Section I: Monitoring program development in six easy steps · integrity (Fig. Intro.1). The six...
32
5 Section I: Monitoring program development in six easy steps T his section describes how to design and implement a long-term ecosystem-based monitoring program at the landscape level (an area > 400 ha or 1000 acres; Fig. 0.1). It is based on the assumption that one of the primary objectives of the monitoring program will be to detect long-term changes in the status of three basic attributes of grassland, shrubland and savanna ecosystems: soil and site stability, hydrologic function and biotic integrity (Fig. Intro.1). The six steps Each of the first six steps illustrated in the flow chart (Fig. 0.2) and listed in the Monitoring Program Design Checklist (found at the end of this Introduction to Section I) is described in its own chapter (Chs. 1-6). The steps are listed in the order they are normally completed. Because there is no “single” way to design a monitoring program, revisiting earlier steps is often helpful. For example, the assessments completed in Step 3 often reveal issues that lead to new management and monitoring objectives (Step 1). State and transition models can be helpful here by focusing attention on areas that are at risk, or have a high potential for recovery. It is also helpful to redefine management and monitoring objectives (Step 1) for specific monitoring units identified in Step 2. Use the Monitoring Program Design Forms I (Ch. 1) and II (Ch. 4) to organize information about your monitoring program. Use the Monitoring Program Design Checklist to ensure that you have completed each step. The system allows maximum flexibility to address objectives and long-term changes, including monitoring for adaptive management, additional objectives, short-term monitoring, and monitoring threats and drivers. Figure 0.1. Landscape-scale monitoring programs should be responsive to the most important drivers, and sensitive to interactions among landscape units. by Rob Wu
Transcript of Section I: Monitoring program development in six easy steps · integrity (Fig. Intro.1). The six...
5
Section I: Monitoring programdevelopment in six easy steps
T his section describes how to design andimplement a long-term ecosystem-basedmonitoring program at the landscape level
(an area > 400 ha or 1000 acres; Fig. 0.1). It isbased on the assumption that one of the primaryobjectives of the monitoring program will be todetect long-term changes in thestatus of three basic attributes ofgrassland, shrubland and savannaecosystems: soil and site stability,hydrologic function and bioticintegrity (Fig. Intro.1).
The six stepsEach of the first six steps illustratedin the flow chart (Fig. 0.2) andlisted in the Monitoring ProgramDesign Checklist (found at the endof this Introduction to Section I) isdescribed in its own chapter(Chs. 1-6). The steps are listed inthe order they are normallycompleted. Because there is no“single” way to design amonitoring program, revisitingearlier steps is often helpful. Forexample, the assessmentscompleted in Step 3 often revealissues that lead to newmanagement and monitoringobjectives (Step 1). State andtransition models can be helpfulhere by focusing attention on areasthat are at risk, or have a highpotential for recovery. It is alsohelpful to redefine managementand monitoring objectives (Step 1)for specific monitoring unitsidentified in Step 2.
Use the Monitoring ProgramDesign Forms I (Ch. 1) and II(Ch. 4) to organize informationabout your monitoring program.Use the Monitoring Program
Design Checklist to ensure that you havecompleted each step. The system allows maximumflexibility to address objectives and long-termchanges, including monitoring for adaptivemanagement, additional objectives, short-termmonitoring, and monitoring threats and drivers.
Figure 0.1. Landscape-scale monitoring programs should be responsiveto the most important drivers, and sensitive to interactions amonglandscape units.
by Rob Wu
6
Figure 0.2. Monitoring program design and implementation (Steps 1-6) and integration withmanagement (Steps 7-10).
Monitoring for adaptivemanagement and managementby hypothesisIn addition to long-term monitoring data, adaptivemanagement requires three types of information:short-term monitoring data, knowledge of potentialthreats or drivers, and clearly defined hypotheses(predictions) of management effects (Steps 1, 3 and 7of the checklist). State and transition models (Ch.24) can be used to integrate assessment andmonitoring data with current knowledge aboutpotential management effects (based onmanagement experience, scientific studies andsimulation models) to generate these predictions.
