Transcript AG-GEM model - University of Missouri

Adaptively Managing
Agricultural
Production for Future
Climate Change in
Montana’s Flathead
County
Zeyuan Qiu, New Jersey
Institute of Technology
Tony Prato, University of
Missouri
Dan Fagre, USGS
Tony Prato (PI), ecological economics, University of
Missouri
Zeyuan Qiu (co-PI), natural resource and
environmental economics, New Jersey Institute of
Technology
Dan Fagre (co-PI), ecology and climate science,
USGS Northern Rocky Mountain Science Center
Duane Johnson, agronomy, Great Plains Oil &
Exploration
1. To construct plausible future climate scenarios for
Montana’s Flathead Valley.
2. To develop an Adaptive Agricultural Management (AGGEM) model to: (a) evaluate the impacts of climate change
on crop production; and (b) determine how best to adapt
crop production systems to future climate change for
representative farms in the Flathead Valley.
3. To create an interactive spatial decision support tool
(ISDST) that makes the AG-GEM model and associated
geospatial databases useable and accessible to farmers and
agricultural technical service providers.
Study Area
Flathead Valley
2,000 square miles
When compared to the 20th
century average, the West has
experienced an increase in average
temperature during the last
five years that is 70 percent greater
than the world as a whole.
 Crop yields are expected to increase in crop
production areas that experience slightly higher
surface air temperature and higher summer
precipitation, and decrease in production areas
that experience significantly higher surface air
temperature, lower summer precipitation, and
reduced irrigation water supplies.
 Other things equal, crop yields increase with the
higher atmospheric CO2 concentration.
 Crop farmers can take advantage of the benefits
and reduce the risks of future climate change by
adapting their crop production systems to actual
or expected climate change.
 Representative farms and producer panels
 Crop production systems
 Evaluation and ranking of crop production systems
 Simulating crop yields, soil erosion, and water quality
 Determination of best crop production systems
 Adaptive management of crop farms for climate change
 Assessing potential benefits of adaptive management of crop
farms
 Two representative farms have been defined for
Flathead Valley.
 Two representative farms:
 164 acres (small; farm at NW Montana Agr Res Ctr)
 600 acres (large)
 Representative farms differ in terms of:
(1) size (small and large);
(2) tenure (acres owned and leased) and asset values;
(3) crop enterprises;
(4) target crop yields;
(5) mix of dryland and irrigated crops;
(6) variable and fixed costs of crop production;
(7) machinery complement and replacement strategy; &
(8) participation in federal farm programs.
 A producer panel was established for each
representative farm that consists of one to three
farmers that operate farms in Flathead Valley that have
features similar to the representative farm.
 Each producer panel assists the research team in:
(1) identifying the eight features of the representative
farm listed above;
(2) providing other information required to implement
the AG-GEM model; and
(3) formulating alternative crop production systems and
adaptive strategies for a representative farm.
 The first meeting of the producer panels was held in
February 2007.
 At that meeting the producer panels were introduced
to the project and asked to provide information about
historical crop production systems and various
parameters of the AG-GEM model.
 A second meeting of the producer panels will be held
this year to present and discuss modeling results for
the future climate period and their willingness to
adopt various strategies for adapting crop production
systems to future climate change.
 A crop production system consists of a particular mix
of crop enterprises (i.e., acreages planted to different
crops) for a representative farm.
 Historical crop production systems incorporate crop
enterprises that are suitable to growing and climatic
conditions in the local production area during the
historical climate period.
 Historical climate period: 1960-2006 (46 years)
Small Farm
30 acres of spring wheat (54 bu/acre)
55 acres of irrigated alfalfa (5 tons/acre)
15 acres of spring canola (40 bu/acre)
4 acres of camelina (40 bu or 2,000 lb/acre)
20 acres of irrigated peppermint (85 lb/acre)
40 acres of permanent pasture
164 acres
Large Farm
200 acres of winter wheat (75 bu/acre)
150 acres of spring barley (80 bu/acre)
100 acres of spring canola (2,000 lb/acre)
150 acres of dryland hay (3.5 tons/acre)
600 acres
 Future crop production systems incorporate specific
strategies for adapting the best historical crop
production system to climate change in the future
climate period.
 Future climate period: 2007-2053 (46 years)
 Crop enterprises and adaptive strategies are identified
by the research team and producer panels.
 Potential adaptive strategies include:
 using later maturing cultivars to take advantage of longer









growing seasons;
planting crops earlier and using higher seeding rates to take
advantage of higher spring temperatures and higher
precipitation;
changing crop enterprises;
adopting new crops, including those that are more drought
tolerant;
shifting the timing and scheduling of field operations to take
advantage of earlier planting dates and a longer growing
season;
reducing use of irrigation water due to higher precipitation;
improving irrigation efficiency;
altering nutrient and pesticide management practices in
response to higher temperatures and greater precipitation;
increasing crop drying and pesticide use in response to hotter
and wetter summers; and
increasing field drainage in response to higher precipitation
 Crops that were used to develop three alternative,
historical crop production systems for the two
representative farms:
barley
peppermint
oats
lentils
spring wheat
permanent pasture
winter wheat
irrigated hay
spring canola
dryland hay
winter canola
dryland alfalfa
camelina
irrigated alfalfa
 The AG-GEM model evaluates and ranks crop
production systems for each representative farm based
on four attributes of those systems:
 annual net farm income;
 variance in annual net farm income;
 annual soil erosion; and
 annual water quality (e.g., nitrogen and phosphorus in
runoff).
