Normalized Difference Vegetation Index (NDVI) 1. Today’s Goal Rancher’s in New Mexico need an...
108
Vegetation Index (VI) Normalized Difference Vegetation Index (NDVI) 1
Normalized Difference Vegetation Index (NDVI) 1. Today’s Goal Rancher’s in New Mexico need an insurance program for their grazing and haying perils RMA
Todays Goal Ranchers in New Mexico need an insurance program
for their grazing and haying perils RMA is committed to meeting
those needs Limited options Pros and Cons to both programs (RI/VI)
Can program improvements be implemented for VI Limit available
Index Intervals to assure production for the year is captured? What
time periods should be offered? Do producers prefer RI? 2
Slide 3
Where we are today? Ten Index Intervals during a year ONLY Four
intervals have been released to date Latest interval released to
date: April-June First three intervals covered winter and early
spring months that normally have very low NDVI readings as plants
are dormant or beginning to green up Above average biomass carry
over from 2010 Drought conditions in New Mexico, Texas, Oklahoma,
etc. with catastrophic impacts Impacts to the industry as a whole
3
Slide 4
History The Agricultural Risk Protection Act of 2000 (ARPA)
mandates programs to cover pasture and rangeland Vegetation Index -
Pasture, Rangeland, Forage (VI-PRF) Rainfall Index - Pasture,
Rangeland, Forage (RI-PRF) 4
Slide 5
Challenges PRF Pasture, Rangeland, Forage Crop 1. Various plant
species 2. Timing of plant growth 3. Lack of individual/industry
data 4. Vast range of management practices across the industry 5.
Publicly announced prices not available 6. Crop continuously
harvested via livestock 5
Slide 6
History Statement of Objectives issued by RMA Contractors put
together ideas and proposals 16 proposals received All were indexes
Rainfall Index Vegetation Index RMA awarded four contracts 2 were
Rainfall Indexes 2 were Vegetation Indexes 2 were implemented
6
Slide 7
Fact or Fiction Vegetation Index utilizes remote sensing
measures for the grid. All biomass in a grid is included Does not
measure grass only Deviation of normal for the interval (1989 to
2009) THIS IS NOT DROUGHT INSURANCE (Multi Peril) RMA does not use
the term drought for the Vegetation Index program nor for the
Rainfall Index program 7
Slide 8
Program Overview Area Plan of insurance Not individual coverage
Losses are area based, not producer based Index based on NDVI (a
proxy for vegetation biomass) Not measuring actual individual
production No loss adjustments, records, etc. More timely payments
Does not reward poor management practices 8
Slide 9
Program Overview Rating Each grid, index interval, and coverage
level is individually rated Encourages producers to select a
scenario that best mitigates their operation/production risks
Critical that producers select the correct interval for RI or VI
Encourage producers to view rates, BUT that should not be the
determining factor in selecting which index interval(s) to insure.
