An Approach to Habitat Data Summaries:

Decision Support for DAU Planning

 

N. T. Hobbs

 

Introduction

 

During the last 30 years, the Colorado Division of Wildlife (CDOW) used a process known as Data Analysis Unit (DAU) planning to set population objectives for ungulates based on two types of input.  Part of that input included data on animal abundance and model projections of population growth.  The other part included public opinion about desired population sizes. Throughout this period, data on habitat impacts of ungulates have been notably absent from DAU planning, despite widespread interest in including them.  This interest can be seen in efforts during the 1970’s to estimate carrying capacities of ungulate ranges on a nutritional basis (Carpenter 1980).  Moreover, in 1992, the enabling act for the Colorado Habitat Partnership Program (hereafter, HPP) gave local HPP committees the responsibility to provide “habitat data summaries” to inform DAU plans.  These summaries were intended to describe habitat conditions throughout a DAU to assess the influence of ungulates on habitat. Soon thereafter, the CDOW Long Range Plan committed the organization to “…set objectives for numbers of big game animals within the productive capability of their habitats” (Colorado Division of Wildlife 1994).   However, all of these efforts have failed to bring habitat information into DAU planning.

 

This failure has several causes.  Habitat data are expensive and time consuming to obtain and so these data are often neglected when agencies set budget priorities.  More importantly, it remains unclear how to relate site-specific data on habitats to populations over large areas.  In addition, research findings on ungulate nutritional requirements, foraging behavior, and the impacts of herbivores on plants are appreciated primarily by researchers and have rarely been applied in management.  However, in 1998 the HPP statewide council supported work using ecosystem modeling to inform decisions on population objectives for elk in North Park.  This effort was successful in clearly demonstrating that research findings could be combined with contemporary spatial data on vegetation and animal distribution to predict habitat consequences of different population objectives.  However, the effort failed because the approach was so detailed that its widespread application seemed unlikely.

 

What is needed then, is an approach that takes advantage of accrued research knowledge and contemporary, site-specific data at a level of detail allowing routine application in DAU planning. Here I outline such an approach.  This outline will follow in four parts.  First, I will discuss the general problem of setting objectives for ungulate abundance relative to the vegetation that supports them.  I will show that this problem depends, in a fundamental way, on estimating forage supplies and animal forage requirements at the proper level of resolution and detail.  I will then sketch an approach for estimating forage supplies for ungulates at landscape scales and for estimating the impacts of variation in weather on those supplies.  Next, I will describe how forage removal by ungulate populations of different size can be estimated.  Finally, I will relate estimates of forage supplies and animal requirements to population objectives in an adaptive management framework.

 

Setting Population Objectives Using Habitat Information

 

It is tempting to hope that given adequate data, science could provide a crisp, unambiguous answer to the question of appropriate objectives for ungulate numbers.  However, even in the best of worlds, where decisions were not limited by the availability and quality of data, there is not a single “correct” population objective.  This is because habitat conditions vary enormously over time and space, and because any decision on managing populations includes both objective science and subjective values.  The importance of human values in determining appropriate objectives makes it impossible to provide a single estimate of “carrying capacity” based on science.

 

Alternatively, it is possible for science to inform decision makers about the consequences of alternative population sizes.  That is, given a set of competing objectives for population numbers, it is feasible to provide a reasonable assessment of the relative impacts of each alternative on habitats.  Consideration of alternative objectives, alongside their habitat impacts will enhance decisions in the DAU process, but it will not make those decisions. This is the approach we will offer here.

 

The key to this approach is relatively simple.  The impacts of ungulates on habitats can be estimated as the proportion of net primary production that is consumed by populations of different sizes.   Population objectives can be set by choosing a target population size that can be sustained without causing excessive use of forage.  Specifying what constitutes “excessive” can be achieved in part by results of research and, in part, by local management goals.  Implmenting this approach requires a way to estimate total forage supplies under a variety of weather conditions and animal demand for forage assuming a range of population sizes.

 

 

Estimating Forage Supplies

 

We will estimate biomass of palatable shrubs and live and dead herbaceous vegetation using vegetation maps, digital elevation models, and a database of plant community characteristics.  These elements are described in more detail as follows:

 

 

We will use these data to calculate the standing crop of forage in each 30 x 30 map cell using:

 

                                                                                                        

 

where,

 

Bijkt = aboveground biomass (g/m2) of forage class i in map cell j at time t,

 

Xik = The mean biomass of forage class i in vegetation type k.  A lookup table would connect each specific map cell to the vegetation community database based on the plant community in the map cell.   The biomass value taken from the database can be a static mean or it can be a dynamic prediction based on the ANPP vs precipitation regression and user input on weather,

 

f(djk) = a linear function that adjusts the productivity of map cell j based on the number of cells in plant community type k that drain into map cell j,

 

gi(t) = a plant growth equation that gives the proportion of peak production for forage class i as a function of the number of weeks since the beginning of the growing season,

 

lvj = a weighting factor based on local knowledge of site potential.