Monitoring for additionalobjectivesMonitoring for the three basic attributes can serveas the foundation for use-specific monitoring, asillustrated in Figure Intro.1.
The basic measurements (described in QuickStart) were selected in part because they also canbe used to generate indicators related to specificuses. For example, the Line-point interceptgenerates vegetation cover and compositionindicators that are related to the quantity andquality of forage production. These indicators,together with spatial structure indicators from theGap intercept method, can be used to assess and
7
monitor wildlife habitat quality, as well as plantcommunity changes in response to fire.
The value of the basic measurements can oftenbe increased at a relatively low cost through slightmodifications (see Section IV). For example,vertical vegetation structure can be measured byadding plant height measurements to the Line-point intercept protocol (Ch. 15), or by adding theVegetation structure method (Ch. 11). In somecases, such as riparian monitoring, supplementarymeasurements (Section II) may be required.Section IV also provides recommendations foraddressing specific monitoring objectives.
Short-term monitoring (AnnualUse Records)Short-term monitoring data (listed at the end ofQuick Start) are used to make short-termmanagement changes (Steps 7 and 8). Forexample, information on residual cover or biomassis often used to decide when to move livestock to anew pasture. This information is also used tointerpret long-term monitoring data.
Monitoring threats and driversInformation on potential threats and drivers, suchas development of new roads or a change in firefrequency, is used to help identify areas where achange in management and/or monitoring will berequired. Threats and drivers are identified in Step 3.
What if I don’t have enoughtime?Nearly any monitoring is better than no monitoring.Using management and monitoring objectives toguide monitoring program design can reducemonitoring costs. A few days of careful planningoften can reduce monitoring costs by 50 percent ormore and result in much more useful data.• Use photo points where few changes are
expected (see description of state andtransition models in Ch. 24) or where yourequire only a qualitative record.
• Select measurements that are sensitive tochanges defined in the management andmonitoring objectives.
• Select measurements that generate indicatorsthat are relevant to multiple objectives. Themeasurements included in Quick Start wereselected in part because they are sensitive to
changes in the three key attributes, whilegenerating numerous indicators that arerelevant to many other objectives.
• Match monitoring frequency to expected ratesof change based on minimum detectablechange. If the smallest change in basal coveryou can detect is five percent (Ch. 4) and ittakes at least five years for this change tooccur, it’s a waste of time to repeatmeasurements more frequently.
Using State and TransitionModels for Monitoring DesignState and transition (S&T) models (Chapter 24) areconceptual models that describe the soil andvegetation dynamics for a particular type of landwith similar soils and climate. Applying S&Tmodels to monitoring program design helps a)define ecological potential, benchmarks, orreference conditions and b) specify predictionsabout the possible future change of different landunits in a landscape. This approach allowsmonitoring site selection to be based on objectivesand the ecological processes involved in landchange. Designing a monitoring program within astate and transition model framework helpsspecify the ecosystem attributes to be monitoredand other details that may vary among states andecological sites.
Applying S&T conceptual models to monitor-ing site selection minimizes monitoring expendi-tures in highly degraded states where all availableevidence suggests they will not change; and focusesmonitoring efforts in ‘at risk’ states and plantcommunities where management has the potentialto limit degradation or promote recovery. With thislogic in place, monitoring can be treated as a seriesof tests matched to specific parts of a landscape.Key components of this test are the steps used toapply S&T models to a monitoring program design.
Steps for S&T Monitoring DesignFirst, stratify the landscape (Chapter 2) into eco-logical sites or potential-based land classes. This isdone using soil surveys, landform maps, digitalelevation models and knowledge of key soil gradi-ents. Next, stratify each ecological site into statesbased on S&T models using aerial photography,remote sensing and/or field surveys. Finally, selectmonitoring methods that detect changes in focalpatterns and processes within each specific ecologi-cal site and state.
