Summary (Hess)

Question: In our study area, development pressures are strong and tend to drive property values. Can payments of ecosystem services compete with those property values?

Our answers: The closer you are to already developed areas, the less likely ecosystem services payments are to compete with selling property for development. Ecosystem services payments are most likely to compete for wetlands, and agriculture and forest more distant from developed areas. Payments for carbon seq are most likely to compete, because it occurs anywhere there are trees including areas of relatively low property values.

Take-home: Need to get out ahead of development and establish contracts for ecosystem services before land prices rise and those services are actually needed.

Land use analysis
  • Note that 99% of wetlands are in top quartile and 80% highly competitive (Table 1); compare to others.
  • Use also figure 2
  • Show any of the full landscape maps (I'd use the stacked, because that's what table 1 and fig 2 refer to) adjacent to a land cover map to show the pattern relative to where development is.

ES analysis (Figure 1)
  • Carbon seq drives - we suspect this is because carbon seq is aspatial and occurs (largely) anywhere there are trees, which includes areas distant from development that have lower property values
  • N retention and pollination are very spatial and high values tend to occur near development (N retention) and valuable ag lands (pollination) where property values are higher.

Presentation note:
I'd recommend presenting in the order shown above - state the question we addressed, tell them the answer and the take-home, then show the supporting material. I've found that people seem to pay more attention to the methods and details when they know the answer first - especially if they don't like the answer. Then repeat the take-home.

One question of interest is whether payments for ecosystem services can compete with the opportunity cost of not selling property for development. This question is of particular import in areas near expanding cities. We examined this issue by comparing property value to payments for ecosystem service in our study area.

We used the output from InVEST models for carbon, nitrogen retention, and pollination to estimate payments for ecosystem services. For carbon, we considered both storage and sequestration. The estimated value ($) of carbon storage is an InVEST model output; we evaluated $<<social cost>> and $<<middle cost, based on ...>>. For sequestration, we ran the InVEST model using estimates for annual net primary productivity for each land cover in our study area to provide an estimate of one year of payments for sequestration at both the social and middle cost. We used this information to estimate the value ($) of 20 annual payments for sequestration, with a 1% discount rate reflecting current economic conditions. For nitrogen retention, we used output from the InVEST model assuming a cost of $97 per kilogram nitrogen removed for a 20-year period with a 1% discount rate. <<For pollination, we estimated the value of natural pollination services by calculating the replacement cost using managed pollinators. To do so we assumed managed pollinators cost $300 an acre for the entire growing season (####) and that all agriculture areas are pollinated by managed pollinators. We normalized the replacement cost based on the pollination ecosystem service value to distribute the costs proportionally across the landscape. These values were used to estimate the cost of 20 annual payments for managed pollination services with a 1 % discount rate. - bigsby>>

We used the tax value of each parcel of land as an estimate of property value. Triangle Land Conservancy compiled tax value data from the most recent tax appraisal for the six counties in April, 2010, and shared them with us. Tax appraisals are not coordinated among counties, the the appraisal dates varied: Durham (2008), Franklin (2004), Granville (2010), Orange (2009), Person (2005), Wake (2008). We adjusted all property values for inflation to 2010 dollars using factors from the Bureau of Labor Statistics inflation calculator (Table, 2010 Nov 24). To bring these data into alignment with the spatial scale of the InVEST model's output, we calculated the value of property for each 30mx30m pixel in our study area.

Table: Factors to bring property values to 2010 dollars.
County / year of appraisal
Inflation factor
Franklin / 2004
Person / 2005
Durham & Wake / 2008
Orange /2009 & Granville / 2010

We divided property value of the land by the value ($) of ecosystem services, per 30mx30m pixel. Thus, values larger than one indicate places where property value exceed the value of ecosystem services; where values are less than one, the value of the ecosystem services exceeds the land value. The result was arranged in order-of-magnitude categories, with the exception of the category 0.95 - 1.05 for values near one (where property value and ecosystem service value are nearly equal).

0.0 - 0.001
0.001 - 0.01
0.01 - 0.1
0.1 - 0.95
0.95 - 1.05
1.05 - 10
10 - 100
100 - 1,000
1,000 - 10,000
10,000 - 100,000
100,000 - 1,000,000

We compared property value to ecosystem service value ($) for the following combinations:
  • Nitrogen retention (20 years) - because this is generally the highest value ($) ecosystem service in our study area and because nutrient retention is an important issue here because of our drinking water supply and because markets are developing <<rough language>>
  • Carbon sequestration (20 years) - because there is a market for this, and we can imagine property owners taking a 20-year option rather than selling their property now IF the option price we high enough
  • Carbon storage + sequestration (20 years) - because markets may be developing for storage (is this true?)
  • Stacked, high values: N-retention (20 years), C-storage (social cost), C-sequestration (social cost, 20 years), pollination - to evaluate what paying the highest possible values in a system that allow stacking reveals.

