Computer models predict the future? Garbage in, garbage out

Computer modeling is an increasingly popular tool used in ecological studies.  The rapidly changing climate is putting pressure on scientists to predict the trajectory of the change and the impacts those changes will have on the environment.  However, a computer model is only as predictive as the assumptions used to build it.  In other words, “garbage in, garbage out.”

That sets the stage for a study published in 2018 that predicted that “grassland may be a more reliable carbon sink than forests in California.”(1) The study was quickly adopted by native plant advocates as a weapon in their battle to destroy non-native trees in favor of grassland they prefer. (2) They prefer grassland because it was the pre-settlement coastal landscape.  They don’t acknowledge that burning by Native Americans and grazing by native ungulates were the primary reasons why grassland did not succeed to shrubs and forests prior to settlement. Pre-settlement grassland was as much a human creation as any modern landscape.

Source: US EPA, 2018

Most carbon storage is below ground, in roots and soil.  That is true of both grassland and forests. If the forest burns, the carbon it has stored in soil remains, just as the below ground carbon sink of grassland remains. 

The study (1) that claims grassland may be a more reliable carbon sink than forests reaches its erroneous conclusion by comparing below ground carbon storage in grassland with above ground carbon storage in forests. It’s a classic case of inappropriately comparing apples with oranges to the disadvantage of forests.  It seemed such an unlikely comparison that I asked the study’s authors to confirm they had compared below ground carbon storage in grassland with above ground carbon storage in forests.  They confirmed that they did, indeed, make that inappropriate comparison.

The study also bolsters its mistaken conclusions by erroneously claiming that forests are more likely to burn than grasses:

“The fire resistance for grasses is 0.5 while that of trees range from 0.1−0.3, making grasses more resistant to wildfires than trees, which is roughly consistent with field-observations since in the event of a wildfire, when compared to trees, a smaller fraction of the biomass of grass is damaged.” (1)

However, the study cited as the source of this statement (3) says exactly the opposite:

“The fraction of individuals killed depends upon the prescribed PFT fire resistance, which represents the PFT survivorship during a fire (see Table 1). In the fire model, grasses and litter are fully consumed.” (3)

Table 1 PFT parameter values for fire resistance
PFTFire Resistance (%)
Tropical broad-leaved evergreen12.0
Tropical broad-leaved raingreen50.0
Temperate needle-leaved evergreen12.0
Temperate broad-leaved evergreen50.0
Temperate broad-leaved summergreen12.0
Boreal needle-leaved evergreen12.0
Boreal summergreen12.0
C3 grass100.0
C4 grass100.0

Table 1 is consistent with this statement in the abstract of the cited study:  “Estimated litter moisture is the main driver of day‐to‐day fire probability.”  (3) Forests retain more moisture in the soil and leaf litter because of the shade provided by the tree canopy.  I wrote to the study author again, asking “where is the source of your statement that grasses are more fire resistant than trees?”  He did not reply.

If a study doesn’t seem to make sense, or it contradicts other sources of information, it is worthwhile to look under the hood.  What is driving the model?  Is it fueled by hot air?  Is it serving an activist agenda? Are cited studies accurately quoted? 

Some truth emerges from the model’s black box

Despite the erroneous assumptions of the computer model used by this study, there is some truth in the conclusions it reaches.  Vegetation type conversions are occurring now and they will continue as the climate continues to change because when the climate changes, the vegetation changes. We are presently witnessing the transition of native conifers at high altitudes to lower altitude hardwood trees. Although these changes will occur gradually and there will be many intermediary transitions, the fact is that grassland is more likely to survive than forests in a warmer, drier climate in the long run. 

The Guardian has published a comprehensive report about the loss of forests all over the world.  In the Rocky Mountains, one-third of places where trees burned 20 years ago are now occupied by shrubs and flowers.  About 15% of forests in the Rocky Mountains are not expected to grow back if killed by fire because the climate is no longer suitable for them.  About half of existing forests in Alberta, Canada are expected to vanish by 2100.  The “megadrought” in south-western US is expected to convert 30% of forests to shrubland or another type of ecosystem.