8
Monitoring Program Design ChecklistStep* Task Completed?
Develop monitoring program1 Define management and monitoring objectives
Define management objectives ..................................................................................... ________________Define monitoring objectives .......................................................................................... ________________
2 Stratify land into monitoring units (areas with similar characteristics)Assemble background information (maps, photos, management history) .................... ________________Define stratification criteria (e.g., soils, vegetation, management units) ....................... ________________Complete stratification and list monitoring units on Monitoring Program
Design Forms I and II (Chs. 1 & 4). ........................................................................ ________________3 For each monitoring unit, assess current status; identify threats and drivers; refine long-
term management and monitoring objectives; and develop/modify management strategySelect assessment system (e.g., Pellant et al. 2005) .................................................... ________________Verify that personnel have relevant qualifications ......................................................... ________________Complete assessments .................................................................................................. ________________Identify and record threats, drivers and opportunities ................................................... ________________Refine long-term management and monitoring objectives ............................................ ________________Develop/modify management strategy .......................................................................... ________________
4 Select monitoring indicators, number of monitoring plots, number of measurements, andmeasurement frequency based on objectives and resource availability
Select monitoring indicators ........................................................................................... ________________Define number of monitoring plots ................................................................................. ________________Define measurement frequency ..................................................................................... ________________Estimate time requirements ........................................................................................... ________________
5 Select monitoring plot locationsChoose and apply site selection approach .................................................................... ________________Select “rejection criteria” and use to eliminate unsuitable locations .............................. ________________
6 Establish and describe monitoring plots, and record long-term monitoring data (baseline)Establish and permanently mark monitoring plots ......................................................... ________________Describe monitoring plots and record GPS locations, including coordinate
system, datum and zone......................................................................................... ________________Record long-term data ................................................................................................... ________________Error-check and copy data and keep copies in different locations ................................ ________________
Short-term monitoring (all years)7 Record short-term monitoring data (at least 1x/year) (Quick Start) ..................................... ________________
8 Adjust management, if necessary (Quick Start) ..................................................................... ________________
Repeat long-term monitoring (every 1-5 years)9 Repeat long-term monitoring measurements (Ch. 6), compare data with Year 1 and
interpret changes (Ch. 17)Repeat long-term monitoring measurements ................................................................ ________________Copy data and keep copies in different buildings .......................................................... ________________Calculate indicators ........................................................................................................ ________________Compare with Year 1 (or previous years) ...................................................................... ________________Interpret changes using short-term monitoring data and Section III .............................. ________________
10 Refine management strategy, if necessary ............................................................................. ________________
*Steps 1-6 correspond to Chapters 1-6, except where noted.
9
Why monitor?Monitoring data are used to:• evaluate the effects of past management;• confirm effective management practices;• identify trends that can be used to predict
future changes so management can be adaptedaccordingly;
• learn more about how different factors(drought, fire, management) affect the land.
The most useful monitoring programs helpmanagers achieve long-term managementobjectives by generating relevant data.Consequently, it is essential to clearly define bothmanagement and monitoring objectives beforedesigning a monitoring program.
Use the Monitoring Program Design Form I(end of Ch. 1) to record your objectives as youdevelop them. You may find it easier to completethe stratification process (Ch. 2) before definingspecific short- and long-term objectives.
Step 1.1. Define managementobjectives(a) List the general long-term management
objective(s) for the area to be monitored on thefirst line in Monitoring Program Design Form I.What do you want the land to look like? Whatgoods and services do you want it to be able toprovide now and 100 years from now?
(b) List specific long-term management objectives foreach monitoring unit or type of land in thefifth column of the Monitoring ProgramDesign Form I (see Ch. 2 for a discussion ofmonitoring units). The long-term monitoringprogram will be designed to measure progresstowards meeting these objectives. For example,
Chapter 1
Step 1: Define management andmonitoring objectivesChecklist
1.1. Define management objectives ................................................................... ___________1.2. Define monitoring objectives ...................................................................... ___________
the specific objectives may includemaintaining or increasing the production ofparticular products (e.g., forage for livestock)or services (e.g., filtering water before itreaches streams). State and transition models(Ch. 24) can be used to help define what typesof changes are possible in different areas.
