5.2 Site Descriptors and Covariates
Video Presentation
Learning Guide
In this module we discuss things that you measure and record at a site that are not, in themselves, direct indicators for your study or monitoring. We can collectively call these site descriptors and covariates.
Why Gather Site Descriptive Information?
Plot characteristics influence land condition and changes – and thus provide very helpful context for interpreting monitoring data
The example here shows two different sites only a couple miles apart from each other before and after spraying
What do you think is different about these two sites that might be causing some of the differences?
(Answer: soil in this case, shallow clay soil on the top vs. a loamy soil; also slightly higher elevation for the bottom site)
- Plot characteristics include things that don’t change over time such as Soils and Landscape Position!
- These features help us understand landscape conditions and changes because they affect the potential of the land. For example, soils and where you are on the landscape (i.e., landscape unit) affect the ability of the land to produce particular kinds and amounts of vegetation.
- Soils also affect land response to management and disturbance (e.g., spraying herbicide, wildfire, grazing)
- By using soils information and landscape units, you can determine ecological sites
- Ecological sites also help us figure out what the potential of the land is – in many cases, AIM uses ecological sites as strata for their sample design to ensure points are distributed evenly across ecosystem types.
Land Potential
Both of the examples above describe the effect of land potential on the plant communities and their response to management or disturbance.
Land potential is the capacity of land to support ecosystem services required to meet “the needs of the present without compromising the ability of future generations to meet their own needs” (Bruntland Commission 1987, Herrick et al. 2013). Land potential is defined by a site’s climate, soils, landscape position or landform, and disturbance history. Knowing a site’s land potential helps you interpret monitoring indicators (e.g., is there too much, too little, or the expected amount of bare ground here?)
Many of the site descriptive information we collect are related to land potential.
Site Descriptors – General descriptive information
Site descriptors are pretty much just what the term implies. They are descriptive information you record about the site you are measuring. This could include unstructured information on where the site is located, ownership and access information, indications of recent disturbance or management, and even photographs of the plot. The purpose of this type of unstructured site information is to give you reference information if you need to revisit a site or understand why you got certain values for indicators at the site. In other words, this kind of general site descriptive information can be very valuable to refer back to when you are analyzing your data.
Structured Site Descriptors
Another class of site descriptive information is structured information you collect on the physical properties of the site. This could include measures such as elevation, slope, landform, precipitation, and temperature. For these properties, specific measured values are recorded or are made according to established categories. These properties can either be measured at the site (slope and landform are easily measured in the field), from nearby locations (which may be appropriate for temperature and precipitation if there is a weather station close by), or from GIS layers (such as for elevation or climate information).
Site Descriptors – Guided Observations
Structured site descriptors can also take the form of guided observations. For example, is there evidence of erosion, grazing, fire, or disturbance? If so, a rubric can help categorize these observations into consistent classes in a repeatable way to create data that could be for analysis. Consider the signs of soil erosion that are recorded as part of the plot observation data sheet in the Monitoring Manual for Grasslands, Shrublands, and Savanah Ecosystems (Herrick et al. 2017). In this instance, observers use five classes to evaluate whether the site exhibits any signs of soil erosion for six different aspects. These kinds of observations are generally reliable and quick to implement, and can provide useful information for interpreting monitoring data.
Site Descriptors – Soils information
The soils at a site are one of the primary determinants of land potential and the type and amount of vegetation that can occur at a site. Because monitoring data need to be analyzed and interpreted relative to the land’s potential, having good soils data collected from the site is important for studies and monitoring programs. Collection of soil data does not need to be complex, but it does mean that 1) you need to have some basic training and skill in collecting soils data, and 2) yes, you need to dig a small soil pit. Typically, data are collected by soil horizon or depth for soil texture class, percent rock fragment, percent clay, effervescence, color, and structure. While this may seem like a lot of work, the value of the soils data you will get from it will be worth it!
Soil Coarse Fragment Content
Soil coarse fragment content is the percent rocks (defined as particles greater than 2 millimeters) by volume within the soil. Coarse fragment content affects soil water availability through its effects on infiltration rates, storage, and evaporation. An example of this is that in arid regions, trees often grow better than grasses in soils with lots of coarse fragments in the top meter. Why do you think that might be? A hint is to compare tree roots to grass roots. Water drains quickly through soils with a lot of rock content and so is often stored deeper in the soil profile. Tree roots are longer and can reach pockets of soil where water collects. In contrast, grass roots are short and concentrated near the soil surface, so they cannot access deeper soil water. You can measure soil coarse fragment content by gathering some soil from the horizon in a measuring cup, recording the total volume with rocks and smaller particles, sieving the soil through a 2 millimeter sieve to remove the rocks, and then recording the volume of soil left without the rocks. The percent coarse fragment content is the total volume minus the volume without rocks, divided by total volume.
- Equipment needed: measuring cup & sieve
Soil Texture
Soil texture refers to the percent soil particles of different sizes. The three major size classes are sand, silt, and clay. Sand is the largest at 0.05 to 2 millimeters; silt is intermediate size at 0.002 to 0.05 millimeters; and clay is the smallest at less than 0.002 millimeters. The amounts of sand, silt, and clay can be used to assign texture classes to soils according to the soil texture triangle pictured here. Texture affects all soil processes and is one of the most important soil properties. For example, we have already discussed how water dynamics can differ in a loamy or coarser-textured soil versus a clayey or finer-textured soil. In the Plot Characterization protocol from the Monitoring Manual for Grassland, Shrubland, and Savannah Ecosystems, texture is estimated for each horizon using the texture by feel method. A spray bottle is used to wet a soil sample and the sample is manipulated by hand in various ways to determine the broad texture classes pictured in the texture triangle. Detailed information can be found in the protocol. While this is a quick and reliable method with experience, it is a good idea to practice with soils of known texture each season before going out in the field. You can also perform quality control by determining texture in the lab using the hydrometer method for a subset of samples.