9
Slide 10
Program Overview Index Intervals Minimizes dependency on
subjective pre-determined biomass growing seasons Elevation,
climate, etc. found within an area Maintains consistency across the
country Allows for regional and local variance Allows individual
freedom to select appropriate intervals 10
Slide 11
Program Overview - VI Vegetation Index Program Area Based Plan
Approximately 8 x 8 km grid vs. county Utilizes satellite remote
sensing data Normalized Difference Vegetation Index (NDVI)
Deviation from Normal: 1989 to 2009, captures multiple perils
Review of historical indices and how they relate to your ranch is
critical Critical that peak of growing season is insured and not
time periods outside those months 11
Slide 12
Grid Overview VI Area of insurance = 8 x 8 km (~ 4.9 x 4.9
miles) 12
Slide 13
Program Overview VI (&RI) Coverage Levels Percentages
available: 90, 85, 80, 75, and 70 Consistent with other area
programs Catastrophic Risk Protection (CAT) Not currently available
Producers are eligible for NAP coverage 13
Slide 14
Program Overview Not required to insure 100% of acreage Forage
utilized in the annual grazing or hay cycle can be insured without
insuring all acreage All acres within a property may not be
productive, e.g., rocky areas, submerged areas Provides additional
flexibility for the rancher to design the coverage to their
specific needs Because the program is an area plan, there is no
opportunity to move production Producers cannot affect trigger
14
Slide 15
Index Definitions Expected Grid Index: Expected Grid Index:
Based on the historical mean accumulated NDVI values, by Index
Interval, expressed as a percentage; EGI = 100 Final Grid Index:
Final Grid Index: Based on the current NDVI values for each Index
Interval If current data represents a 40% reduction, then FGI = 60
Trigger Grid Index: Trigger Grid Index: The selected coverage level
multiplied by the Expected Grid Index i.e. - Coverage Level = 85;
then Trigger Grid Index = 85 If the final grid index falls below
the trigger grid index, the insured may be due an indemnity 15
Slide 16
Program Overview Payment Calculations The only insurable cause
of loss is when the final grid index value is less than the trigger
grid index, and only when caused by a natural occurrence If the
cause is determined by FCIC to be an act of man or intentional, a
method of assigning the Vegetation Index value from the nearest
unaffected grid will be used to establish a final grid index value
for the grid affected 16
Slide 17
2011 Changes: Filed 6/30/2010 Addition of Total Loss Factor (VI
ONLY) Accelerates the level of loss at which the maximum indemnity
amount would be made allows producers to obtain 100% payouts more
frequently VI Program expanded to balance of counties in Idaho,
Oregon and South Dakota and all counties in Arizona, New Mexico,
and Utah NO CHANGES ARE MADE TO THE PROGRAM AFTER THE CONTRACT
CHANGE DATE (CCD) (6/30) 17
Slide 18
18
Slide 19
VI Program Overview Index Intervals Crop Year divided into 10,
3-month index intervals Must select at least one interval Currently
can select up to 4 intervals Crop Practice = Index Interval Ability
for producers to manage appropriate timing risks Correlate to
individual growth patterns and production seasons The 3-month
intervals provide for greater reaction to biomass reduction events
vs. a yearly average 19
Slide 20
Technology VI (2011) USGS EROS Data: Historical Data can be
retrieved from: http://edcsns17.cr.usgs.gov/EarthExplorer/ Select
the AVHRR composites Bi-weekly composite http://ivm.cr.usgs.gov/
NDVI is band 6 in the binary image Information about the data
http://ivm.cr.usgs.gov/Metadata.dochttp://ivm.cr.usgs.gov/Metadata.doc
NDVI images are processed by the EROS data center and are not
further processed by RMA 20
Slide 21
Characteristics of the NDVI used for PRF Data from AVHRR
satellite are processed by USGS EROS and made available from 1989
to present. AHVRR data is collected daily, however the product used
is the 14-day maximum NDVI composite image Resolution of the data
is 1-km, but aggregated to 8 km for the group insurance.
Slide 22
1x1 km NDVI Grids Averaged to 8x8 km Grids 22 4 km 2 = 1.5
sections
Slide 23
Data Processing for PRF Vegetation Index NDVI data are acquired
from EROS Data Center every 14 days At the end of each indexing
interval, the NDVI images are staged for the insurance indexing.
The data are screened to remove negative NDVI values (clouds,
water, etc). Negative NDVI values are not used in the index
calculations
Slide 24
Vegetation Index Calculations Calculation of the Final Grid
Index has 3 stages: Calculation of the daily index values Averaging
the daily index to calculate the interval index Standardizing the
current interval index to the long- term average of the historical
interval indices
Slide 25
Daily Index Calculations A daily vegetation condition index is
calculated for each grid where: Daily Index i = daily vegetation
condition index for day i NDVI i = NDVI for day i NDVImin i = the
minimum NDVI across all years for day i NDVImax i = the maximum
NDVI across all years for day i Note: 200 is an arbitrary scalar...