(c) List short-term management objectives that arenecessary to achieve each of the long-termobjectives for each type of monitoring unit inthe same (fifth) column of the MonitoringProgram Design Form I. Use of short-termmonitoring indicators helps ensure the short-term objectives are being met, and helpsinterpret long-term monitoring data.
Examples of management objectives are listedin Table 1.1.
Step 1.2. Define monitoringobjectivesMonitoring objectives follow directly from themanagement objectives. Additional monitoringobjectives may result from plot assessments (Ch. 3).Where possible, the monitoring objectives shouldbe quantitative. Use Appendix C to help decide ifmonitoring objectives are realistic.(a) List the general long-term monitoring objectives for
the area to be monitored in the second row ofthe Monitoring Program Design Form I. Theseshould be based on the general long-termmanagement objectives. There are threegeneral types of monitoring objectives: (i)change in average status, (ii) change in thestatus of areas with a high degradation risk, and(iii) change in the status of areas that have ahigh recovery potential. Monitoring programs
10
lareneG:tnemeganaM dnaleziminiM.snoitpoesudnalforebmunehtdnaytivitcudorpdnalesaercnironiatniaM
.ksirnoitadarged:gnirotinoM otredronilaitnetopyrevocerro/dnaksirnoitadargedhgihahtiwsaeranognirotinomsucoF
.elbissopsanoitamrofnitnaveler-tnemeganamhcumsaedivorp
foepyTtinugnirotinom sevitcejbomret-gnoL sevitcejbomret-trohS
ylhgih,peetS-htuoselbidore
sepolsgnicaf
:tnemeganaM.noisorelioseziminiM)1(
.efildliwrofytisrevidtatibahesaercnI)2(:gnirotinoM
yllaicepse,revocdnuorgnisegnahctceteDtceteD.revocrailofburhsdnalasabssarggnidulcni,seicepsevisavnifoecneserpeht
.ssargtaehc
:tnemeganaMtneiciffusniatniamotgnizarglortnoC)1(
.noisoreeziminimdnarevocdnuorgssarglainnerepetomorpotgnizargemiT)2(
elihwtnemhsilbatsednanoitcudorperrofrevochsurbegastneiciffusgniniatniam
.tatibahefildliw:gnirotinoM
dnagnirudrevocdnuorgnisegnahctceteDehtdroceR.doirepgnizargehtfodneehtta
gnizarghcaefoetaddnednagninnigeb.doirep
nairapiR :tnemeganaM.revoceertesaercnI)1(
.ytilibatsknabesaercnI)2(:gnirotinoM
gnolarevoceertni%01>fosegnahctceteD.aeranairapirehttuohguorhtdnamaertseht
forevocehtni%5>fosegnahctceteD.maertsehtgnolaseicepsgnizilibats-knab
:tnemeganaMlitnuseertfoesunosaes-gniworgtimiL)1(
.enilesworbnahtrellaterayehtsseccalanoitaercerdnakcotseviltimiL)2(
noitcapmoc/noisoreotsgnissorcdna.levargekil,setartsbustnatsiser
-knabfohtworgetomorpotgnizargemiT)3(.seicepsgnizilibats
:gnirotinoMdnaetadnoitelpmoctnemucoD
noitubirtsidlaminawenfossenevitceffedenedrah,gnicnef,.g.e(serutcurtslortnoc
yltcerid,elbissoperehW.)sgnissorchtiw,.g.e(noitubirtsidkcotseviltnemucod
dnagninnigebehtdroceR.)stnuoctapgnud.doirepgnizarghcaefodne
Table 1.1. Examples of management and monitoring objectives for a mid-elevation ranch in an area dominatedby sagebrush and perennial bunchgrasses. Similar objectives can be generated for areas in which recreation,mining and/or biodiversity conservation are the primary land uses.
