Surface deformation associated with continental earthquake ruptures includes localized deformation on the faults, as well as deformation in the surrounding medium though distributed and/or diffuse processes. However, the connection of the diffuse part of the surface deformation to the overall rupture process, as well as its underlying physical mechanisms are not yet well understood. Computing high-resolution optical image correlations for the 2021/05/21 Mw7.4 Maduo, Tibet, rupture, we highlight a correlation between the presence of faults and fractures at the surface, and variations in the across-fault displacement gradient, fault zone width, and amplitude of surface displacement. We show that surface slip along primary faults is systematically associated with gradients greater than 1%, and is dominant in regions of greater coseismic surface displacement. Conversely, the diffuse deformation is associated with gradients ≤0.3%, and is dominant in regions of lesser surface displacement. The distributed deformation then occurs for intermediate gradients of 0.3-1%, and at the transition between the localized and diffuse deformation regions. Such patterns of deformation are also described in laboratory experiments of rock deformation, themself supported by field observations. Comparing these experiments to our observations, we demonstrate that the diffuse deformation along the 2021 Maduo rupture corresponds to kilometer-wide plastic yielding of the bulk medium occurring in regions where surface rupture is generally missing. Along the 2021 Maduo rupture, diffuse deformation occurs primarily in the epicentral region, where the dynamic stresses associated with the nascent pulse-like rupture could not overcome the shallow fault zone frictional strength.

We develop an algorithm to integrate GPS and InSAR data for a 3-dimensional crustal deformation field at the Earth’s surface. In the algorithm discrete GPS data points are interpolated to obtain a 3-dimensional continuous velocity field, which is then combined with the InSAR line-of-sight (LOS) velocity data pixel by pixel using the least-squares method. Advantages of our method over previous ones are that: 1) The GPS data points are optimally interpolated by balancing a trade-off between spatial resolution and solution stability. 2) A new algorithm is developed to estimate realistic uncertainties for the interpolated GPS velocities, to be used as weights for GPS data in GPS-InSAR combination. 3) Realistic uncertainties for the InSAR LOS rate data are estimated and used as weights for InSAR data in GPS-InSAR combination. 4) The ramps and/or offsets of the InSAR data are globally estimated for all the images to minimize data misfit, particularly at regions where the data overlaps. Application of this method to real data from southern California shows its capability of successfully restoring 3-dimensional continuous deformation field from spatially limited GPS and dimensionally limited InSAR data. The deformation field reveals water withdrawal induced subsidence and drought caused uplift at various regions in southern California.

California’s Central Valley is responsible for $17 billion of annual agricultural output, producing 1/4 of the nation’s food. However, land in the Central Valley is sinking at a rapid rate (as much as 20 cm per year) due to continued groundwater pumping. Land subsidence has a significant impact on infrastructure resilience and groundwater sustainability. It is important to understand subsidence and groundwater depletion in a consistent framework using improved models capable of simulating in-situ well observations and observed subsidence. Currently, groundwater well data is sparse and sampled irregularly, compromising our understanding of groundwater changes. Moreover, groundwater pumping data is a major missing piece of the puzzle. Limited data availability and spatial/temporal uncertainty in the available data have hampered understanding the complex dynamics of groundwater and subsidence. To address this limitation, we first integrated multimodal data including InSAR, groundwater, precipitation, and soil composition by interpolating data with the same spatial and temporal resolutions. We then identified regions with different temporal dynamics of land displacement, groundwater depth, and precipitation. Some areas (e.g., Helm) with coarser grain soil compositions exhibited potentially reversible land transformations (elastic land compaction). Finally, we fed the integrated data into the deep neural network of a gated recurrent unit-based sequence-to-sequence generation model. We found that the combination of InSAR, groundwater depth, and precipitation data had predictive power for soil composition using deep neural networks (correlation coefficient R=0.83, normalized Nash-Sutcliffe model efficiency NNSE=0.84). A random forest model was tested as baseline (R=0.65, NNSE=0.69). We also achieved significant accuracy with only 40% of the training data (NNSE=0.8), suggesting that the model can be generalized to other regions for indirect estimation of soil composition. Our results indicate that soil composition can be estimated using InSAR, groundwater depth and precipitation data. In-situ measurements of soil composition can be expensive and time consuming and may be impractical in some areas. The generalizability of the model sheds light on high spatial resolution soil composition estimation utilizing existing measurements.