Standardizing Landsat park cooling contrasts for reliable prioritization of urban green spaces

Satellite land surface temperature (LST) is often used to compare urban green spaces for heat mitigation planning, but park cooling rankings depend on how the park and non-park comparator are defined. This study tested the stability of a Landsat park to buffer thermal contrast (PBTC) for 13 parks and forested green spaces in Kraków, Poland. Four quality screened late spring Landsat scenes from 1990, 2000, 2013 and 2018 were analyzed using fixed cores, two reference rings, quality assurance masking, aggregation to nominal Landsat thermal support, balancing and fixed-effects models. Observed PBTC was positive in all scenes and decreased after standardization for impervious context, elevation and slope. Paired site-year bootstrap tests showed that observed PBTC exceeded context-standardized PBTC in every distance-by-year contrast, while site rankings also changed after standardization. NDMI conditioning reduced PBTC further and is reported as a diagnostic rather than the primary ranking metric. The delivered 30 m LST grid produced higher PBTC than the support-aggregated workflow. These findings concern surface thermal contrasts only and should not be read as air temperature cooling. PBTC can support screening and prioritization only when the park core, reference surface, Landsat support, quality mask, conditioning variables and weighting rule are reported with the value.