Note
Go to the end to download the full example code.
Canary Islands: nine PVDeg degradation-mechanism maps#
This complements the 100 m thermal map with hourly-weather mechanism screening at eleven locations across eight islands. PVGIS supplies modeled TMY irradiance, humidity, temperature and wind, selected from 2005-2023. Every result below is recomputed offline with PVDeg 0.7.4.
The maps cover thermal, damp-heat, broadband light-assisted and combined stress; module-surface, front/back encapsulant and backsheet moisture; and solder fatigue. They are scenario metrics, not annual power loss or remaining lifetime. Do not add the metrics together: they have different units and overlapping physics.
Markers show the actual sampling locations. The coastlines use the existing
100 m climate masks for context only. No humidity/irradiance downscaling or spatial
interpolation is performed. These are not 100 m degradation predictions.
See the guide for parameters, coverage,
missing mechanisms and reproducible input acquisition. Inputs and metadata are
available in the repository’s examples/analysis/pvdeg_canary_weather directory.
from pathlib import Path
import matplotlib.pyplot as plt
import numpy as np
import rasterio
from matplotlib.colors import ListedColormap
from rasterio.plot import plotting_extent
from rasterio.warp import transform
from energy_analysis.pvdeg_mechanisms import METRICS, analyze_manifest
Calculate every mechanism using hashed hourly inputs#
results = analyze_manifest(Path("pvdeg_canary_weather"))
print(results.round(3).to_string(index=False))
print("Weather: modeled PVGIS TMY; material / module parameters: scenarios")
print("No calibrated annual degradation rate, UV damage, PID or LeTID prediction")
x, y = transform("EPSG:4326", "EPSG:32628", results.longitude.tolist(), results.latitude.tolist())
name longitude latitude thermal damp_heat photothermal combined surface_rh front_rh back_rh backsheet_rh solder_fatigue
El Hierro -17.980 27.750 1.134 1.119 0.542 0.350 60.587 54.941 4.300 32.443 0.000
La Palma -17.850 28.680 0.810 0.956 0.268 0.216 66.665 63.356 59.331 62.998 0.000
La Gomera -17.240 28.110 1.135 1.062 0.529 0.309 59.796 53.784 58.132 58.964 0.000
Tenerife south -16.590 28.080 1.253 1.108 0.663 0.356 61.477 52.771 55.514 58.496 2.684
Tenerife north -16.520 28.390 0.938 0.731 0.475 0.201 61.773 49.282 55.405 58.589 0.000
Teide -16.644 28.272 1.142 0.710 0.728 0.237 60.013 43.053 44.214 52.114 10.169
Anaga -16.230 28.550 1.015 1.015 0.418 0.270 59.630 56.063 4.396 32.013 0.000
Gran Canaria -15.600 28.000 1.254 0.850 0.728 0.261 58.795 44.056 41.504 50.149 8.270
Fuerteventura -14.010 28.430 1.281 1.010 0.678 0.308 57.165 47.235 42.378 49.772 5.925
Lanzarote -13.600 29.040 1.035 1.047 0.468 0.307 61.563 56.653 49.192 55.377 0.000
La Graciosa -13.510 29.250 1.037 1.067 0.442 0.289 62.739 57.359 50.815 56.777 0.000
Weather: modeled PVGIS TMY; material / module parameters: scenarios
No calibrated annual degradation rate, UV damage, PID or LeTID prediction
Nine separate maps with their own units#
Numbered sites can be identified in the following table. Gray land is context, not a computed degradation surface. Color is only assigned to sampled points.
land = []
for path in sorted(Path(".").glob("pvdeg_canary_*.tif")):
with rasterio.open(path) as dataset:
mask = dataset.read(1, masked=True)
land.append(
(
np.ma.array(np.ones(mask.shape), mask=mask.mask),
plotting_extent(mask, dataset.transform),
)
)
for group in (list(METRICS)[:3], list(METRICS)[3:6], list(METRICS)[6:]):
fig, axes = plt.subplots(3, 1, figsize=(12, 15), layout="constrained")
for ax, key in zip(axes, group, strict=True):
title, unit, cmap = METRICS[key]
ax.set_facecolor("#edf6fb")
for mask, extent in land:
ax.imshow(
mask, extent=extent, cmap=ListedColormap(["#d4dbdf"]), interpolation="nearest"
)
artist = ax.scatter(
x, y, c=results[key], cmap=cmap, s=100, edgecolors="black", linewidths=0.6, zorder=3
)
for i, (east, north) in enumerate(zip(x, y, strict=True), start=1):
ax.annotate(
str(i),
(east, north),
xytext=(5, 7),
textcoords="offset points",
fontsize=8,
zorder=4,
)
ax.set(
xlim=(160000, 690000),
ylim=(3040000, 3290000),
title=f"{title} · hourly TMY scenario · sampled sites",
xlabel="UTM 28N easting (m)",
ylabel="UTM 28N northing (m)",
)
fig.colorbar(artist, ax=ax, label=unit, shrink=0.9)
Site comparison and interpretation#
The exposure ratios use 60 kJ/mol and a constant 25 C / 50% RH / 1000 W/m² reference. The light term uses broadband POA, not a UV spectrum. Moisture uses upstream EVA W001 / PET W017 defaults. Solder fatigue uses the upstream hourly calibration; it is not a percentage of failed joints or module power.
fig, ax = plt.subplots(figsize=(12, 6), layout="constrained")
ax.axis("off")
columns = ["thermal", "damp_heat", "photothermal", "combined", "solder_fatigue"]
labels = [f"{i}. {name}" for i, name in enumerate(results.name, start=1)]
table = ax.table(
cellText=results[columns].round(3).to_numpy(),
rowLabels=labels,
colLabels=["Thermal", "Damp heat", "Light + heat", "Combined", "Fatigue (kPa)"],
loc="center",
cellLoc="center",
)
table.auto_set_font_size(False)
table.set_fontsize(9)
table.scale(1, 1.7)
ax.set_title("Site lookup · exposure ratios and solder-fatigue scenario metric", pad=25)
plt.show()

Total running time of the script: (0 minutes 3.428 seconds)


