The Hidden Networks Shaping Our Changing Landscapes
Walk through the arid landscapes of the American Southwest or the sprawling savannas of Africa, and you might notice something subtle yet significant changing. Grasses that once dominated these regions are gradually giving way to woody shrubs and trees in a phenomenon scientists call woody plant encroachment (WPE). This quiet transformation represents one of the most dramatic landscape changes of the past century, affecting nearly 40% of Earth's surface—regions we classify as drylands that support over 38% of the global population 1 .
At first glance, more trees might seem beneficial. But in water-limited ecosystems, this shift sets off a complex chain reaction that reshapes how water moves through the landscape, potentially altering groundwater recharge, streamflows, and the very resilience of these ecosystems.
For decades, scientists struggled to explain why seemingly similar areas responded differently to WPE—in some cases becoming more efficient at capturing water, in others spiraling toward desertification. The key to unlocking this mystery lies in understanding ecohydrological connectivity—the hidden networks that govern how water, nutrients, and energy move through and transform these critical landscapes 1 .
Woody plant encroachment describes the gradual conversion of grasslands and savannas to woodlands through the expansion of trees and shrubs. This transformation has been accelerated by various factors including overgrazing, disrupted fire regimes, climate change, and even rising atmospheric CO₂ levels that favor woody species over grasses 1 .
The consequences extend beyond botanical shifts. As woody plants claim more territory, they can reduce forage production that sustains livestock-based economies, diminish biodiversity adapted to open ecosystems, and fundamentally alter the water cycle—a concern for the billions who depend on dryland resources 1 .
The effects of WPE aren't uniform across drylands. Research across a climatic gradient from subhumid to arid regions reveals strikingly different outcomes:
| Climate Zone | Primary Process | Ecosystem Outcome |
|---|---|---|
| Subhumid | Increased evapotranspiration | Reduced streamflow and groundwater recharge |
| Semiarid | Increased soil infiltrability | Variable effects on streamflow depending on soils |
| Arid | Xerification with bare, interconnected patches | Increased surface runoff and groundwater recharge 1 |
In wetter drylands (subhumid regions), the additional leafy surface area from woody plants leads to greater water loss through evapotranspiration, reducing water availability in streams and aquifers. Meanwhile, in drier regions, the story grows more complex—woody plants can either enhance or diminish water availability depending on how they rearrange the landscape's connectivity 1 .
Ecohydrological connectivity provides a powerful lens for understanding these divergent outcomes. Simply put, it describes how readily water, sediments, nutrients, and organisms can move across a landscape. This connectivity operates in multiple directions—longitudinally (along stream channels), laterally (across hillslopes), and vertically (between surface and subsurface) 1 .
Connectivity along stream channels and river networks
Connectivity across hillslopes and floodplains
Connectivity between surface and subsurface
When woody plants invade grasslands, they fundamentally rearrange this connectivity by creating a mosaic of resource islands (under their canopies) surrounded by bare soil patches. The canopy areas typically feature higher infiltrability thanks to root channels and organic matter, while the intercanopy spaces often become more compacted and connected. This rearrangement controls how rainfall is partitioned between runoff and infiltration, and ultimately determines the water future of the entire ecosystem 1 .
The ecohydrological impacts of WPE emerge from the tension between two countervailing feedback loops:
As vegetation cover decreases and bare patches become more connected, the landscape loses more water and nutrients through runoff, further reducing plant productivity in a self-reinforcing cycle that can lead to degradation .
The same bare soil connectivity that causes global resource loss also directs more runoff to remaining vegetation patches, creating local resource concentration that enhances plant growth and helps preserve vegetation cover .
The balance between these opposing forces—one destructive, the other protective—determines whether a dryland ecosystem trends toward degradation or maintains its functioning despite transformation.
To test how these connectivity-mediated feedbacks influence dryland vulnerability, researchers developed a spatially explicit model that simulated vegetation dynamics under different feedback strengths . This computational experiment offered crucial insights into what drives resilience and collapse in dryland ecosystems.
The research team built upon existing dryland models by creating a probabilistic cellular automaton—essentially a grid where each cell could exist in one of three states: vegetated, bare soil, or degraded .
| Parameter | Description | Role in the Model |
|---|---|---|
| Cell States | Vegetated, Bare Soil, Degraded | Represented different landscape conditions |
| Global Feedback Strength | Controlled runoff and resource loss from the ecosystem | Simulated the positive feedback loop |
| Local Feedback Strength | Controlled resource redistribution to vegetation patches | Simulated the negative feedback loop |
| Environmental Pressure | External factor like drought or grazing | Tested ecosystem resilience |
The model uniquely incorporated both the global positive feedback (where increased bare soil connectivity enhances resource loss from the system) and the local negative feedback (where the same connectivity channels resources to vegetation patches). By adjusting the relative strength of these feedbacks, researchers could observe how different dryland systems responded to increasing environmental pressure .