adapted from Kogan (1990, 1995) Vegetation Condition Index
Slide 26
Daily Index Calculation Historical Maximum NDVI on June 1 NDVI
on June 1, 2011 Historical Minimum NDVI on June 1 Daily Index for
June 1= 49.7
Slide 27
Daily Index Calculation The Daily Index equation is basically
answering the question of How does todays vegetation compare to the
best and the worst conditions for this day historically as seen by
the satellite. Daily Index values near zero indicate relatively
poor condition of the vegetation compared to the history for that
day Does not mean that no vegetation is present! So, if the worst
day historically for a given day had evergreen vegetation present
such as cholla, creosote bush, and juniper, then this greenness
does not influence the daily vegetation condition index because
that greenness is the minimum value. High values indicate
relatively good vegetation condition compared to the history on
that day
Slide 28
Final Grid Index Calculation For each interval, the daily index
values are averaged for the interval of interest to calculate the
Index Interval. The Final Grid Index is then calculated by dividing
the Index Interval by the long-term average of the historical
indices for the interval in question.
Slide 29
NDVI Conditions December 2010 NDVI Image for December 14 28,
2010 EROS Data Center Indicates the gradation of greenness across
New Mexico Greener areas indicate higher levels of
photosynthesizing leaf area NDVI Departure from Long-Term Average
US Forest Service - Wildland Fire Assessment System Compares
current NDVI to long-term average Greener areas indicate NDVI is
greater than long term average. Yellow to Red areas indicate the
opposite
Slide 30
NDVI Conditions February 2011 NDVI Image for February 8 to 21,
2011 EROS Data Center Green area has declined compared to December
NDVI Departure from Long-Term Average Feb 21 US Forest Service -
Wildland Fire Assessment System Majority of New Mexico for this
period is showing above average NDVI conditions Some of the
forested areas showing below average conditions
Slide 31
NDVI Conditions April, 2011 NDVI Image for April 5 to 18, 2011
EROS Data Center NDVI Departure from Long-Term Average April 18 US
Forest Service - Wildland Fire Assessment System Majority of New
Mexico for this period is showing average to above average NDVI
conditions Eastern New Mexico is showing declining NDVI
conditions
Slide 32
NDVI Conditions June 2011 NDVI Image for May 31 to Jun 13, 2011
EROS Data Center NDVI Departure from Long-Term Average June 13 US
Forest Service - Wildland Fire Assessment System Majority of New
Mexico for this period is showing below average NDVI conditions
Eastern New Mexico is showing large departures from average
Slide 33
NDVI Conditions August 2011 NDVI Image for August 8 to 23, 2011
EROS Data Center NDVI Departure from Long-Term Average August 1 US
Forest Service - Wildland Fire Assessment System Majority of New
Mexico for this period is showing below average NDVI
conditions
Slide 34
Daily NDVI Trends Union County New Mexico Grid
Slide 35
70% 75% 80% 85% 90% Interval 645 Jan 1 to Mar 31 Final Index =
155.1 Interval 646 Feb 1 to Apr 30 Final Index = 174.6 Interval 647
Mar 1 to May 31 Final Index = 159.7 Interval 648 Apr 1 to Jun 30
Final Index = 102.2 Interval 649 May 1 to July 31 Final Index =
40.95
Slide 36