For each, we present the distribution of order of magnitude categories for the whole study area and for the top quartile of ecosystem services (defined biophysically). <<Alicia, I will also get the graph you sketched out with the ES on the x-axis and the proportion of land in each category on the y-axis, stacked to 100%>>


Need to check - some of the property values might reflect being in an use value program (ag, forestry, horticulture) which artificially lowers property value because the land is being used for ag, forestry, or horticulture. The values would be less than the market value for development. <<This is an error in the "right" direction, in that it underestimates the property value so that the full value of the property is even higher than reflected in this analysis. Using full value would skew the property value / ES comparison even further toward property value.>

IMAGES are large, so I linked to the JPGs instead of displaying. These are the 20-year 1% discount rate values.

Property values ($/ha - legend needs cleaning): PropertyValues.jpg

Full Landscape
Carbon Seq (Med): CSeqMed.jpg
Pollination: Pollination.jpg
Nitrogen retention: Nitrogen.jpg
Stacked (the above 3): Stacked.jpg

ES4 Only (top quartile)
Carbon Seq (Med): CSeqMedES4Only.jpg
Nitrogen retention:NRetentionES4Only.jpg
Stacked (the above 3): StackedES4Only.jpg

Analysis files are on velocity:

Spreadsheet with analyses: PropertyValueAnalyses.xlsx

Figure 1. Comparison of property value to value ecosystem service (20 annual payments with a 1% discount rate), shown for carbon (at <<$medium value>>), nitrogen retention, pollination, and all three stacked. Results shown proportion in each category (color / legend) for the full study area (full) and the area in top quartile of ecosystem services, defined biophysically (top). Property value was divided by ecosystem service value, so that results around one occur when the values are approximately equal (yellow outlines in read); results < 1 occur when ecosystem service value are greater than property values (bottom portion of bars). Areas of NoData occur when tax value is undefined (government-owned land) or ecosystem service value is zero (dividing by zero is an undefined result).


Results / Discussion

  • Payments for carbon sequestration are competitive with property values for a much larger portion of the landscape than payments for nitrogen retention and pollination (Figure 1, compare carbon full and top quartile to N full and top quartile). This is because C-sequestration occurs throughout the landscape, including many places distant from development where property values are relatively low. Nitrogen retention, in contrast, attains its highest value proximate to development, where property values are high. As a result, the $value of C-seq is higher than the $value of N-retention for a large portion of the landscape.
  • Stacking increases the portion of the landscape for which ecosystem services payments are competitive with property values. From 34% to 42% for full landscape and from 52% to 64% for the top quartile (Figure 1).
  • Interesting that pollination is worth more than N retention - do we believe this? <<I think that pollination is more than N retention because there is a larger area where the pollinators are of high value. In that case I do we believe this. It is an interesting result, because perhaps we have to come to terms with the fact that the services people are willing to pay for (or the ones that markets are being developed for) are not the same services that would pay enough to offset the cost of selling the land>>

George 11/29 - working on one time payments (C-med and stormwater):

LULC analyses layers:

LULC results - data in the ByLULC tab of this spreadsheet: PropertyValueAnalyses.xlsx

Figure 2. Comparison of property value to ecosystem service value (20 annual payments with a 1% discount rate), shown for stacked carbon (at <<$medium value>>), nitrogen retention, and pollination. Results shown proportion in each category (color / legend) for each land use / cover category; barren and open water are omitted. Property value was divided by ecosystem service value, so that results around one occur when the values are approximately equal (yellow outlines in read); results < 1 occur when ecosystem service value are greater than property values (bottom portion of bars). Areas of NoData occur when tax value is undefined (government-owned land) or ecosystem service value is zero (dividing by zero is an undefined result).


Table 1. Proportion and area of land in each cover that is in the top quartile of ecosystem services, defined biophysically.


Proportion in

Proportion in Top Quartile
Prop/ES <= 1.05



1% (3 sq-km)


38% (1,112 sq-km)


1% (2 sq-km)


99% (43 sq-km)


Results / Discussion

Eighty percent of wetland in the study area have values for ecosystem service (20-year, 1% discount, stacked Carbon Med, pollination, nitrogen retention) greater than or comparable to property value (Figure 2), and 99% of wetlands are in the highest quartile of ecosystem service value, defined biophysically (Table 1). Simply put, purchasing wetlands is a good bet in terms of providing high-value ecosystem services. In contrast, only 10% of developed land has an ecosystem service value greater than or comparable to property value and only 1% of develop land is in the highest quartile of ecosystem services.

Neal's stuff:

Excel analysis of prop values and ES value 11/30


Here is value of ES for Pollen, Carbon-medium, and nutrient retention per lulc (Ag, Urb, Grass, For, Wetlnd). Since arc only reports the mean cell value and does not seem capable of adding all the cells I multiplied the mean times the cell count to get total value for each lulc. The downside to this is there is lots of room for rounding error. Neal updated 11/30

Results are pretty much what you would expect. :( I was hoping for something like, the best strategy for ES enhancement in Urban areas is Nutrient Retention, and Forested areas is carbon seq., and agricultural areas is pollination. Not sure I can say that with these results.

Here are extracts of tax value per LULC (Ag, Urb, Wet, For, Grass). NW- 11/30