National Resources Inventory OASys Use Report

Author: Marty Holko

National Resources Inventory OASys Background:

The 1997 National Resources Inventory (NRI) is a compilation of natural resource information on non-Federal land in the United States.

Conducted by the U.S. Department of Agriculture's Natural Resources Conservation Service (NRCS) in cooperation with the Iowa State University Statistical Laboratory, this inventory captures data on land cover and use, soil erosion, prime farmland soils, wetlands, habitat diversity, selected conservation practices, and related resource attributes. The NRI is a statistically based survey that has been designed and implemented using scientific principles to assess conditions and trends of soil, water, and related resources. For the 1997 NRI, data were collected from about 800,000 sample sites in all 50 States, Puerto Rico, the U.S. Virgin Islands, and some Pacific Basin locations.

In general, access to the NRI information was made available to users through three primary means:

• Public access to the NRI was predominantly through NRCS national, regional or state offices. A database of NRI point sample data was made available in February 2002 (on CD-ROM) and distributed to about 180 users many of which were NRCS staff.
• NRCS national program staff and some public data requests were processed using NRI databases located at the Resources Assessment Division (Washington DC), the Resources Inventory Support Branch (Fort Worth, Texas) and the Natural Resources Inventory and Analysis Institute (Texas A&M University).
• NRCS state and regional offices accessed the NRI data through an intranet application called OASys.

Purpose of this report:

Because of the multi - use aspect of the NRI, NRCS staff have often wondered exactly what and how the NRI results were used. From the perspective of survey design and functionality there are two interesting aspects of this question which are of concern.

1. Is the information content of the survey (questions asked) relevant to current demands for natural resource information?
2. Are the results of the survey being presented or made available to potential users of the survey?

Often these questions are not only complex but relatively hard to quantify. Seldom are statistics on data access and survey use either collected or analyzed. Even if there is an attempt to quantify the survey use these attempts are often incomplete. In this vane the 1997 NRI is similar yet unique. From the perspective of  use in NRCS national offices and the general distribution of the NRI sample data to the public there is little information on the survey use. However a significant amount of the use of the survey was processed through OASys. In fact for most of NRCS's state and regional offices OASys was the only means to obtain survey results. Designed into the OASys access was a mechanism to record every request that was made through the system. Although this design was done mostly  for system monitoring and debugging it quickly became apparent that the continued collection of request information may be useful to monitor the survey access. The purpose of this report is to summarize the OASys use statistics for the 1997 NRI, to identify the information in the survey that was most often requested and to identify potential areas of limited use. Although this is not a complete analysis of the use of the 1997 NRI survey it does cover a significant portion of the use by NRCS state and regional offices.

OASys access design and functionality:

In general the OASys system can be thought of as a system that allowed the user to design a statistical table. The cells of this table were completed through a request to the NRI database. There are five components that the user must specify to make a valid table request. These are:

• The state(s) that the data represents.
• The unit that the user would like to estimate.
• The row elements that the data are categorized into.
• The column elements that the data are categorized into.
• The year(s) that the data request is for.

States:

For each request the user must select at least one of the states for which the NRI data is available.

Unit:

For each request the user must have selected one of the eleven units available for the cells in the statistical table. For the purpose of this analysis these units were grouped as follows.

Row or Column Elements:

Appendix I shows the complete list of elements that were available through the OASys system. The specific elements were grouped into categories in  OASys to aid the user in navigating to the specific element desired. An element could be selected for the roe, column or both the row and column of the desired table. This was done when the user desired to get a table indicating change from one year period to another. In this case the user would select an element such as "Broad Cover Use" for both the row and column. However for the row they would select say year 1982 and for the column they would select year 1997. In this fashion the user could get the acres of 1982 forest that converted to say cropland for 1997.

Years:

The NRI survey allowed the user to compare the condition of the natural resource characteristics for the years 1982, 1987, 1992, and 1997. As mentioned previously prudent use of the OASys system allowed the user to track the changes in the natural resource condition over two time frames by submitting only one request to through the OASys system.

Analysis:

OASys recorded every NRI information request placed to the system from late November 2000 through October 2002. The total number of requests processed was 9,463. Figure 1 shows the number of request by month as a percent of the total.

Figure 1

Appendix II shows the percent of requests by state. The state indicated in these statistics is not necessarily the state that requested the data. The state is the one for which the data was requested. The request could have been made by anyone with access to OASys. The multi state category are those requests that were for information pertaining to more than one state. Figure 2 shows the percent of the requests by NRCS regional office. 

Figure 2


Because of the many possible analysis combinations available through the OASys system it is easy to become confused when trying to categorize these requests. The approach used in this analysis is to examine the NRI requests from a more general categorization to a more specific one. Using this approach it is possible to identify categories of the NRI information that were used infrequently. As we eliminate segments of the NRI data we can then examine the specific information that was requested often enough to warrant a more detailed analysis. In light of this approach the first aspect of this analysis was to look at the estimation units that were requested.