 

Three components of equation 1 require additional explanation.  First, it is possible to predict the  average production based on the mean values in the plant community database.  However, this average is not likely to be terribly informative because of the high levels of temporal variability in forage supplies.  In most cases, the ability to examine extreme years (drought or severe winter) will be more revealing.  In this case, we can predict a time series of ANNP (e.g., Xik for several years) using historic weather records.

 

There is also an important temporal component within years.  Use of forage by wild ungulates during the early growing season is likely to be far more important to stakeholders than use of forage during the winter.  To account for these seasonal differences we can predict aboveground standing crops at a weekly time step during the growing season.  This is the purpose of g(t), which is simply a plant growth function.  There are several simple forms which can be used to represent g(t) (Thornley and Johnson 2000).

 

Finally, it is important to adjust estimates of production based on local knowledge of spatial variation in site potential.  The lvi term could be derived from “site potential maps” that are sketched by local land managers to reflect differences in productivity within gross plant community types.

 

 

Estimating Forage Removal by Populations of Different Sizes

 

We can estimate total forage removal from each map cell using maps of seasonal distribution of ungulates (WRIS maps), a database containing species specific diet composition, data on sex and age composition of the population, and estimates of the average body mass of each sex and age class.

 

We first estimate daily dry matter (I) intake of forage consumed from each forage class by the total population using:

 

                                                                                                       

where,

 

N = the target population size,

 

c = the number of age and sex classes estimated in the population,

 

pw = the proportion of the wth age/sex class in the population,

 

mw = the average body mass of the wth age/sex class,

 

Dit = the proportion of forage class i in the diet at time t,

 

0.025 = a standard estimate of the daily dry matter intake (kg) consumed by ruminants per kg of body mass (Cordova et al. 1978).

 

The mass of forage removed during each time step is calculated using maps of ungulate distribution and estimates of the number of days that they use the area bounded by the distribution map.  As above, these maps can be modified by local knowledge of distribution patterns.  Equipped with local knowledge, we can allocate use levels to reflect areas that are used heavily or lightly.  In the absence of local knowledge, we can assume uniform use across the vegetation map.   

 

The amount of forage removed from each cell in the map is calculated as

 

                                                                       

                                                                                                                         

                                                                       

 

where

 

Rij = the mass of forage class i removed from map cell j during the time step,

 

I = total mass (kg) of forage removed daily by the target population (calculated in eq. 2)

 

d = the duration of the time step (days),

 

laj = a weighting factor based on local knowledge of differential uses of areas with the distribution map. In the absence of local knowledge, lai = 1,

 

h = the number of cells within the distribution map.

 

 

Estimating Impacts on Habitat

 

Given the calculations above, we are able to create a landscape map portraying the intensity of use of each forage class during different seasons where the value in each map cell is:

 

                                                                                                                            

 

This map will reveal, in a general way, areas of the landscape that may be used excessively and areas that are used within acceptable limits.  The data contained in these maps can be summarized in a variety of ways.  For example, we can plot the total area of the landscape where use is considered to be excessive as a function of the target population size.  

 

 

Decision Support and Adaptive Management

 

There are a variety of ways this modeling approach can be used to support DAU planning.  Perhaps the simplest way is to contribute a scientific framework for discussion, in much the same way that population data and projections are used. Although this is useful and informative, more sophisticated approaches are available.  For example, as with all decisions in managing natural resources, one of the difficulties of setting population objectives is that the future is uncertain, primarily as a result of variation in weather.  Thus, the population objective that might be appropriate for a series of years with mild winters and wet summers will be entirely different from the objective appropriate for years with harsh winters and dry summers.  However, given a set of goals and constraints specified by managers, there is an optimum policy for managing ungulates (or any other resource) in the face of this uncertainty and there are numerical techniques for specifying that policy.  The model described above could be used to identify an optimum harvest regime given goals for the condition of the habitat.

 

Adaptive management uses models like the one above to make initial, educated guesses about the best way to manage natural resources.  In so doing, these models provide a starting point for management.  However, a more important contribution of models is to identify variables that need to be monitored to allow adjustment of management decisions as data and understanding improve.  One of the glaring shortcomings of habitat monitoring efforts in the past was the absence of an analytical framework to relate habitat data to animal populations.  The approach offered here provides that framework.

 

Carpenter, L. H. 1980. Big game investigations: Nutritional basis for quantifying the capacity of winter ranges to support deer. COLO. W-126-R-03/WK.PL.02/JOB 01, Colorado Division of Wildlife.

Colorado Division of Widlife.  1994.  Long Range Plan: approved by the Colorado Wildlife Commission March 11, 1994.  [Denver, CO : Colo. Dept. of Natural Resources.]  appendix p.4.5.

Cordova, F. H., J. O. Wallace, and R. D. Pieper. 1978. Forage intake by grazing livestock: A review. J. Range Manage 31:430-438.

Milchunas, D. G., and W. K. Lauenroth. 1993. Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecological Monographs 63:327-366.

Thornley, J. H. M., and I. R. Johnson. 2000. Plant and crop modelling: A mathematical approach to plant and crop physiology. The Blackburn Press, Caldwell New Jersey, U.S.A.