All work
Understanding history · Change detection

Where the snow retreats, the mountains green

A satellite time-series read of Lebanon's high mountains: as snow cover has steadily retreated over 36 years, vegetation has advanced up the same slopes: the NDVI and NDSI analysis behind ORG's Land Degradation Neutrality diagnostic.

The same mountains, two indices: vegetation gaining (left) where snow is retreating (right). The greening and the snow loss trace the same high-mountain spine.

The same mountains, two indices: vegetation gaining (left) where snow is retreating (right). The greening and the snow loss trace the same high-mountain spine.

01The idea

ORG was preparing a Land Degradation Neutrality (LDN) diagnostic for the High Mountains of Lebanon: the terrain above 1,500 m. My part was the remote-sensing analysis: use the satellite record to test what was happening to vegetation and snow across the range, and whether the two were connected. A comparison of 1998 and 2017 land-cover maps hinted that the mountains were gaining vegetation; the question was whether that was real, and if so, why.

02The analysis

I built the whole thing on the Landsat archive through Google Earth Engine:

  • Vegetation (NDVI): a 2000–2020 median-composite baseline (only scenes under 20% cloud), then each year's percentage deviation from it, classified into seven bands and mapped nationwide. I vectorised the result to quantify vegetation gain and loss per municipality above 1,500 m.
  • Snow (NDSI): a 36-year time series (1985–2021) of the Normalized Difference Snow Index (Landsat 5 → 7 → 8, January–June each year) to measure the long-term trend in snow cover.

03What it shows

The two records move in opposite directions. Snow cover shows a clear, sustained decline across the 36 years, while vegetation shows an equally clear rise.

Snow cover (NDSI) falling and vegetation (NDVI) rising over the same period: the inverse trends that framed the question.

Snow cover (NDSI) falling and vegetation (NDVI) rising over the same period: the inverse trends that framed the question.

Mapped in space, the coupling is unmistakable: the vegetation gain and the snow loss trace the same high-mountain spine. Where the snow line has retreated above 1,500 m, vegetation has moved in behind it: the classic signature of a warming mountain.

NDVI deviation from the 2000–2020 average across Lebanon. Green marks vegetation gain, red degradation; the strongest greening runs along the high-mountain spine.

NDVI deviation from the 2000–2020 average across Lebanon. Green marks vegetation gain, red degradation; the strongest greening runs along the high-mountain spine.

Departure from average snow cover (NDSI) in the High Mountains. Red marks snow loss, concentrated along the same spine that is greening.

Departure from average snow cover (NDSI) in the High Mountains. Red marks snow loss, concentrated along the same spine that is greening.

It gave the diagnostic a defensible, evidence-based answer: the greening is real, it concentrates at altitude, and it is bound up with the retreat of the mountain snowpack: a signal worth monitoring closely.

36
years of Landsat record analysed (1985–2021)
1,500 m
elevation above which the greening concentrates
7
NDVI deviation classes mapped across Lebanon
2
indices moving in opposite directions: snow & vegetation
Next project
Weighing five layers to site solar farms →