Great difficulty in characterizing the spatial pattern and scale of regional surface urban heat islands (RSUHIs) over urban agglomerations is mainly attributed to the RSUHI extent exceeding urban boundaries and strong variations in the background climate state and land cover. A new method that successfully addresses these problems by applying the spatial similarity of the land surface temperature (LST) annual cycle in urban agglomerations is proposed to characterize the RSUHI spatial pattern and scale in the Yangtze River Delta of China. This method has good adaptability to complex terrain and climate variations and simultaneously identifies the RSUHI spatial pattern and scale. The RSUHI spatial extent averaged 3.7 × 10⁴ km² in 2003–2020, which is 1.5–2.0 times the urban extent, and had a fluctuating but increasing tendency. Rapid urbanization and the spatial pattern‐shifting LST anomalies were mainly responsible for the fluctuating RSUHI extent during 2003–2020.

Surface urban heat islands (SUHIs) can extend beyond the urban boundaries and greatly affect the thermal environment of continuous regions over an agglomeration. Traditional urban-rural dichotomy depending on the built-up and non-urban lands is challenged in characterizing regional SUHIs, such as how to accurately quantify the intensity, spatial pattern, and scales of SUHIs, which are vulnerable to SUHIs, and what the optimal scale for conducting measures to mitigate the SUHIs. We propose a machine-learning-assisted solution to address these problems based on the thermal similarity in the Yangtze River Delta of China. We first identified the regional-level SUHI zone of approximately 42,328 km² and 38,884 km² and the areas that have no SUHI effects from the annual cycle of land surface temperatures (LSTs) retrieved from Terra and Aqua satellites. Defining SUHI as an anomaly on background condition, random forest (RF) models were further adopted to fit the LSTs in the areas without the SUHI effects and estimate the LST background and SUHI intensity at each grid point in the SUHI zone. The RF models performed well in fitting rural LSTs with a simulation error of approximately 0.31 °C/0.44 °C for Terra/Aqua satellite data and showed a good generalization ability in estimating the urban LST background. The RF-estimated daytime Aqua/SUHI intensity peaked at approximately 6.20 °C in August, and the Terra/SUHI intensity had two peaks of approximately 3.18 and 3.81 °C in May and August, with summertime RF-estimated SUHIs being more reliable than other SUHI types owing to the smaller simulation error of less than 1.0 °C in July–September. This machine-learning-assisted solution identified an optimal SUHI scale of 30,636 km² and a zone of approximately 23,631 km² that is vulnerable to SUHIs, and it provided the SUHI intensity and statistical reliability for each grid point identified as being part of the SUHI. Urban planners and decision-makers can focus on the statistically reliable RF-estimated summertime intensities in SUHI zones that have an LST annual cycle similar to that of large cities in developing effective strategies for mitigating adverse SUHI effects. In addition, the selection of large cities might strongly affect the accuracy of identifying the SUHI zone, which is defined as the areas that have an LST annual cycle similar to large cities. Water bodies might reduce the RF performance in estimating the LST background over urban agglomerations.

Compared with atmospheric urban heat islands, surface urban heat islands (SUHIs) are easily monitored by the thermal sensors on satellites and have a more stable spatial pattern resembling the urban and built-up lands across single cities, large metropolitans, and urban agglomerations; hence, they are gaining more attention from scholars and urban planners worldwide in the search for reasonable urban spatial patterns and scales to guide future urban development. Traditional urban–rural dichotomies, being sensitive to the representative urban and rural areas and the diurnal and seasonal variations in the land surface temperature (LST), obtain inflated and varying SUHI spatial footprints of approximately 1.0–6.5 times the urban size from different satellite-retrieved LST datasets in many cities and metropolitan areas, which are not conducive to urban planners in developing reasonable strategies to mitigate SUHIs. Taking the Yangtze River Delta urban agglomeration of China as an example, we proposed an improved structural similarity index to quantify more reasonable spatial patterns and footprints of SUHIs from multiple LST datasets at an annual interval. We identified gridded LST anomalies (LSTAs) related to urbanization by adopting random forest models with climate, urbanization, geographical, biophysical, and topographical parameters. Using a structural similarity index of the LSTA annual cycle at a grid point relative to the urban reference LSTA annual cycle in terms of average values, variances, and shapes to characterize the SUHIs, cross-validated SUHI footprints ~1.06–2.45 × 10⁴ km² smaller than the urban size and clear transition zones between urban areas and the SUHI zone were obtained from multiple LST datasets for 2000–2022. Hence, urban planners can balance urbanization’s benefits with the adverse effects of SUHIs by enhancing the transition zone between urban areas and the SUHI zone in future urban design. Considering that urban areas rapidly transformed into SUHIs, with the ratio of the SUHI extent to the urban size increasing from 0.43 to 0.62 during 2000–2022, urban planners should also take measures to prevent the rapid expansion of high-density urban areas with an ISA density above 65% in future urban development.

Accurately characterizing grid‐level vegetation dynamics in urban agglomerations remains challenging due to high spatial heterogeneity and complex temporal responses to urbanization. We developed a threshold‐free clustering framework that integrates the LinCoIndex and structural similarity (SSIM) index to address this issue. Applied to the Yangtze River Delta urban agglomeration in China, it reveals three findings that challenge conventional threshold approaches: (a) L‐shaped urbanization‐driven browning in new urban areas versus V‐shaped policy‐linked greening in urban cores, with a browning‐to‐greening transition around 2009–2011; (b) low‐vegetation footprints occupying only 30%–58% of urbanized area while impervious surface plus annual peak vegetation exceeds 100%; (c) above‐impervious canopy increasing from 10% to 48% through tree planting and high‐rise building restrictions, enabling the region to meet China's 43% urban vegetation coverage target. These results indicate that coordinated greening interventions can offset net vegetation loss under sustained urbanization pressure.

Analysis of ground-based and remotely retrieved precipitation data reveals that heavy Meiyu precipitation events (HMPEs) produce a relatively independent rain-belt over eastern China. A rotating calipers algorithm is applied to quantify the spatial scales of HMPEs. We find that HMPEs have regular spatial scales with an average length, width and extent of about 1400 km, 500 km and 40.00 × 104 km2, respectively, through a comprehensive assessment of different types of HMPE, illustrating that HMPEs have a size similar to that of the sub-synoptic-scale Meiyu front (1500–2000 km). Convective activities along the Meiyu front zone and the upper westerly jet stream strongly affect the position and orientation of rain-belts of HMPEs. The Meiyu front zone, strong vertical motions and large transport of warm moisture have a comparable spatial scale to the HMPE rain-belts over eastern China.