HowTo – Framework for long term sites
Priority: high
Updating: rare
This documents the project’s strategy for recruiting voluntary participants in long term site roles, contract task 2.1. This supersedes certain detailed terms in the contract scope of work. This substitution became effective 8/30/2021 upon agreement among DEC and Cornell BEE SWL, then was refined on 6/20/2023 after reviewing the recruitment results so far.
Change log:
| When | Who | Comment |
|---|---|---|
| 2021 08 31 | Sp17 | First version launched to reflect revised strategy with DEC agreed on 8/30. |
| 2021 09 15 | Sp17 | Brian has done some screening of earlier sampled sites in five of our six counties. We exclude Schenectady. |
| 2021 10 01 | Sp17 | Reformatted for export in DOCX |
| 2023 01 04 | Sp17 | Updated for how we are pursuing this in late 2022. Actually not many changes. Bumped from 50% complete to 75% complete. |
| 2023 04 20 or later | Sp17 | Updated for major progress in 2023Q1. We had to relax many criteria. [incomplete] |
| 2023 06 05 | Sp17 | Updated for DEC review. Content remains weak in how we are recruiting. Added a snapshot of the status tracker bar graph. |
| 2023 06 21 | Sp17 | Yesterday NYSDEC decided that we should spread long term sites out farther than the first set of ~20. This edit incorporated that as a goal but we do not yet have a strategy. |
| 2023 09 28 | Sp17 | Updated recruitment strategy to final as materialized. To 100%. |
Cross references to other HowTos:
1. Objectives
- “Long term” sites should have wells that tap recharge from a variety of upgradient pesticide users, not necessarily including the site owner themselves.
- The sites should represent a mix of most vulnerable to less vulnerable aquifer areas (biased toward more vulnerable), and focus on average to higher area pesticide uses.
(In practice the above two criteria became infeasible given our staffing and the difficulty of direct recruiting of mostly homeowners.)
- The sites should be geographically diversified thus an ecoregional and county diversity are relevant.
Since the data yield from this type of site will be much smaller than from categorical sites, because of having a single sample per year and a single sampling point per site, the long term sites as an ensemble are a lower priority for recruitment effort and site quality considerations than the categorical sites as an ensemble.
2. Desirable attributes of sites for the long term roster
The original contract wording is:
Assess long-term trends in pesticide detections. Monitor groundwater once annually at 24 wells spread across the six intensity/vulnerability areas of 1.2. The wells will be selected from the NYSDEC Water Well Program database to the extent that such wells represent probable upgradient pesticide uses and the region’s pattern of ground water extraction (such as deep rock wells or shallow wells in principal aquifers). While it is desirable to have four wells in each of the six areas, if four appropriate monitoring wells cannot be located within each of the six areas, then Cornell will identify and substitute wells in other geographic areas of similar intensity/vulnerability. Wells already included in Cornell’s past monitoring agreements would be evaluated as candidates, especially given an earlier history of sampling and analysis. These wells are intended to be sampled once per year for several years. Final well selection is subject to Department approval.
We substituted ecoregional diversity, combined with principal aquifers, for “six areas” in an August 2021 refinement of the framework for categorical sites. A similar diversity approach is appropriate for populating the long term roster.
We prefer that a recruited site have been involved in earlier sampling projects, thus bringing some prior data and an early comfort level with Cornell visiting their property annually to sample.
The early long-term sites recruited were focused in Cayuga, Cortland, Orange, St. Lawrence, and Jefferson Counties. This proved to be insufficient diversity in presumed aquifer vulnerability and geographic spread.
Following a June 2023 review, we updated the diversity criteria to emphasize counties over ecological regions, partly because counties are better understood to the public than ecological regions. Later recruits are much more spread out, adding sites that bring the span of this type of site to 18 counties as of September 2023.
We are treating the original contract target of 24 sites as nonbinding, having recruited over 30 to increase the spatial spread.
We are not necessarily fixed to keeping a site in the network for the full time period. These sites can be changed year to year more easily than the categorical because of not needing well installations and our being less intrusive to a business operation by not trying to understand the owner’s own pesticide use.
3. Relationship between categorical sites and long term sites
The long term sites would be sampled once per year, one well per site, thus would yield many fewer samples per year than the categorical sites with multiple wells each and two visits per year. The priority of long term sites is clearly much lower.
By striving for wide ecoregional diversity of the categorical sites we deliberately aim for a spatially sparse network. This sacrificed sample collection efficiency by having more road miles per site than if we were to attempt to cluster the categorical sites closer together or closer to Ithaca.
A project revision in June 2023 made us aim for county diversity of long term sites. This had the effect of spreading out the long term sites even farther afield than the categorical sites. Together the 58 categorical, long term, and lake sites reach about half of upstate New York’s counties.