Objectives
designed to primarily address the firstobjective are usually the least cost-effectivebecause a lot of effort is devoted to monitoringareas with a low probability of change.Selecting one or both of objective types (ii)and (iii) allows resources to be focused on areaswhere management is most likely to have aneffect. See Chapter 5 for more information onsite selection.
(b) List the specific long-term monitoring objectivesfor each type of monitoring unit in the sixthcolumn of the Monitoring Program DesignForm I. The potential for degradation andrecovery varies both within and amongmonitoring units. State and transition models(Ch. 24) can be used to help select appropriatemonitoring objectives for each type ofmonitoring unit.
11
Objectives
(c) List the short-term monitoring objectivesnecessary to ensure the management plan isbeing followed and to document managementchanges. Record objectives in the same sixthcolumn of Monitoring Program Design Form I.
Figure 1.1. Tallgrass prairie functioning at its highest potential, Kansas, USA. Arrow reflects lack of significantchange over time. Long-term management objective(s): Maintain biodiversity and productivity. Long-termmonitoring objective(s): Detect changes in plant cover and production by plant functional group; detect changesin plant species richness.
Figure 1.2. Overgrazed rangeland on left side of fence (b), and appropriately grazed rangeland on right side offence (c), and conversion to rain fed agriculture (a), Zacatecas, Mexico. Arrows reflect desirable and undesirablechanges from a long-term ecological sustainability perspective. Long-term management objectives: (1) Increasegrass cover for livestock forage production. (2) Avoid cultivation, which leads to a relatively irreversible thresholddue to increased soil degradation and erosion. Long-term monitoring objectives: (1) Detect changes in plantcover and production by plant functional group and vegetation spatial distribution. (2) Collect sufficient data todetect 5% change in bare ground.
X
a b c
Examples of monitoring objectives are listed inTable 1.1. Figures 1.1 and 1.2 show two additionalexamples, where arrows indicate desirable changes.
12
Imr
oF
ngise
Dmar
gor
Pg
niroti
no
M.
.tinufoepyt
hcaerof
sevitcejbognirotino
mdnatne
meganam
dnastinu
gnirotinomfo
sepyttnereffidtsiL.
mrofsihtfo
kcabeht
noseton
droceR
:)s(evitcejb
ot
neme
gana
mlarene
G
:)s(evitcejb
og
niroti
no
mlarene
G
gnir
otin
oM
*ema
Nti
nU
)2.h
C(
/lio
Se
pacsd
nal**
noitis
op
)2.h
C(
tna
nim
oD
noitate
gev)2.
hC(
tnerr
uC
sutats
)3.h
C(
)1.h
C(sevitcej
bo
tne
mega
naM
mret-g
no
L–
mret-tro
hS
–
)1.h
C(sevitcej
bo
gnir
otin
oM
mret-g
no
L–
mret-tro
hS
–
aeralato
Tf
osti
nulla(
)ema
nsi
ht
levelg
niroti
no
M)4.
hC(
)4.hC(II
mroF
ngiseD
margorP
gnirotinoMfo
nmuloctsrif
ehtot
esehtypo
C*.n
wonkfi,tnelaviuqero
etislacigolocer
O**
13
Chapter 2
Step 2: Stratify land intomonitoring unitsChecklist
2.1. Assemble background information (maps, photos, management history) ___________2.2. Define stratification criteria (e.g., soils, vegetation, management units) .. ___________2.3. Complete Stratification ............................................................................... ___________2.4. Complete Monitoring Program Design Forms I (Ch. 1) and II (Ch. 4) ...... ___________
T his chapter describes how to stratify the areainto monitoring units and decide which unitsto monitor. Data from individual monitoring
plots can be more reliably extrapolated to representlarger areas if the area of interest is stratified.
