Ecological Systems of the Galuiro Mountains, Safford Ranger District, Coronado National Forest, Arizona

Jim Malusa
School of Natural Resources and the Environment
University of Arizona
malusa@email.arizona.edu

For a printable version of this report, click here.

The following is a summary of the methods used in mapping the ecological systems of Galiuros Mountains within the bounds of the Coronado National Forest. Because this map is part of a larger effort to provide maps for landscape-wide planning in the Coronado, I also provide a brief history detailing the evolution of our mapping in other districts of the Coronado, from an emphasis on soils and landform in the Catalina and Rincon Mountains, to an emphasis on existing vegetation in the Galiuro Mountains.

Background

The Coronado National Forest encompasses a dozen mountain ranges in southeastern Arizona and southwestern New Mexico. In 2008, Cleland et al. began mapping the land type associations of the Coronado and adjacent lands (Land Type Associations of Southeastern Arizona, available at http://www.azfirescape.org/home/related_activities/). This was a large-scale map covering almost 2 million acres, and subsequent efforts focused on a smaller areas and finer scale.

For instance, in 2009 we mapped the land type associations and, where possible, the smaller ecological land types, of the Catalina and Rincon Mountains. (Land Type Associations capture repeatable patterns of soil complexes and plant communities, while Ecological Land Types are based on finer scale topographic, geologic, soil, and plant associations. See http://www.azfirescape.org/catalina/mapping_methods/). The study area included a considerable expanse of alluvial fans and terraces outside the Coronado's boundaries. In all, we mapped over 100 land type associations and 300 ecological land types (http://www.azfirescape.org/catalina/landscape_types/).

Because the Coronado NF was considering landscape-scale treatments for restoring sustainable, more natural fire regimes across entire mountain ranges, these units were summarized into 30 ecological units, or 'ecounits', based primarily on potential vegetation and landform. For example: Upper Sonoran desert on fans, terraces and weakly consolidated pediments with limy soils. However, these ecounits were problematic when applied to modeling exercises such as determining the Fire Regime Condition Class, because of the difficulty in cross-walking between the ecounits we used, and the ecounits used by LandFire (http://LandFire.cr.usgs.gov/)

LandFire is a uniform system for landscape data nationally, and is created by interpretation of data from remotely sensed 30-meter pixels. There are GIS layers for both potential vegetation ('biophysical setting, or BpS) and existing vegetation (existing vegetation type, or EVT). Each of these is named for the 'ecological systems' listed and described on NatureServe (http://www.natureserve.org/explorer/index.htm/).

In 2010 our mapping efforts shifted to the Chiracahua, Dragoon, and Dos Cabezas Mountains, where we mapped not the land type associations but the existing vegetation using the accepted ecological systems names whenever possible (http://www.azfirescape.org/chiricahua-dragoons-dos-cabezas). Of the 12 ecological systems mapped, three of the ecological systems were not accepted names in the NatureServe repository: (1) Madrean Oak/Conifer/Manzanita on Rocks, (2) Madrean Pinyon-Mt.Mahogany on Limestone, and (3) Madrean Pinyon-Juniper-Oak Woodland. The first two encompass elements of the Land Type Associations, in which substrate/landform is paramount. The last, Madrean Pinyon-Juniper-Oak Woodland, is a composite of three NatureServe ecological systems: PJ woodland, Juniper Savannah, and Madrean Encinal. These were combined because of the time and money required to tease them apart without adequate imagery (such imagery was thankfully available for the Galiuros.) Mapping was done without the benefit of soils maps (most of the study area was mountainous, and lacked soil surveys), but with geologic maps to delimit limestones and associated sedimentary rocks. Available imagery, mostly 2006 NAIP, was fair but not great. The resulting ecosystem map was based primarily on field observations and photos during the summer of 2010, combined with data from other vegetation maps, including:

Southwestern ReGAP Landcover

LandFire EVT(Existing Vegetation Type)

LandFire EVC(Existing Vegetation Cover)

