Sentinel-2 true-colour image of Merlin Woods and Doughiska, 26 September 2026, showing felled patches in the dark conifer canopy

Felled patches

Sentinel-2 and Sentinel-1 · March to October 2026

Merlin Woods from orbit

What optical and radar satellites show of the conifer clearance in Doughiska, Galway

Contains modified Copernicus Sentinel data 2026

1 / 13

The story so far

From storm damage to felling

1

2

3

4

5

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

2025

Sentinel-2

26 Sep: latest clear pass

solid = clear, hollow = partly cloudy; shaded = felling works

no clear optical scene

1 · Jan 2025

Storm Eowyn damages the Norway spruce, about half the woodland

2 · 8 Jul

Council announces a restoration plan and seeks NeighbourWood funding

3 · 26 Aug

Plan published: fell and replant 3.5 ha; 54% of the woodland kept

4 · 9 Sep

Clearance works begin; squirrel feeders go in

5 · 17 Sep

Friends of Merlin Woods call for a pause

Sources: see final slide

2 / 13

Where it is

Merlin Woods, between the hospital and Doughiska

Sentinel-2 true colour, 13 July 2026: Merlin Woods outlined, north of the Galway–Dublin railway, west of Doughiska

Merlin Woods

Merlin Park Hospital

Doughiska

Railway to Dublin

Detail on next slides

58.5 ha

mapped as Merlin Woods in OpenStreetMap

46.5 ha

of tree canopy inside that outline, measured from the imagery. The western 5.7 ha lies just off this satellite tile

8%

of that canopy is the 3.5 ha the council plans to fell, so the reported “46% of the wooded area” must refer to a smaller area

Outline © OpenStreetMap contributors (ODbL) · imagery 13 July 2026

3 / 13

Before and during felling

The northern block, July and September

The northern block on 13 July 2026: continuous dark-green conifer canopy

13 July

The same area on 26 September 2026: brown patches of bare ground where canopy was felled

26 September

Unbroken dark-green conifer canopy before the works.

Brown bare ground where stands have been felled, with works still under way.

S2C vs S2A, orbits R123 / R023

4 / 13

Every clear view

Six cloud-free scenes, July to September

Animation stepping through 6 cloud-free Sentinel-2 scenes of the northern block from 13 July to 26 September 2026; brown felled patches appear in the later frames

13 July · before

16 July · before

18 July · before

7 August · before

21 September · felling under way

26 September · felling under way

Only scenes with at least 95% of the wood cloud-free are shown; cloudy passes are skipped.

Sentinel-2 true colour, same colour stretch in every frame

5 / 13

Where the canopy went

Five felled patches, 2.22 ha in all

26 September 2026 with 5 detected felled patches outlined in orange

B

A

F

C

D

PatchHectaresGreenness change
B1.01−0.21
A0.75−0.18
F0.28−0.17
C0.12−0.19
D0.06−0.23
Total2.22

Greenness change is the drop in NDVI from the June–July baseline to 26 September, after removing the wood-wide seasonal shift.

Detection: NDVI and moisture index both fall; 10 m pixels

6 / 13

Progress by 26 September

Most of the plan so far

2.22 ha

felled of the 3.5 ha planned (63%)

range 1.22–3.34 ha across detection thresholds

0 ha

2.22 ha felled

3.5 ha planned

By clear pass: 21 Sep 1.24 ha · 26 Sep 2.22 ha

Plan figure: Galway Daily, 26 August 2026

7 / 13

The signal over time

Only the felled patches lose greenness and moisture

Greenness (NDVI)

0.5

0.7

0.9

Mar

May

Jul

Sep

0.38
felled patches

0.47
rest of canopy

In March, before leaf-out, these stands were greener than the rest of the wood: the mark of evergreen conifer.