In the short run, the loss of forests can be mitigated by reforestation with tree species that are better adapted to a warmer, drier climate.  The study (1) acknowledges the potential for mitigation to preserve forest ecosystems:  Factors such as species traits, biodiversity, rapid evolution, and human management intervention could alter our model-based findings from the projections provided here. Consequently, our results indicate the potential direction of change as opposed to predictions that consider the full ensemble of ecological, physiological and management factors that can alter pathways and responses of ecosystems to climate change.”

From the standpoint of carbon storage, it is not good news that grassland is likely to inherit hot, dry lands previously occupied by forests.  Forests and wetlands store more carbon than grasslands, as the above chart in a USDA publication about carbon storage shows.  Sustaining below ground carbon sinks will depend on carbon sequestration by above-ground plants and trees.  Because above-ground carbon sequestration is primarily dependent upon the biomass, forests will always do a better job than grassland in the long run.  In the short-run, grassland will grow back more quickly than forests, but it will never achieve comparable biomass. 

Forests are presently absorbing about one-quarter of all human carbon emissions annually. Forests make a significant contribution to reducing carbon emissions, but planting trees is not a panacea as long we continue to burn fossil fuels to generate energy. The loss of carbon-sequestering capabilities of forests will exacerbate climate change in the long-run.  It’s one of many dreaded feedback loops that are reaching tipping points:  the impacts of climate change are destroying the mechanisms that mitigate climate change. 

The study (1) acknowledges that by the end of the 21st Century, under current climate conditions (warming limited to 0.3⁰ – 1.7⁰ Centigrade) forests will have removed 5 times more net carbon (carbon storage minus carbon loss) per hectare from the atmosphere than grassland in California.  See Table 1 in the study (1).  Thus, the study agrees that forests store more carbon than grassland.

From the standpoint of wildlife, it is not good news that grassland is likely to replace forests in a warmer climate. The insects, birds, and animals that live in the forest will lose their habitat. Forests are home to over 80% of terrestrial species.  We will lose our shade in a warming climate and our windbreak. 

Not an argument for destroying forests

This study (1) is unfortunately being used by the native plant movement to advocate for the preemptive destruction of healthy urban forests that are not more likely than native forests to burn in wildfires.  Virtually all wildfires in California occur in native vegetation. There is no advantage to destroying healthy forests that are expected to live for another 100-200 years.  We don’t amputate our limbs to avoid breaking them.  Nor should we destroy our forests before they die.

(1)“Grasslands may be more reliable carbon sinks than forests in California,” Pawlok Dass, Benjamin Z Houlton, Yingping Wang and David Warlind, 10 July 2018, Environmental Research Letters, Volume 13, Number 7 

(2) “Importance of Grasslands for Carbon Storage,” Yerba Buena Chapter of California Native Plant Council, Quarterly Newsletter, March 2021, page 6. 

(3) “The role of fire disturbance for global vegetation dynamics: coupling fire into a dynamic global vegetation model,” Thonicke K, Venevsky S, Sitch S and Cramer W 2001,  Glob. Ecol. Biogeogr.10 661–77

Invasion or Natural Succession?

In a recent post we considered the changes in our landscape that have occurred as a result of climate change.  In this post we examine more historical sources of change in the landscape.

Native plant advocates in the Bay Area choose to replicate the “pre-settlement” landscape that existed in the late 18th century.  The arbitrary selection of this date does not take into account that Native Americans had lived in the Bay Area for approximately 10,000 years.  Throughout that period Native Americans altered the landscape by setting fires to promote food production as well as to provide materials for cultural activities such as basket weaving.  Fires were used to improve forage for the animals they hunted and visibility during the hunt, and to funnel animals into their hunts.  Fires also promoted the growth of their food sources such as acorn production.  (1)

Unlike some parts of California, fire ignition in the San Francisco Bay Area is rarely caused by lightening, making this anthropogenic (caused by man) source of fire the predominant cause of fire historically.  (2)

After the arrival of the Spanish in the late eighteenth century, cattle and sheep grazing was the predominant economic activity in California and continued to be an important activity into the early 20th century.  These early ranchers also introduced non-native grasses which had greater nutritional value for their herds.  The non-native annual grasses out-competed the native bunch grasses, resulting in California grassland that is 99% non-native today (3).