Because rangelands are among the mostdiverse ecosystems in the world, it is impossible todesign a monitoring system that perfectly reflectschanges in all landscape units. However, theaccuracy and precision of any monitoring systemcan be improved by carefully dividing the areainto relatively uniform monitoring units.
Monitoring units are areas located in aparticular part of the landscape (e.g., flood basinor hill summit), within which vegetation, soiltype, management and current status are relativelysimilar. All sections within a given monitoringunit are expected to respond similarly to changesin management and to catastrophic disturbances,such as a combination of drought and fire.Monitoring units may range in size from less thanan acre to several square miles or more.
Multiple monitoring units of the same type(e.g., hill backslope in Fig. 2.1) often repeat acrossthe landscape, geographically separated from oneanother by other monitoring units. Figure 2.1shows how a landscape unit (floodplain) wasdivided into two types of monitoring units basedon management (grazed vs. ungrazed).
Not all monitoring units will necessarily bemonitored (Fig. 2.1). For example, highly stabletypes of monitoring units (such as bedrock) mightnot be included in a monitoring program if theprimary objective is to monitor for degradationrisk or recovery (see Ch. 1). Use MonitoringProgram Design Forms I (Ch. 1) and II (Ch. 4) tokeep track of potential monitoring units.
Figure 2.1. Example of how monitoring units aredefined using landscape, soil, vegetation andmanagement criteria. In this example, threemonitoring plots, shown here as three sets of threetransects (spokes), were located in the summer-grazed floodplain monitoring unit, which has a highpotential for both degradation and recovery. Nomonitoring plots were located on monitoring units onthe adjacent slopes because they did not meet theselection criteria, which included livestock use.
Stratification: How to do itLandscape stratification is a three-step process:2.1 Collect background information, maps and
photographs.2.2 Define stratification criteria.2.3 Divide the area into monitoring units:
(a) divide the area into soil-landscape units;(b) subdivide the soil-landscape units into soil-
landscape-vegetation units (if necessary);(c) subdivide the soil-landscape-vegetation
units into monitoring units based on typeof management.
Record each type of monitoring unit from 2.3 inMonitoring Program Design Forms I and II.
14
Step 2.1. Collect backgroundinformationThe following resources are helpful in stratifyingthe landscape into monitoring units and selectingthe units to monitor. See Table 2.1 for sources ofbackground information (regularly check http://usda-ars.nmsu.edu for the most up-to-date list). Insome instances, there is a fee for these resources,but many of them can be downloaded free fromthe Internet. New sources are constantly becomingavailable.
Aerial photographs. One of the easiest ways toorganize information is on a map or recent aerialphotograph of the area, or through a GeographicInformation System (GIS). Ideally, use one or moreaerial photographs with fences and roads markedon them.
If you want to be able to locate yourself on theaerial photo using GIS and a GPS (GlobalPositioning System) unit, you will need a digitalimage that has been modified so that the distanceson the photo correspond directly to distances onthe ground (orthorectified). The most widelyavailable photographs of this type are the USGSDigitial Orthophoto Quarter Quadrangles, orDOQQs. Each of these images covers one quarterof a 7.5 minute USGS topographic map.
Satellite imagery. High resolution satellite imagerycan be used for stratification. See Chapter 25 formore information on the use of remote sensing inmonitoring.
Written and oral histories. Information onhistoric changes can help predict which parts ofthe landscape are most likely to change in thefuture. Sources of information on historic changesinclude old monitoring records (often stored inthe local Bureau of Land Management [BLM] orUnited States Forest Service [USFS] offices), oldaerial photographs and survey records. Interviewswith current and previous land managers areamong the most valuable sources of information.
Property maps. Conservation plan maps (availablefrom NRCS offices) locating current and historichomesteads, fence lines, corrals, roads, wateringholes, supplemental feeding locations, and areas
seeded, herbicided or where vegetation wasremoved are valuable when stratifying thelandscape into monitoring units. All of these havethe potential to affect the way land will respond tofuture management.