USFS Mid-scale Dominance Units

Ecological Systems of the Galiuros

Overview of the Study Area

Earlier mapping efforts in the Catalina, Rincon, Dos Cabeza, Chiricahua, and Dragoon Mountains included both public and private lands, taking in a good deal of the alluvial fans and valleys surrounding the mountains. For the Galiuros, we mapped only those lands within the Coronado National Forest. The Coronado includes about 135,000 acres of the Galiuro Mountains, within an area that runs about 27 miles from southeast to northwest, and 8 to 10 miles wide. Elevations range from 7663 feet atop Bassett Peak, to 3500 feet at the mouth of Kielberg Canyon. Most of the Galiuros are built of rhyolite of Tertiary age, block faulted into two roughly parallel ranges, the 'east divide' and 'west divide.' Both of these ranges hold peaks in excess of 7000 feet, with big conifers clinging to the north slopes. However, much of the high country, as seen in this view from Bassett Peak (below), is rock. The outstanding vegetation is border pinyon, typically 3 to 5 m tall, with associated silverleaf oak, netleaf oak, and manzanita.

Figure 1. Photo Point 879. The view north over the East Divide of the Galiuros, from 7660 feet (2330 m) on Bassett Peak. Virtually the entire rocky habitat within view is home to the Madrean Pinyon-Juniper Woodland, featuring border pinyon (Pinus discolor), alligator juniper (Juniperus deppeana) silverleaf and netleaf oak (Quercus hypoleucoides, Q. rugosa), and manzanita (Arctostaphylos pungens). Mountain mahogany and beargrass were common associates. October 3, 2011.

The hallmark of the Galiuros is the large galley forests of pine and cypress that trace Rattlesnake Canyon and Redfield Canyon to elevations all the way down to 4700 feet. In the same drainages, Douglas fir is an important species above 5500 feet.

The far northern study area, in the Four Mile Creek watershed, is very different (see figure 2 below). The bedrock geology is largely deeply eroded sedimentary rocks of Pliocene to middle Miocene age. The climate is characterized by more winter precipitation, at least relative to the southern end of the Galiuros. The vegetation is single-needle pinyon, not border pinyon, along with scrub oak and single-seed juniper. There are no Ponderosa or Chihuahuan pine, nor Douglas fir, despite elevations up to 6600 feet.

Figure 2. In this view looking south over the Galiuros, the red polygon represents the range of single-needle pinyon/juniper ecosystem, on the deeply eroded sedimentary rocks of Pliocene to middle Miocene age. Further south, the majority of the Galiuros are built of Tertiary volcanics.

Naming the Ecological Systems

For the Galiuros map, every named ecosystem corresponds to an ecosystem described by NatureServe. This means that a pixel in a LandFire layer can be reassigned based on the vegetation map generated in this project. Further, in the field you can use the vegetation keys generated by the NatureServe Terrestrial Ecology Dept. (e.g., Field Key to Ecological Systems and Target Alliances of LandFire Map Zones 25 and 26).

Eleven ecosystems were described, based on existing vegetation:

Sonoran Paloverde-Mixed Cacti Desert Scrub

Apacherian-Chihuahuan Semi-Desert Grassland

Apacherian-Chihuahuan Mesquite Upland Scrub

North American Warm Desert Riparian

Rocky Mountain Montane Riparian

Madrean Juniper Savannah

Great Basin Pinyon-Juniper Woodland

Mogollon Chaparral

Madrean Pinyon-Juniper Woodland

Madrean Encinal

Madrean Lower Montane Pine-Oak Forest and Woodland

For some of these ecosystems, the NatureServe description and the observed vegetation were happily congruent, as in the case of the Madrean Pinyon-Juniper Woodland. Other ecosystems were more problematic. For instance, the northern study area, characterized by single-needle pinyon, single-seed juniper, and scrub oak, was most similar to the Great Basin Pinyon-Juniper Woodland, yet the NatureServe description reads: 'This ecological system occurs on dry mountain ranges of the Great Basin region and eastern foothills of the Sierra Nevada south in scattered locations throughout southern California'. There is no mention of Arizona, but there is no other ecological system that approximates the observed vegetation.

In other cases, the ecosystem is poorly defined in the NatureServe description. For instance, the entry for Mogollon Chaparral merely mentions that the vegetation is typically evergreen broadleafed shrubs such as scrub oak, mountain mahogany, and manzanita. However, these species occur in a variety of settings, and the description makes no mention of the cover values, only that there is 'moderate to dense' shrub cover.