Table 2 shows the number of requests (and percent of total) by estimation unit grouping. The table shows that requests for Area estimates make up the vast majority of all requests made to the OASys system. Request for Universal Soil Loss (USLE) statistics are a distant second. These numbers however may be slightly misleading. Information for the categories other than Area can also be selected as row or column elements. In particular soil erosion characteristics were requested either in the row element or column  element for about 1000 of the 8403 Area requests. Including these reports would indicate that soil erosion information was requested in about 22 percent of all of the OASys requests instead of the 11 percent indicated in Table 2.

Figure 3

The confusion associated with the evaluation of the use of soil erosion points out one of the more difficult aspect of trying to analyze NRI use through OASys. Like erosion information it is possible to use many of the NRI elements in more than one way. In particular any of the elements can be used in either the row, column or both row and column of the requested table. In light of this it seems that the clearest way to view element use is to look at the number of requests which contained each category (no matter how it is used) as a percent of total NRI requests. Table 5 indicates the percent of the total 9,463 requests by each general category. Because of the duplicated use the percentages do not add to 100 percent.

Figure 4

From Table 3 we can see that "Cover Use" and "Geographic Area" were the dominant categories requested from OASys. To get a more detailed review Appendix III presents figures that show the detailed element use within each general category. In nearly every case requests specifying a "Geographic Area" in the row or column specified some other NRI element as the other component of the request. Figure 5 shows the percent of requests ignoring "Geographic Area". The elements listed in Figure 5 include only those that accounted for more than one percent of the requests. From Figure 5 we can see that "Broad Cover/Use" dominated the requests presented to OASys with almost 68% of the requests. With "Specific Cover Use" accounting for 6 % of the requests cover use statistics was requested in over 75 % of the OASys use. Figure 6 is included to give a clearer picture of the elements requested excluding the cover use elements. Of the roughly 65 elements that could have been requested in OASys one accounted for 68% of the requests, one accounted for 6% and eleven others accounted for 1 to 5% the remaining 50 or so elements were relatively unused.

Figure 5
(Columns are listed from left to right)

Figure 6
(Columns are listed from left to right)

In Appendix III (Figure A) geographic requests by "County" or "Water Resources Subbasin (8-digit Hydrologic Unit)" amounted to almost 50% of the data requests made to OASys. Considering that the precision of the estimates obtained for individual counties or 8-digit hydrologic units were at best marginal, it is important that we at least look at the number of these requests. Table 4 shows that 28 percent of the requests made were of statistics for an individual county and 8 percent were of statistics for an individual 8-digit hydrologic unit.

Discussion:

As one might expect this quick review of OASys use may present us with more questions than answers. First it is apparent that OASys was used during this 2 year period. OASys averaged over 400 requests per week and still continues to receive 50 to 100 requests a month even though the data is 5 years old. The perplexing part of this summary is why a significant portion of the NRI (over 50 elements) was used so little. As mentioned earlier one could expect the limited use to be a combination of either lack of interest in the information or the lack of the knowledge as to how to use the information. In either case this situation should be of concern for the design of future NRI surveys.

Is it that the information was not requested because of lack of interest? If so then one might conclude that a review of  the content of the NRI is necessary. One must be careful in concluding this. It must be remembered that this analysis does not cover all of the use of the NRI. This analysis does point out that some attempt should probably be made to assess the use and value of the elements surveyed. Changes in the content of the survey could have significant impacts on the NRI process.

Is it that the presentation of the NRI information is insufficient in some means? If this is the case then how can NRCS present the data or NRI results such that users can gain the full value of the inventory? Would it be more appropriate to present an analysis of the NRI? It may be more appropriate to generate summary tables that depict key natural resource information in stead of providing resources to build systems like OASys. These systems rely on the knowledge and interest of the user to first "think of" and then generate relevant natural resource statistics. Given the relatively specific information requested it is possible to generate the statistical tables directly and not have to maintain a system like OASys. 

One final concern highlighted in this analysis is the significant use of the NRI to obtain small area statistics. Over one third of the requests to OASys were to obtain statistics for small sub state areas such as an individual county or 8 digit hydrologic unit. NRI documentation has repeatedly pointed out that these estimates are often unreliable. The fact that these estimates continue to be requested would indicate that there is considerable demand for these statistics. The requests for these statistics point out that the NRI process needs to continue to caution users. In addition if the demand is great enough NRCS  may need to look at a means to meet this need for natural resource information.

Appendix I - Complete Element List by Generalized Category

Appendix II - Number and Percent Requests by State for Which Data is Requested

Appendix III - Percent of Requests by Specific Element

Figure A
(Columns are listed from left to right)

Figure B
(Columns are listed from left to right)

Figure C
(Columns are listed from left to right)

Figure D
(Columns are listed from left to right)

Figure E
(Columns are listed from left to right)

 Back to Top