2.90, SE=0.12, p=0.00379)
and variation increased with spatial scale at a faster rate for nest
sites at the smaller spatial scales (nest:scale2, β =
2.621, SE=0.04, p= 0.0089 | nest:scale, β = -2.029, SE=0.13, p=
0.042). Post hoc comparison of site management strategies showed the
nests of sheep grazed fields had significantly lower 3D variation
compared to other sites, while wet grassland sites had significantly
higher 3D variation (see supplemental material) For scales below the
size of the clutch, 3D energy originates from deviation in height
between small vegetation (grasses) or from the substrate (large stones,
gravel). The 3D energy of pastoral nest sites at higher spatial scales
was more similar to those of the arable sites than their null sites,
except for at sheep grazed sites. At scales above the size of the nest,
high energy results from large clumps/mounds of weedy vegetation,
trampling and sloping terrain (hills). On average, clutches were
elevated 4.5cm above their local surroundings. There was no significant
difference between management type and nest elevation; nest elevation
was instead predicted by 3D energy of the surroundings (energy :
elevation, β = 2.894, SE= 53.816, p = 0.00493).
Nest Predation
Over the 2 years, we sampled the Avon Valley and Sussex Sites we
photographed 115 lapwing nests, 86 of which were scanned. Of the nests
found, 13 (8 in 2021, 5 in 2022) were predated. The proportion of nests
predated varied widely between county and site, with no predation events
of scanned nests recorded in the Sussex sites. Though nest predation of
unscanned nests did occur. Predation was the most common cause of nest
failure, followed by abandonment. On average, predated nests had poorer
colour matches and lower surrounding luminance complexity (SD). However,
no significant result was observed and none of the camouflage metrics
used were able to predict nest failure from predation (see supplementary
material).
Clutch Occlusion &
Camouflage:
The percentage visibility (un-occluded) of the clutch (eggs only)
increases with the observer viewing angle in a sigmoid fashion. On
average, a viewing angle of 15o elevation (equivalent
Horizontal Distance: Fox 1.5m, corvid [6.0m, 11.9m, 23.9m, 47.8m,
95.5m]) is required for 25% visibility and an angle of
27o (Horizontal Distance: Fox 0.8m, corvid [3.14m,
6.2m, 12.6m, 25.1m, 50.2m]) to see 50% (Figure 4). Increased 3D
energy across spatial scales increases nest occlusion at low viewing
angles (10 o -30 o), particularly at
spatial scales below the clutch size, with the lower scales to the
clutch having the largest effect (scale of grasses) on occlusion (See
Supplemental Material).
The JND colour and luminance difference of the clutches from the local
surround are in line with those of highly camouflaged animals (fox, lum
ΔS μ 1.10 ± 0.02 SE | col ΔS μ 0.85 ± 0.02 SE) (Corvid, Lum ΔS,
μ 0.9 ± 0.02SE | Col ΔS μ 1.58 ± 0.02 SE). Clutches were of a
better colour match to bare crop and fallow sites as opposed to
vegetated wet grassland sites for both visual systems (Sussex-Hampshire
: corvid colour ΔS β= -6.33, SE= 0.87, p < 0.0001)
(Sussex-Hampshire : fox colour ΔS, β= -7.43, SE= 0.80, p <
0.0001). ΔS colour and ΔS luminance follow a negative exponential with
increasing viewing angle (Figure 4). For a decrease in viewing angle to
drop ΔS colour and/or ΔS luminance by just 0.1 JND, the clutch would
already be 75% occluded from most viewing heights. The exceptions were
for corvid vision from a height of 12.8m (22.5 o for
-0.1 JND) and 25.6m (32.5 o for -0.1 JND).