Mapping the Ecological Systems

No soils maps were available for the study area, which lay entirely within the bounds of the National Forest. Bedrock geology maps from the Arizona Geology Survey served to identify the rhyolite that characterizes most of the study area from the sedimentary rocks in the northern study area. Field mapping was with eight days of hiking and one of driving FS roads between May and October of 2011. A Garmin 550t GPS camera recorded several hundred photo points; data on species composition and structure (height and cover) were recorded for about 100 of these photo points. Even photo points without vegetation notes proved valuable when drawing the boundaries of the ecosystems. For instance, when trying to decide, whether a shrub on an aerial photo is a small juniper or a large manzanita; in many cases, it's plain to see from the photo, because the camera not only records the location, but the direction in which it is pointed. When viewed in ArcGIS, an arrow on the screen indicates the location and direction of the photo.

I did not revisit the Redfield drainage in 2011, but instead used photographs and notes from earlier trips in the Redfield watershed, some dating as far back as 1975. Geo-referenced photos downloaded off Google Earth, after checking via landmarks to make sure they were correctly positioned, supplemented these. Also useful was a 1985 map from an unpublished report to the Nature Conservancy, "Vegetation Associations of the Muleshoe Ranch Preserve," by L. Susan Anderson, Peter Warren, and Frank Reichenbacher.

In addition, National Agriculture Imagery Program (NAIP) imagery from 2006, 2010, and 2011 was available online, via both ArcGIS online, and Google Earth. The 2011 imagery was superb, as shown below in a view of Powers Garden, in which the cabin and tack room are clearly shown.

Figures 3 and 4.The remarkable recent imagery of the Galiuros can be appreciated in this view of Powers Garden in June of 2011, with the watercourse of Rattlesnake Creek to the right. The red arrow points to the fire pit outside the cabin. The May 2011 photo below shows the same fire pit, to the far right near the bullet-scarred chair, with a red bottle and plant press atop a plank bench that is only a foot wide, yet is clear in the image above.

This excellent imagery also benefited from a particularly dry spring in 2011, which resulted in the oaks being scantily clad in reddish leaves that make them easy to distinguish from the conifers. (Manzanita also appeared reddish, which made it difficult to discern encinal from chaparral). Below is an example of how this was exploited to help distinguish different ecosystems.

Figure 5. Two images of the same place, of about one square mile near Horse Mountain Tank in the northern study area. The upper photo is from September, 2010, while the lower is from June 2011. The lower photo shows that the dense vegetation near the bottom of the images stays green; in other words: conifers, in this case, pinyon and juniper.

Imagery from November of 2011 (below) from the northern study area very clearly showed the deciduous riparian vegetation, making it possible to map some areas, such as upper Pipeline canyon, without having ever visited.

Figure 6. The swath of fall colors reveals a gallery forest of deciduous riparian trees, likely cottonwood, sycamore, and ash, at 5000 feet in Pipeline Canyon. This was subsequently mapped as Warm Desert Riparian.

Comparing this Study to Other Mapping Efforts

This study was preceded by several others, all of which sought to describe the vegetation of the Galiuros. These earlier efforts include USFS Potential Natural Vegetation Type (PNVT, based on 24 Oct 2011 draft map), USFS Mid-scale Dominance Type, Southwest Regional GAP Ecological System, Landfire Existing Vegetation Type, and Landfire Biophysical Setting. All of these maps are remotely sensed raster data in 30 meter pixels, and all have strengths and weaknesses. In an effort to show the correspondence – or lack thereof – between this study and the previous maps, each of the ecosystem descriptions that are part of this mapping package includes a quantitative comparison to other mapping efforts.

In the example below, the Madrean Lower Montane Pine-Oak polygons created by this study were used as a 'cookie-cutter' on the Landfire Existing Vegetation Type (EVT) layer (see methods).

Figure 7. In this view of the upper two miles of Rattlesnake Canyon, between 5500 and 6500 feet, Douglas fir, ponderosa, and cypress trace the watercourses and darken the north-facing slopes. Pinyon-Juniper is on the relatively pale south-facing rocky slopes. Holdout Spring is in the upper right tributary. Image from June, 2010, NAIP.

