Slopes rarely out-warm the floor a 2003–2025 MODIS assessment of deep surface-based night-time inversions over the Malang basin, East Java
Abstract
Nocturnal cold-air pools can trap heat and pollutants in basin cities, but tropical basins are seldom monitored. We assessed deep surface-based night-time inversions over the Malang basin, East Java, Indonesia, from 2003–2025 MODIS Terra and Aqua night-time land surface temperature (LST), using the difference ΔT between a slope band (900–1300 m; mean 1079 m) and the rural basin floor (≤ 600 m; mean 449 m). Of 8395 nights, 2017 had sufficient clear-sky coverage. The slope band was colder than the floor on 98 % of clear nights (median ΔT −3.1 °C, against −4.1 °C for a standard atmosphere). ΔT > 0 occurred on 1.7 % of clear nights (35 nights; 95 % confidence interval 1.3–2.4 %) and on 1.3–2.2 % under a view-angle limit, removal of valley-floor pixels from the slope band and a fixed-pixel anomaly index, with median ΔT changing by less than 0.3 °C. Night-to-night variation in ΔT was not reproduced by the two overpasses about three hours apart (Spearman ρ = 0.05 on 351 shared nights), implying a per-overpass error of about 1 °C, as large as the night-to-night spread. A noise-only model accounts for the nights with ΔT > 0 observed by Terra alone or by both sensors, but not for those observed by Aqua alone (29 observed, 3–14 expected, a lower bound because the error estimate comes from the clearest nights), which include all 14 nights with ΔT > 1 °C. These nights were cloudy at ~22:30 and clear at ~01:30, wetter than average, and had an anomalously cold floor (median −4.5 °C) beneath near-normal slopes, consistent with post-rain cold pools, fog or residual cloud. After season adjustment, nights with ΔT > 0 were not calmer by ERA5-Land wind and only weakly wetter. Their frequency rose after 2015 in the Aqua record only. The night-time surface urban heat island was 0.9 °C and did not affect the result. At the MODIS overpass times and 1 km scale, slopes 630 m above the floor rarely out-warm it; MODIS LST constrains the climatology of this contrast but does not identify individual inversion nights.
Related
- Vertical distribution of pollution trapped by wintertime surface inversions in Fairbanks, Alaska, during the ALPACA experiment
- Probabilistic seismic hazard analysis based on microtremor data: A site effect study for earthquake disaster mitigation at Malang Regency, Indonesia
- Comprehensive assessment of historical surface air temperature simulations over iran using 51 CMIP6 models
- Two large unstable slopes in West Greenland illustrates an emerging Arctic tsunami hazard
- Mapping rockfall hazard and detecting precursory damage in rock slopes with passive seismic: Lessons from the La Praz case-study
BibTeX
@misc{wijaya2026slopes,
title = {{Slopes rarely out-warm the floor a 2003–2025 MODIS assessment of deep surface-based night-time inversions over the Malang basin, East Java}},
author = {Ryan Andri Wijaya},
howpublished = {EarthArXiv},
year = {2026},
month = oct,
doi = {10.31223/x5p222},
url = {https://eartharxiv.org/repository/view/15346/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Slopes rarely out-warm the floor a 2003–2025 MODIS assessment of deep surface-based night-time inversions over the Malang basin, East Java
AU - Ryan Andri Wijaya
PY - 2026
DA - 2026/10/03
PB - EarthArXiv
DO - 10.31223/x5p222
UR - https://eartharxiv.org/repository/view/15346/
AB - Nocturnal cold-air pools can trap heat and pollutants in basin cities, but tropical basins are seldom monitored. We assessed deep surface-based night-time inversions over the Malang basin, East Java, Indonesia, from 2003–2025 MODIS Terra and Aqua night-time land surface temperature (LST), using the difference ΔT between a slope band (900–1300 m; mean 1079 m) and the rural basin floor (≤ 600 m; mean 449 m). Of 8395 nights, 2017 had sufficient clear-sky coverage. The slope band was colder than the floor on 98 % of clear nights (median ΔT −3.1 °C, against −4.1 °C for a standard atmosphere). ΔT > 0 occurred on 1.7 % of clear nights (35 nights; 95 % confidence interval 1.3–2.4 %) and on 1.3–2.2 % under a view-angle limit, removal of valley-floor pixels from the slope band and a fixed-pixel anomaly index, with median ΔT changing by less than 0.3 °C. Night-to-night variation in ΔT was not reproduced by the two overpasses about three hours apart (Spearman ρ = 0.05 on 351 shared nights), implying a per-overpass error of about 1 °C, as large as the night-to-night spread. A noise-only model accounts for the nights with ΔT > 0 observed by Terra alone or by both sensors, but not for those observed by Aqua alone (29 observed, 3–14 expected, a lower bound because the error estimate comes from the clearest nights), which include all 14 nights with ΔT > 1 °C. These nights were cloudy at ~22:30 and clear at ~01:30, wetter than average, and had an anomalously cold floor (median −4.5 °C) beneath near-normal slopes, consistent with post-rain cold pools, fog or residual cloud. After season adjustment, nights with ΔT > 0 were not calmer by ERA5-Land wind and only weakly wetter. Their frequency rose after 2015 in the Aqua record only. The night-time surface urban heat island was 0.9 °C and did not affect the result. At the MODIS overpass times and 1 km scale, slopes 630 m above the floor rarely out-warm it; MODIS LST constrains the climatology of this contrast but does not identify individual inversion nights.
ER -