 The ranking procedure requires the weights for the
four attributes.
 The producer panels were asked to assign weights to
the attributes, such that the sum of the weights equals
one.
 The EPIC model is being used to simulate crop yields,
soil erosion, and water quality for each representative
farm for two soil regimes: a moderately favorable soil
regime; and a highly favorable soil regime.
 EPIC-simulated crop yields for the future climate
period are increased 2% per annum to account for
increases in crop yields due to technological change.
 The daily weather inputs for the EPIC model include
precipitation, maximum and minimum temperatures,
relative humidity, solar radiation, and wind velocity.
These data are readily available for the historical
climate period.
 Daily precipitation and temperature for the future
climate period are derived by adjusting daily
precipitation and temperature data for the historical
climate period using monthly precipitation and
temperature data for the future climate period. The
latter are from a 24-model, 12-km resolution, dataset
for the future climate period developed by the Coupled
Model Intercomparison Project (CMIP3).
• The CMIP3 dataset covers three CO2 emission scenarios:
 Daily relative humidity and solar radiation for the
future climate period are derived from the estimated
daily precipitation and temperature for that period
and geographic data for the study area.
 Daily wind velocity in the future climate period is
assumed to be same as in the historical climate period.
 The best crop production systems for a representative
farm in both the historical and future climate periods
will be determined using a multiple-attribute decision
model that integrates the Technique for Order
Preference by Similarity of Ideal Solution (TOPSIS)
method and the stochastic dominance with respect to
a function (SDRF) criterion.
 The TOPSIS method is used to calculate the relative
closeness of feasible crop production systems to the
positive-ideal solution (closeness coefficient for short)
for each representative farm.
 The SDRF criterion ranks the feasible production
systems for a representative farm by applying the
SDRF criterion to sample values of the closeness
coefficients for those systems.
 The ranking assumes the risk attitudes indicated by
the two producer panels, namely risk loving for the
small representative farm and risk neutral for the large
representative farm.
 The best crop production system for a representative
farm is the highest ranked feasible crop production
system for that farm.
 The adaptive management framework in the AG-GEM
model incorporates both ex ante and ex post adaptive
management.
 Ex ante adaptive management determines how a
farmer adapts the best historical crop production
system for a representative farm to each of the three
climate scenarios in the future climate period (i.e.,
prior to future climate change).
 The alternative future crop production systems
evaluated with ex ante adaptive management
incorporate one or more strategies identified by the
producer panel for adapting the best historical crop
production system to future climate change.
 The best future crop production system for a climate
scenario is determined by applying a multiple-attribute
decision model to the alternative future crop
production systems for that scenario.
 Ex post adaptive management dynamically adjusts
crop production systems for a representative farm to
emerging changes in climate in consecutive time
intervals (e.g., every five years).
 Ex post adaptive management takes place in real time
and, as such, is applicable only as future climate
change occurs.
 The potential benefits of adapting best historical crop
production systems for representative farms to future
climate change is assessed by ranking two alternative crop
production schemes:
(1) using the best crop production system for the historical
climate period in the future climate period, which implies
no adaptation to future climate change (first scheme); and
(2) using the best crop production system for each climate
scenario for the future climate period (second scheme),
which is equivalent to ex ante adaptive management.
 If, for a particular climate scenario and representative
farm, the best crop production system with adaptation
(scheme 2) ranks higher than the best crop production
system without adaptation (scheme 1), then
adaptation is a superior to non-adaptation for that
scenario and farm.
 Conversely, if, for a particular climate scenario and
representative farm, the best crop production system
without adaptation (scheme 1) ranks higher than the
best crop production system with adaptation (scheme
1), then adaptation is not superior to non-adaptation
for that scenario and farm.
 To date, no studies have evaluated: (1) the potential
impacts of future climate change on net farm income and
the quality of natural resources used in crop production for
representative crop farms in a local production area; (2)
how best to adapt those farms to future climate change;
and (3) whether such adaptation is advantageous.
 This situation exists due to: (1) a lack of integrated models
with which to assess the economic and environmental
impacts of climate change on representative farms and
adaptation of representative farms to climate change; and
(2) a lack of higher resolution spatial data with which to
simulate the crop production impacts of future climate
scenarios.
 The AG-GEM model alleviates the first deficiency and
the CMIP3 dataset alleviates the second deficiency.
Questions/Comments
http://cares.missouri.edu/agclim-montana