Species lists. Lists of plant species commonlyfound in the area are helpful. Vegetationmeasurements are usually recorded to the specieslevel. At a minimum, lists of potential invasivesand exotics should be acquired for all monitoringprograms.
Ecological Sites and Site Descriptions (ESDs).Each ecological site includes several similar soils.Each ESD includes partial species lists and basicsoils information and state and transition modelsthat can be used to help plan and interpretmonitoring (see end of Introduction).
Soil maps. Soil maps are commonly available inthe form of county soil surveys. Soil maps areoften drawn on aerial photos. In addition to maps,soil surveys have a wealth of information on soilproperties and the suitability of soils for differentuses. GIS layers of soil surveys can be obtained formost counties from the local NRCS office.
Soil maps of pastures and rangelands rarelyinclude map units named with a single soil seriesdue to the complexity of most rangelandlandscapes (a soil series is like a plant species).Instead, individual areas are mapped as“complexes” or “associations” of two or more soilmap unit components. Soil map unit componentsare phases of soil series. Phases of soil series areusually identified based on features important formanagement, such as slope, soil surface texture,surface rockiness and salinity. A soil map unitcomponent is like a plant subspecies. The soilsurvey (or a professional soil scientist) can helpyou decide if the components in a particular mapunit are sufficiently similar to be treated uniformlyfor monitoring purposes.
Soil series are distinguished based on soilprofile characteristics. These characteristics areusually, but not always, directly related to soilfunction. Soil series allow us to access referenceinformation included in Ecological SiteDescriptions and other databases.
Land stratification
15
Land stratification
.1.2elbaT ,91rebmetpeSfosadilaveraesehT.egnahcnetfoskniltenretnI.secruosernoitacifitartsepacsdnaL.8002
secruoseR secruoS
*sotohplaireA
•••
•
taSGSU rerolpxEhtraE/vog.sgsu.rc.71snscde//:ptthtasotohpSGSUgnillesseinapmoC lmth.enilnoweiv/oig/vog.pamlanoitan//:ptth
slis/vog.sgsu.ksa//:ptth _ lmth.xedni , lmth.secruos/vog.sgsu.ksa//:ptth llacro,yhpargotohPlaireAlanoitaNehtmorfdeniatboebnac6991nahtrewensegamI.SGSU-KSA-888-1
rerolpxEhtraEnoelbahcraeseradna,)PAHN(yhpargotohPedutitlAhgiHlanoitaNdna)PPAN(margorPta rerolpxEhtraE/vog.sgsu.rc.71snscde//:ptth
)108(ro,0202-91148,TU,ytiCekaLtlaS,htuoS0032tseW2222,OFPAASFADSU,hcnarBselaSADSUro,0092-448 gnidnal=cipot&gnidnal=tcejbus&emohofpa=aera?ppaofpa/ASF/vog.adsu.asf.www:ptth
latigiD:sotohplaireAretrauQotohpohtrO
*)QQOD(elgnardauQ
•
••
ehtllatignivig,deifitceroegdna)retupmocaotnidennacs(dezitigidneebsahtahthpargotohplaireanA.sepacsdnalyfitartsotygolonhcetSIGgnisunehwlufpleherasQQOD.pamafoseitreporp
tasrentrapssenisubstiroSGSU lmth.xedni/vog.sgsu.sore//:ptthtaSCRNADSU lmth.xedni/stesatad/stcudorp/vog.adsu.scrn.cgcn.www//:ptth
spamcihpargopoT••