Daily NDVI Trends Torrance County New Mexico Grid
Slide 37
70% 75% 80% 85% 90% Interval 645 Jan 1 to Mar 31 Final Index =
160.5 Interval 646 Feb 1 to Apr 30 Final Index = 165.58 Interval
647 Mar 1 to May 31 Final Index = 144.8 Interval 648 Apr 1 to Jun
30 Final Index = 9.06 Interval 649 May 1 to July 31 Final Index =
53.61
Slide 38
Daily NDVI Trends Chaves County New Mexico Grid
Slide 39
70% 75% 80% 85% 90% Interval 645 Jan 1 to Mar 31 Final Index =
170.0 Interval 646 Feb 1 to Apr 30 Final Index = 152.2 Interval 647
Mar 1 to May 31 Final Index = 121.9 Interval 648 Apr 1 to Jun 30
Final Index = 77.38 Interval 649 May 1 to July 31 Final Index =
38.8
Slide 40
Daily NDVI Trends Grant County New Mexico Grid
Slide 41
70% 75% 80% 85% 90% Interval 645 Jan 1 to Mar 31 Final Index =
165.05 Interval 646 Feb 1 to Apr 30 Final Index = 145.0 Interval
647 Mar 1 to May 31 Final Index = 128.9 Interval 648 Apr 1 to Jun
30 Final Index = 106.4 Interval 649 May 1 to July 31 Final Index =
75.3
Slide 42
42
Slide 43
Program Overview - RI Rainfall Index Program Area Based Plan
0.25 degree grid vs. county Utilizes NOAA daily reported weather
data NOAA: Climate Prediction Center (CPC) Deviation from Normal:
1948 to 2009 Review of historical indices and how they relate to
your ranch is critical Critical that critical precipitation periods
are insured and not time periods outside those months 43
Slide 44
Program Overview - RI Crop Year divided into 11, 2-month index
intervals Must select at least two intervals Currently can select
up to 6 intervals Crop Practice = Index Interval Ability for
producers to manage appropriate timing risks The 2-month intervals
provide for greater reaction to biomass reduction events vs. a
yearly average 44
Slide 45
Grid Overview - RI Area of insurance = 0.25 o grids 45
Slide 46
Technology - RI NOAA CPC data NOAA wants to use the best data
available for their programs too NOAA rainfall data based on the
Optimal Interpolation (OI) methodology Historical data (1948 to
2006) currently can be retrieved from NOAA at the following
website:
ftp://ftp.cpc.ncep.noaa.gov/precip/CPC_UNI_PRCP/GAUGE_CONUS/V1.0/
ftp://ftp.cpc.ncep.noaa.gov/precip/CPC_UNI_PRCP/GAUGE_CONUS/V1.0/
Near real-time data (2006 to present) is currently accessed from
ftp://ftp.cpc.ncep.noaa.gov/precip/CPC_UNI_PRCP/GAUGE_CONUS/RT/
ftp://ftp.cpc.ncep.noaa.gov/precip/CPC_UNI_PRCP/GAUGE_CONUS/RT/
46
Slide 47
Technology - RI For 2011 results: the historical period for
calculating the long term average is Jan.1, 1948 through Dec. 31,
2009 Precipitation is interpolated to the grid and not measured
within a grid Producers MUST understand that even if there is a
weather station that reports daily to NOAA CPC inside their grid,
the results will NOT equal that weather station Similar to NASS
data used for GRP crop policies Producers reporting to NASS unknown
Surveys NASS eliminates in their quality control unknown 47
Slide 48
What we hear - RI Ranchers believe RMA is using a single point
specific weather station Ranchers provide NWS, NCDC, WFO, or other
NOAA/USGS/NASA data sets, airport weather reports, etc. Ranchers
use their own rain gauges Ranchers believe grid results will always
reflect exact conditions on their ranch Purpose: to provide general
rainfall conditions in a grid, not measure a single gauge 48
Slide 49
NAP and PRF Clarification FSA NAP Coverage and RMA PRF Pilot