- Equipment needed: spray bottle, your hand, & soil triangle/chart
Image source: NRCS
Percent Clay
Percent clay is a particularly important aspect of soil texture. Clay, again, is the smallest soil particle size at less than 0.002 millimeters (or less than 2 micrometers). The amount of clay in the soil is especially important for understanding the processes of water infiltration and availability, soil stability and erodibility, and plant rooting. Look at the picture from eastern Colorado during the Dust Bowl, a major wind erosion event. How do you think soil texture and clay content influence wind erosion potential? Coarser textured soils (more sand, less clay) are significantly more erodible by wind. Similar to soil texture, the method for estimating percent clay is by feel, using the same flowchart.
- Equipment needed: spray bottle, your hand, & soil triangle
Effervescence
The next soil property is effervescence, which is an indicator of calcium carbonate content. Calcium carbonate often accumulates in soils in arid areas. Large amounts of calcium carbonate appear as white specks, nodules, or even a cemented horizon in the soil like the one in the picture. How do you think this cemented calcium carbonate horizon might affect plant growth? For one, it physically stops many plant roots from growing past it. However, it can be beneficial also. Water often collects along the surface of a cemented calcium carbonate horizon, and is available to plant roots. Calcium carbonate amount is estimated by dropping a weak hydrochloric acid solution on a piece of soil and rating how “effervescent” the response is. A violent foam of bubbles means that there is a high concentration of calcium carbonate. No effervescence means that there is no calcium carbonate.
- Equipment needed: weak solution of HCl
Soil Structure
The next property is soil structure. Structure is a description of the size, shape, and strength of soil peds or pieces. You would describe it in the field by matching the soil structure to one of several classes: for example, granular in the left hand picture versus platy on the right. Soil structure affects soil water availability through the depth that water can penetrate and residence time that water is available to plants. In the example, which surface structure do you think is superior for seed germination—granular or platy? In general, granular structure is better for seed germination. It provides better seed to soil contact and does not present a barrier to buried seeds trying to push up through it. A recent paper in Landscape Ecology found this is true for the success of sagebrush seedings as well – soil aggregates (which usually have granular surface structure) was a significant predictor of sagebrush seeding success in the Soda Fire in Idaho Describing structure is optional in the protocol described by the Monitoring Manual for Grassland, Shrubland and Savannah Ecosystems. However, it is easy and quick to measure once trained.
- Equipment needed: Field Book for Describing and Sample Soils
Covariates
A covariate is a variable that is used to help explain the value of or change in an indicator. Covariates are not indicators themselves, but things that we record about a site that are related to, or in some cases control, the value of an indicator. Many of the site descriptors, especially structured ones like elevation, precipitation, or soils, can be used as covariates. We can also use other site measurements as covariates. For example, cover of perennial grasses at a site might be influenced by the cover and height of encroaching shrubs like juniper or mesquite. Here, perennial grass cover is the indicator, and encroaching shrub cover (also measured at the site) is a covariate. A helpful rule of thumb for determining covariates is, “If it can be measured, but it’s not one of your indicators, then it is potentially a covariate.” Many of the properties we measure or record as site descriptors are often used as covariates in analyzing monitoring data. You can find more information on covariates at https://www.theanalysisfactor.com/confusing-statistical-terms-5-covariate/Links to an external site..
Using Covariates: an Example
Consider this example for how we might use covariates to help interpret or explain patterns we see in our monitoring data. Monitoring data were collected on 32 sample sites in a sagebrush ecosystem. One of the primary indicators for this monitoring was cover of perennial grasses because of the strong relationship between perennial grasses and rangeland health. From this graph, we can see that the majority of the sites have more than 20% cover of perennial grasses, which looks pretty good, right? However, there are a handful of sites with lower perennial grass cover. It would be tempting to conclude that these sites are in bad shape because they lack perennial grasses, but let’s take a closer look first.
One of the covariates measured as part of this effort was precipitation from the PRISM climate layers. The precipitation information might give us a window into what is going on with perennial grasses in this area.
This graph is a scatterplot of perennial grass cover by site precipitation with each dot representing one of the monitoring plots. Looking at this graph, we can see there is a fairly strong relationship between precipitation and perennial grass cover. So, what does this mean? Well, one explanation is that the amount of precipitation a site receives helps determine the potential of that site to produce perennial grass. Thus, just because a site has little perennial grass does not mean we can just conclude that it is in poor shape.
Let’s take this analysis another step further, and look at the influence of soils on perennial grass cover.
This graph is the same scatter plot as the previous slide, but the points have been colored according to each plot’s ecological site. Ecological sites will be discussed in more detail in the next module, but briefly, ecological sites are determined primarily by soil composition and landform and are an expression of the site’s potential to produce a type or amount of vegetation. In other words, ecological sites tell us about land potential. If we look at the sites with low perennial grass cover, we can see that almost all of them are Black Sagebrush sites, which are inherently lower productivity sites than the other sagebrush ecological sites in the area.
The take-home message here is that we need to interpret our monitoring data within the context of land potential and other management activities or ecological processes that may be occurring at the site. Covariates help us control for these other factors in our data analyses.
Module Review
- Site descriptors are information recorded about a site that can aid in analysis or interpretation of monitoring data. Includes:
- Unstructured data like general descriptions, location information, and notes
- Structured information like elevation, slope, precipitation, and soils data.
- A Covariate is a variable that is used to help explain a value or change in an indicator.