4. Drawing from earlier sampled sites in six counties
The earlier six counties were biased toward higher area pesticide use intensity, particularly Genesee, Orange and Wayne Counties. That is consistent with the “categorical” and “long term” criteria.
The individual sites in these counties were not recruited with aquifer sensitivity in mind. Their wells are generally weakly characterized (no stratigraphic logs), and water tables were not measured. In this project, most of the original 40 per county were screened out before we began contacting potential re-recruits.
We originally characterized land use near the well sites that we thought may influence the well. There is one major advantage to re-recuriting an earlier owner: our earlier sampling results.
In one unusual case, a prior County samplee returned as a categorical site cooperator.
Mailing to a selection of the owners involved prior to 2021 yielded a disappointing response rate; attempts to reach the non-respondents via telephone encountered disconnected numbers, two outright refusals, and a couple of additional yeses. Evidently it had been too long since the prior contact.
All of the earlier sampled sites are in six layers within the QGIS map set, for convenient overlay of principal aquifers, soils, Cropscape agricultural land use, and committed categorical sites.
Patience ultimately yielded rejoined sites in Wayne, Genesee, Cayuga, Cortland, and Orange Counties, five of the original six.
5. Additional Long Term sites outside the six counties
The disappointing results from mail and telephone contact requests led to an eclectic strategy for recruitment. We did an experiment of peer-to-peer and drop-in connecting that yielded several cooperators with samples in one week. These sites are clustered in Jefferson County, with one in St. Lawrence.
We reached into two emerging (or re-emerging) farm types for fresh long term cooperators: cannabis and hops. Cannabis growers, using both fields and greenhouses, proved to be very willing to be involved. Their peer-to-peer effects yielded some. Hops farmers were not as easily reached. We accepted one organic cannabis grower whose upgradient neighbors apply pesticides.
Reconnaissance trips, drop-ins and phone contacting, and the cannabis peer networking yielded cooperators in 13 counties beyond the five earlier-involved counties, a total of 18.
6. Site characterization
Because of the lower priority of long term sites relative to categorical sites, as reflected in their samples/year quota, we devote limited effort to characterize such a site. There are no separate GIS maps per site; instead we use the three substate swath maps (east, central, west) and pan/zoom when needed such as to identify soil, surficial geology, and ecological region. The substate maps include county-level SSURGO soil maps for all counties covered by the swath.
An important aspect is the well characteristics. We attempt to discover the total depth, construction (drilled, dug, driven), and what material it taps (bedrock, unconsolidated). A few owners do not know the depth of their well or much about its construction. Until 2000 upstate well installations did not require submission of a well completion report with stratigraphy to NYSDEC. Indeed, only two of the categorical and none of the long term sites have wells in the NYSDEC registry. One of the two has a well that is 500 feet into bedrock, thus unusable in sampling for their own pesticide residues. The other registered well is one we actually sample.
One trusting and sophisticated long term owner sent us their own (irrigation) well completion report including the stratigraphic log.
7. Site identifiers
The tracking spreadsheet for recruiting long term sites is the place where anonymous identifiers are created for sites. The anonymous identifier consists of a prefix string symbolizing the category, in this case all “Long-”, a dash, and a sequence number that reflects the order when a site became a candidate. For example, “Long-43” is the forty third candidate that was approached about participating.
These identifiers are used as lookup keys in the tabular database, thus must be unique. They are the preferred identifier to put on sample containers, hiding the actual identify from casual browsers of freezers and fridges.
8. Tracking progress of long term site roster building
The HowTo Framework for Categorical Sites specifies a tracking system for building that roster. An analogous approach applies to Long Term Sites.
The various stages for a long term site are, generally in order:
- POTENTAL. Identification of a possible site worth contacting. May be reconned without contact with owner. This is when they receive a site ID number in the Long-nn series.
- CONTACTED. Initial contact by phone or drop-in.
- CONSIDERING. Only if they asked for more information before agreeing.
- AGREED. (Waiting to be sampled.)
- SAMPLED. First sampled.
- (received first data report, tracked in a different way)
- CLOSED. Eventual closeout of site.
In common between both types of site are:
- Open and confidential site identifiers
- Ecoregion name
- Presumed aquifer sensitivity rating
- Status of acquisition (most recent stage)
- Owner contact information: name, mail address, email, phone(s) (confidential); preferred contact medium for scheduling.
- Site coordinates for mapping general location (confidential).
- Qualitative rating of surrounding land use as an indicator of pesticide use intensity upgradient.
The tracking file is Box: DEC/Long term sites/DEC Longitudinal sites 2023 tracker.xlsx . This is confidential because it includes contact information and latitude/longitude. A snapshot, redacted excerpt from its tab for status tracking shows the anonymous identifiers, ecoregion, county, status as a bar graph (full width means sampled), and more site attributes included in quarterly reports to NYSDEC.
Tracking concludes, except for contact information updating, when the first sample has been taken.