Insurance Program Coverage Policy Producers can obtain both a PRF
policy (VI or RI as applicable) and NAP coverage on the same acres
for the same intended use Eligible to earn a PRF indemnity payment
and NAP benefit on the same acres for the same intended use 49
Slide 50
Web Based Tools 50 www.rma.usda.gov
Slide 51
51
Slide 52
52
Slide 53
53
Slide 54
54
Slide 55
55
Slide 56
Historical Indices and DST Actuarial information will not
change Actual Final Grid Index for past years Tools are designed to
be fluid and will change Updated annually Final Grid Index values
will reflect the change in average 56
Slide 57
Summary: Technology & Questions RI & VI Critical that
agents and producers understand the Historical and Decision Support
Tools Must spend time reviewing the historical records and
comparing those results to past production experienced by the
producer FOCUS MUST BE ON GROWING SEASON Decision to purchase MUST
be based on an analysis comparing the historical results of the
grid to a producers experience for past years production As with
any area plan results may not track 100% of the time 57
Slide 58
Growing Seasons It all comes back to growing seasons! When is
grass normally grown in a specific area? Many policies purchased in
intervals that may not be conducive to optimum forage growth Does
NRCS ecological site information help? 58
Slide 59
Growing Seasons NRCS Example Ecological Site Characteristics
Site Name: Limestone Hills (R070CY107NM) Major Land Resource Area:
070C-Central New Mexico Highlands HCPC Mixed grassland/shrubland
with scattered trees 59 Percent Forage Production by Month (%)
JanFebMarAprMayJunJulAugSepOctNovDec 0057101525 8500
Slide 60
Growing Seasons NRCS Example Ecological Site Characteristics
Site Name: Sandy Plains (R070BY055NM) Major Land Resource Area:
070B-Upper Pecos River Valley HCPC Warm-season tall and
mid-grassland mixed with shrubs and forbs 60 Percent Forage
Production by Month (%) JanFebMarAprMayJunJulAugSepOctNovDec
0035510253015700
Slide 61
Growing Seasons NRCS Example Ecological Site Characteristics
Site Name: Shallow Upland(R070AY003NM) Major Land Resource Area:
070A-Canadian River Plains and Valleys HCPC Mid-grassland with
minor components of shrubs and forbs 61 Percent Forage Production
by Month (%) JanFebMarAprMayJunJulAugSepOctNovDec 003510
253012500
Slide 62
Growing Seasons NRCS Example Ecological Site Characteristics
Site Name: Pine Grassland (R039XA012NM) Major Land Resource Area:
039-Arizona and New Mexico Mountains HCPC Grassland with ponderosa
pine overstory and scattered forbs 62 Percent Forage Production by
Month (%) JanFebMarAprMayJunJulAugSepOctNovDec 003510
253012500
Slide 63
Growing Seasons NRCS Example Ecological Site Characteristics
Site Name: Draw (R042XC008NM) Major Land Resource Area:
042-Southern Desertic Basins, Plains, and Mountains State
Containing Historic Plant Community Swale Type 63 Percent Forage
Production by Month (%) JanFebMarAprMayJunJulAugSepOctNovDec
003387182825620
Slide 64
Growing Seasons 2011 Insurance Experience Jan-Mar (645): 14% of
acres Feb-Apr (646): 03% of acres Mar-May (647): 03% of acres
Apr-Jun (648): 29% of acres May-Jul (649): 06% of acres Jun-Aug
(650): 08% of acres Jul-Sep (651): 24% of acres Aug-Oct (652): 03%
of acres Sep-Nov (653): 01% of acres Oct-Dec (654): 09% of acres
64
Slide 65
What preceded 2011?