Canopy moisture (NDMI)

0.0

0.2

0.4

Mar

May

Jul

Sep

0.30
felled patches

0.39
rest of canopy

Felled patches (2.22 ha)

Rest of the canopy (44.3 ha)

Felling under way (from 9 Sep)

Median per group, clear passes only; each dot is one satellite pass

8 / 13

Sentinel-1 radar · a pass every two days

Radar sees the felled patches brighten

−2

−1

0

+1

+2

Jun

Jul

Aug

Sep

Oct

Sentinel-2

normal pass-to-pass range

+1.0 dB
on 8 Oct

felling under way

Felled patches minus intact canopy, dB (one dot per pass; line = 3-pass mean)

±2 sd, June–August

Sentinel-1 GRD, VV and VH averaged, three orbits, each normalised to its own June–August baseline

9 / 13

What the radar adds

Cloud-proof, but blunter than optical

COVERAGE

Sees through cloud

Three Sentinel-1 orbits cover the wood; since July a usable pass has come every 2 days, filling the optical gap from 7 August to 21 September.

2 days

median gap between passes

SIGNAL

Brighter, not darker

Freshly felled stems, brash and newly exposed forest edges scatter radar strongly. A drop usually follows only once the site is cleared.

+1.0 dB

felled patches, 8 October

CHECK

Not the weather

The 5 October pass came after 0.3 mm of rain in 12 hours. The wettest pre-works pass (3 August, 7.3 mm) is at -1.1 sd.

0.3 mm

rain in the 12 h before 5 Oct

Rainfall: Open-Meteo modelled hourly precipitation for the wood

10 / 13

Method

How the felling was measured

STEP 1

Map the tree canopy

Inside the Merlin Woods outline, keep pixels that are dark in near-infrared in March (not grass) and green all summer.

46.5 ha

canopy observed

STEP 2

Compare before and after

Median of four clear June–July passes against 26 September. Greenness and moisture must both fall, after removing the wood-wide seasonal shift.

4 → 1

clear passes compared

STEP 3

Test for false alarms

Run the same test on two July passes, before any works. Drop specks under 4 pixels and hedgerow slivers beside mown fields.

0.00 ha

flagged by the no-works control

Sentinel-2 L2A surface reflectance · 10 m bands B04, B08; 20 m B11

11 / 13

Caveats

What the imagery cannot tell us

10 m pixels.

Small openings along paths, and single trees, are below what this can resolve reliably. The smallest patch here is 6 pixels.

Two clear optical passes since works began.

Felled trees left lying, or brash, can still read partly green, so the area may be undercounted. Later passes firm it up.

No clear optical scene from 7 August to 21 September.

Radar passes every two days fill the gaps, but radar alone cannot map patches this small.

Species are inferred, not seen.

Conifer is read from winter greenness; the satellite does not identify Norway spruce directly.

Storm damage predates the imagery.

Any openings left by Storm Eowyn in January 2025 were already there in March 2026, so they sit in the baseline and are not counted.

Merlin Woods from orbit

12 / 13

Sources

Press coverage and data

Galway Bay FM, 8 Jul 2026
City Council begins plans for removal of fallen trees in Merlin Woods

Galway Daily, 26 Aug 2026
Plans to restore Merlin Woods with native trees

Galway Bay FM, 28 Aug 2026
City Council facing criticism for Merlin Woods felling plan

Galway Bay FM, 9 Sep 2026
Squirrel feeders being installed in Merlin Woods as clearance works underway

RTÉ News, 17 Sep 2026
Call to pause Merlin Woods felling over ecosystem impact

Imagery
Contains modified Copernicus Sentinel data 2026. Sentinel-2 Level-2A, tile T29UNV; Sentinel-1 IW GRD, orbits 23, 74 and 96.

Woodland outline
© OpenStreetMap contributors, ODbL. Relation 14174436, “Merlin Woods”.

Analysis
Python with GDAL and NumPy. Rainfall: Open-Meteo. Thresholds and controls are in the speaker notes.

Merlin Woods from orbit

13 / 13