The modeling revealed several critical patterns:
When the global positive feedback dominated, ecosystems were more vulnerable to catastrophic shifts—sudden transitions from vegetated to degraded states with small increases in environmental pressure. Conversely, when the local negative feedback was stronger, vegetation cover decreased more gradually, and systems showed greater resilience to pressure .
Systems with stronger local negative feedbacks required less improvement in environmental conditions to recover from degraded states. This has profound implications for restoration—enhancing conditions that strengthen the local negative feedback could make ecosystem recovery more achievable .
The models showed that stronger local negative feedbacks promoted the formation of banded vegetation patterns—a phenomenon observed in drylands worldwide that represents an efficient adaptation to capture and retain scarce water resources .
| Feedback Scenario | Regime Shift Risk | Recovery Potential | Characteristic Pattern |
|---|---|---|---|
| Strong Global Positive Feedback | High catastrophic shifts | Difficult, requires significant pressure reduction | Patchy, disconnected vegetation |
| Strong Local Negative Feedback | Gradual changes, low catastrophe risk | Easier, requires moderate improvement | Banded or spotted vegetation patterns |
| Balanced Feedbacks | Moderate shift risk | Moderate recovery potential | Mixed patterns |
These findings fundamentally reshape our understanding of dryland dynamics. Rather than viewing vegetation patterns merely as symptoms of ecosystem condition, we can now see them as visual representations of the underlying feedback strengths that determine ecosystem resilience.
Understanding ecohydrologic connections requires specialized approaches and tools. Researchers employ a diverse toolkit to measure, monitor, and model the complex interactions between vegetation, water, and landscape structure in drylands.
Computational frameworks like the probabilistic cellular automaton used in the featured experiment allow scientists to simulate how vegetation patterns and hydrological processes interact across large areas and over time, testing scenarios impossible to conduct in the field .
Using instruments like rainfall simulators and infiltrometers, researchers quantify how different vegetation patches (canopy vs. intercanopy) partition rainfall into infiltration and runoff—the fundamental process driving resource redistribution 1 .
Through remote sensing and geographic information systems (GIS), scientists map the connectivity of bare soil areas, tracking how patches link to form networks that channel water across the landscape 1 .
Using stable isotopes or dyes, researchers follow the actual pathways of water movement through dryland systems, revealing where water originates and where it ultimately supports plant growth 1 .
Established research sites like the Jornada Basin in New Mexico or the Kruger National Park in South Africa provide decades of vegetation, soil, and hydrological data that reveal slow processes and sudden shifts in dryland ecosystems 1 .
This multidisciplinary toolkit—spanning computation, field measurement, and long-term observation—enables researchers to decode the complex language of dryland connectivity and predict how these vital ecosystems may respond to future changes.
The emerging understanding of ecohydrological connectivity offers powerful new perspectives for guiding and managing landscape change in drylands:
Rather than viewing woody plant encroachment solely as a problem to be solved, the connectivity framework helps us see it as a dynamic process whose outcomes depend on the balance of underlying feedbacks. This shift in perspective opens new management opportunities—instead of focusing solely on removing woody plants, we might manipulate connectivity to strengthen the negative feedbacks that enhance ecosystem resilience .
The research suggests that restoration efforts could benefit from strategically enhancing local negative feedbacks. This might involve:
These principles extend to more human-dominated landscapes through initiatives like Colorado's Transformative Landscape Change programs, which replace water-intensive turf with native, water-wise vegetation. Such efforts essentially reengineer urban connectivity networks to reduce water loss while maintaining aesthetic and ecological benefits 2 5 .
Understanding how connectivity operates across climate gradients helps planners anticipate how drylands might respond to climate change. As precipitation patterns shift, the balance between positive and negative ecohydrological feedbacks may tip, potentially pushing some ecosystems across thresholds to alternative states 1 .
The true revolution in dryland science lies in recognizing that what connects a landscape—the hidden networks of water, sediment, and nutrient flow—ultimately determines its fate. The simple presence or absence of woody plants matters less than how their arrangement transforms these flow networks.
As we face a future of increasing climatic variability and growing human pressure on drylands, the connectivity concept offers both warning and hope. It warns that small, incremental changes can suddenly reorganize entire ecosystems when connectivity thresholds are crossed. Yet it also offers hope—by learning to read the subtle language of landscape patterns and strategically managing connectivity, we might steer transformative change toward resilient, functioning ecosystems that continue to sustain the billions who call drylands home.
The challenge ahead lies in applying these insights across the vast dryland regions of our planet, transforming not just landscapes but our relationship with these vital and vulnerable ecosystems.