Slide 66
2010 Jan/Feb 66 2010 Feb/Mar RI
Slide 67
2010 Mar/Apr 67 2010 Apr/May RI
Slide 68
2010 May/Jun 68 2010 Jun/Jul RI
Slide 69
2010 Jul/Aug 69 2010 Aug/Sep RI
Slide 70
2010 Sep/Oct 70 2010 Oct/Nov RI
Slide 71
2010 Nov/Dec 71 RI
Slide 72
Slide 73
2011 Jan/Feb 73 2011 Feb/Mar RI
Slide 74
2011 Mar/Apr 74 2011 Apr/May RI
Slide 75
2011 May/Jun 75 2011 Jun/Jul RI
Slide 76
January, February, March Interval -Results Released
Slide 77
Slide 78
Slide 79
Slide 80
February, March, April Interval -Results Released
Slide 81
Slide 82
Slide 83
Slide 84
March, April, May Interval -Results Released
Slide 85
Slide 86
Slide 87
Slide 88
April, May, June Interval -Results Released
Slide 89
Slide 90
Slide 91
Slide 92
May, June, July Interval
Slide 93
Slide 94
Slide 95
Slide 96
June, July, August Interval
Slide 97
Slide 98
Slide 99
Slide 100
VI vs. RI? an example 01 Roosevelt County (Grid 143144): 90%
CL: 150% PF Acres: 1,000 Dollar amount of protection/ac: $10.06
Premium: RI: $1,260 VI: $1,093 Indemnity: RI: $7,421 Grid: 16607
Intervals: Feb-Mar (FIV: 17.9); Apr-May (FIV: 01.6); Jun-Jul (FIV:
51.3) VI: $10,060 (100% payment) Interval: May-Jul (FIV: 18.9)
(Note: 650 FIV: ~38.7) 100
Slide 101
VI vs. RI? an example 02 Torrance County (Grid 135650): 90% CL:
150% PF Acres: 1,000 Dollar amount of protection/ac: $10.06
Premium: RI: $1,141 VI: $ 906 Indemnity: RI: $7,675 Grid: 17496
Intervals: Feb-Mar (FIV: 43.8); Apr-May (FIV: 06.6); Jun-Jul (FIV:
13.6) VI: $5,603 Interval: May-Jul (FIV: 56.6) (Note: 650 FIV:
~40.5) 101
Slide 102
VI vs. RI? an example 03 Lea County (Grid 151175): 90% CL: 150%
PF Acres: 1,000 Dollar amount of protection/ac: $10.06 Premium: RI:
$1,354 VI: $1,096 Indemnity: RI: $9,234 Grid: 15406 Intervals:
Feb-Mar (FIV: 05.3); Apr-May (FIV: 00.0); Jun-Jul (FIV: 16.8) VI:
$9,426 Interval: May-Jul (FIV: 33.8) (Note: 650 FIV: ~7.6) 102
Slide 103
VI vs. RI? an example 04 Harding County (Grid 125917): 90% CL:
150% PF Acres: 1,000 Dollar amount of protection/ac: $10.06
Premium: RI: $1,131 VI: $1,193 Indemnity: RI: $3,986 Grid: 19004
Intervals: Feb-Mar (FIV: 80.8); Apr-May (FIV: 26.5); Jun-Jul (FIV:
55.7) VI: $7,495 Interval: May-Jul (FIV: 45.3) (Note: 650 FIV:
~11.9) 103
Slide 104
VI vs. RI? an example 05 Union County (Grid 119038): 90% CL:
150% PF Acres: 1,000 Dollar amount of protection/ac: $10.06
Premium: RI: $967 VI: $817 Indemnity: RI: $4,023 Grid: 20207
Intervals: Feb-Mar (FIV: 49.1); Apr-May (FIV: 22.9); Jun-Jul (FIV:
96.2) VI: $10,060 (100% payment) Interval: May-Jul (FIV: 28.6)
(Note: 650 FIV: ~20.9) 104
Slide 105
Facts to Remember Possible to be indemnified for your full
guarantee under VI Preliminary results for May, June, and July
period would show 100% indemnities in many grids Full guarantees
(annual) would be very rare under RI due to the requirement that
ranchers must insure more than one interval Dual track processes to
assure data is correct prior to releasing results 105
Slide 106
Feedback and Suggestions for Possible changes Index Interval
selection (VI product) Need to insure period of MAXIMUM growth Do
we need to limit index intervals offered in NM Regional
differences? County differences? Elevation influences? Shorten the
Index Interval periods from 3 month to 2 months? Masking? Other
ideas? 106
Slide 107
Feedback and Suggestions for Possible changes Offer Rainfall
Index instead of Vegetation Index? Rainfall Index issues to think
about Potential arid region issues Could limit available index
intervals Spotty rainfall impacts Single peril lack of rainfall
only Vegetation Index issues All biomass impacts when crops &
trees are prevalent in the grid Hitting the growth season 107