The fires set by Native Americans and the cattle grazing of the early Californians were both instrumental in preventing the natural succession of grassland to chaparral and scrub and subsequently to woodlands.  Modern land use and management policies have suppressed fire and reduced grazing in the Bay Area.  Consequently grasslands are succeeding to chaparral and scrub.  Although managers of public lands often describe these changes in the landscape as “invasions,” Jon Keeley (Ph.D. biologist, USGS) considers them a natural succession:

“These changes are commonly referred to as shrub invasion or brush encroachment of grasslands.  Alternatively, this is perhaps best viewed as a natural recolonization of grasslands that have been maintained by millennia of human disturbance.”  (4)

Serpentine Prairie restoration. East Bay Regional Park District

So, if the succession of grassland to shrubland is natural, why do managers of public lands believe it is necessary to prevent—or even reverse– this succession?  

Serpentine Prairie. 500 trees were destroyed, including many oaks.

For example, the “Wildfire Plan” of the East Bay Regional Park District is even more ambitious than halting natural progression of the landscape.  In many instances it proposes to return the landscape to an earlier version of the native landscape.  Here are a few examples of management actions in the “Wildfire Plan” that are intended to roll back biological time to sustain native landscapes from an earlier period:

  •  “[Native] Grasslands and Herbaceous Vegetation…these widely-spaced trees will not cause an active crown fire because of the discontinuity of tree crowns.  They could, however, provide a seed source for invasion of grassland habitats by woodland species and should be considered for removal to maintain desirable and declining grassland habitat.” (page 131)
  • “[Native] Maritime Chaparral…Favor chaparral community by removing oak, bay, madrone buckeye, and other trees under 8 inches diameter at breast height that are encroaching upon the maritime chaparral.”  (page 136)
  • “[Native] North Coastal Scrub…Shift species composition towards native scrub species or consider conversion to grasslands, where appropriate on historic grassland sites…” (page 140)
  • “[Native] Coyote Brush Scrub…In most treatment areas, encourage conversion to grasslands by reseeding with native grasses…after brush removal.”  (page 149)

Serpentine Prairie being weeded by hand. Mowing will be required during the restoration. Prescribed burns will be required to maintain it as prairie.

The return of the existing landscape to earlier, historical versions requires the removal of native trees and shrubs, as well as dangerous, polluting prescribed burns.  In so doing, a permanent commitment to periodic prescribed burns is made to maintain the landscape as grassland.  And what will this accomplish? If this strategy is successful the landscape would be returned to a version of the landscape in the late 18th century, even though that landscape was actually created by the Native Americans and maintained by subsequent grazing by early European settlers.

As we often do on Million Trees, we ask the managers of our public lands to explain their strategy for artificially maintaining our landscape at an arbitrarily selected point in time.  Should we run the risks of prescribed burns for the sole purpose of replicating an 18th century landscape that was created by Native Americans?  Since California grassland is now almost entirely non-native, what is the point of preventing its succession by destroying native plants?  We don’t understand what would be accomplished by such artificial manipulation of the landscape. 

(1) “The Use of Fire by Native Americans in California,” M. Kat Anderson in Sugihara, Fire in California’s Ecosystems, 2006.  

(2) “Central Coast Bioregion,” Frank Davis & Mark Borchert in Sugihara, Fire in California’s Ecosystems, 2006.

(3) Natural History of California, Schoenherr, UC Press, 1992

(4) “Fire history of the San Francisco East Bay region and implications for landscape patterns,” Jon E. Keeley, International Journal of Fire, 2005.