Figure 8. Above: Same area as Figure 7, with the cross-hatched area showing the Madrean Lower Montane Pine-Oak as mapped in this study.

Figure 9. Above: The same area as in Figures 7 and 8, overlain with the LandFire Existing Vegetation Type (EVT). Red is Madrean Lower Montane Pine-Oak, blue is North American Warm Desert Riparian, dark green is Madrean Pinyon-Juniper, light green is Madrean Encinal (oaks).

Figure 9 shows poor correspondence between the Madrean Pine-Oak as mapped in this study (cross-hatched polygon), and Madrean Pine-Oak as mapped by LandFire EVT (shown in red). A GIS analysis shows that within this study’s Madrean Pine-Oak ecosystem, 29% of the EVT pixels were attributed as Madrean Encinal, 28% as Madrean Pinyon-Juniper, 22% as Madrean Pine-Oak, and 12% as North American Warm Desert Riparian Systems (which is Cottonwood-Willow, and not found in upper Rattlesnake).

Inexplicably, LandFire places much of the pine-oak on the south-facing slopes, which typically hold pinyon-juniper.

This method of comparison can be misleading if the two GIS layers being compared differ significantly in the number of total acres mapped as a particular ecosystem. For example, let's compare the Madrean Pine-Oak as mapped in this study (cross-hatched polygon) to the Madrean Pine-Oak as mapped by Southwest Regional GAP Ecological System layer (see Figure 10, below). A GIS analysis shows that within this study’s Madrean Pine-Oak ecosystem, 64% of the GAP map's Ecological Systems pixels were attributed as Madrean Pine-Oak, a relatively high correspondence. But a closer look at the data reveals that the GAP layer attributed 33,784 acres as pine-oak, far more than the 6842 acres mapped in this study. The 66% correspondence is no longer strong evidence of similar results; with so much land mapped by the GAP layer as pine-oak, it's not surprising there's considerable correspondence with the pine-oak mapped in this study.

Figure 10. The same area as in Figures 7-9, overlain with the Southwest Regional GAP Ecological System. Red is Madrean Lower Montane Pine-Oak, purple is Ponderosa Pine, yellow is Mogollon Chaparral, and green is Madrean Pinyon-Juniper.

Still, the comparison gives an idea of the differences between the various vegetation layers, and should be referred to when modeling fire behavior and Fire Regime Condition Class. In particular, it always pays to check the correspondence between alternative classification schemes when there is a vast difference in the number of acres mapped. For instance, LandFire EVT mapped 17,067 acres of chaparral, compared to the 2612 acres mapped in this study. (Chaparral, in the Landfire scheme, includes the Q. turbinella and Cercocarpus montanus alliances). It’s hard to say why LandFire chaparral is so abundant, but it does appear that, within the Galiuros, every pixel that LandFire attributes as an Existing Vegetation Height (EVT) of either '0.5 to 1.0 meter shrub' or ‘1 to 3 meter shrub’ is subsequently mapped as chaparral. (This is true regardless of the cover values; even pixels with 10-20% cover are mapped as chaparral). When comparing the imagery with the LandFire EVT of 0.5 to 3.0 m shrubs, it's clear that most of the pixels are on very rocky areas with very low cover, or simply in areas of very low cover. LandFire is certainly mapping something, but in this case it can hardly be considered chaparral.

Of course the ecosystem map offered here has its own share of shortcomings, particularly when trying to map the often-similar juniper, oak, and pinyon ecosystems. These problems are discussed in the ecosystem summaries, which provide plenty of photos to help the reader understand the range of vegetation that we've tried to squeeze into eleven neat ecosystems. Because nature often doesn't fit into our schemes, it's good to bear in mind that the author has covered only a fraction of the Galiuros (see Figure 11, below), and much of what is mapped was via binoculars and imagery. If you see a mistake, please let us know at azfirescape@cals.arizona.edu.

Figure 11. Routes traveled by the author within the Galiuro Mountains. The heavy black border encloses the Coronado National Forest. The Coronado includes about 135,000 acres of the Galiuro Mountains, within an area that runs about 27 miles from southeast to northwest, and 8 to 10 miles wide.