spamcihpargopotSGSUetunim5.7 vog.sgsu.spamopot//:ptthtasrentrapssenisubstiroSGSUmorfDCroypocdrahaivdesahcrupebnacspamcihpargopotrehtO
//:ptth lmth.xedni/vog.sgsu.sore
cihparGretsaRlatigiD)GRD(
•
•
,deifitceroegdna)retupmocaotnidennacs(dezitigidneebsahtahtpamcihpargopotSGSUdennacsA.snoitacilppaSIGrofydaer
:tasrentrapssenisubstiroSGSU grd/vog.sgsu.spamopot//:ptth
syevruslioS*spamdna
•
•
•
ADSU,erutlucirgAfotnemtrapeD,tnemnrevoGsetatSdetinUrednukool(eciffoSCRNlacolehttisiVSCRNehtkcehcro,)koobenohpehtfosegapeulbehtniecivreSnoitavresnoCsecruoseRlarutaN
(etisbew yevrus/vog.adsu.slios//:ptth .tseretnifoytnuocehtrofyevrusliosafoypocaniatboot).saeratsomrofelbaliavasi)000,052:1(egarevocpam)esabataDcihpargoeGlioSetatS(OGSTATS
lacolhguorhtelbaliavaeraseipocdraH.dezitigidgniebfossecorpehtnieraspam)000,42:1(OGRUSS.seciffoSCRN
seciffoemoS.tseretnifoaeraehtrofyevruSmetsysocElairtserreTaniatbooteciffoSFSUlacolehttisiV.mroflatigidnielbaliavaatadsihtevahyam
yrotnevnInoitategeV*ataD
•
•
noitategevdetcelloc-dleiffospameraesehT.spam)MIVS(dohteMyrotnevnInoitategeVlioS:dnalMLB.mrofSIGnielbaliavaatadsihtevahyamseciffoemoS.atadyrotnevni
:tadnuoferaatadyrotnevnIecruoseRlarutaNdnaspamsutatsSCRN:dnaletavirPsecruoseratad/lacinhcet/vog.adsu.scrn.www
spamlareneG• dnaLfouaeruB,roiretnIfo.tpeD,tnemnrevoGsetatSdetinUrednukool(spamsutatsdnalMLB
.)koobenohpehtnisegapeulbehtnitnemeganaM
stsilseicepS
•••••
.)sdrocergnirotinomdloyllaicepse(seciffoSCRNdnaMLB,SFSU:stnalpfostsilSCRN secruoseratad/lacinhcet/vog.adsu.scrn.www
.woleb)SCRN(snoitpircseDetiSlacigolocEeeS(tayteicoStnalPevitaNehtforetpahclacolruoypukooL gro.spnan.www )
(esabatadlanoitanSTNALP vog.adsu.stnalp//:ptth )
)egnaR(lacigolocE*snoitpircseDetiS
•
•
otogro,ediuGlacinhceTeciffOdleiFehtnidetsilsasnoitpircsedrofksa(eciffoSCRNlacoLvog.adsu.voge.cs.sise//:ptth )
.bewehtnoebteytonyamsnoitpircseddesiveremoS
spaMcigoloeG • taspaMcigoloeGSGSU vog.sgsu.bdmgn//:ptth
stsiLseicepSevisavnI • SCRN revirDsuoixon/avaj/vog.adsu.stnalp//:ptth
.SCRNdnatcirtsidnoitavresnoclacolehthguorhtdepolevednalPnoitavresnoCnworiehtotreferoslanacsrenwodnaL*
16
Land stratification
Step 2.2. Define stratificationcriteriaThere is virtually an infinite number of strategiesfor stratifying the landscape into functionallysimilar monitoring units. Three criteria useful for awide variety of ecosystems are: soil-landscape,current vegetation and management.
Soil-landscape criteria include topography,landscape position and soils. These criteriadetermine the potential of the unit to supportdifferent plant communities. Incorporating soil-landscape criteria is a very important step,especially in areas where the same plantcommunity currently dominates much of theland. In these areas, knowledge of the underlyingsoils can help identify locations where there is ahigh recovery potential.
In most systems, historic differences inmanagement and disturbance have generatedvariability in current vegetation within soil-landscape units. Historic management anddisturbance can be used as stratification criteria, ascan current and planned future management.
While stratification may sound complex, inreality it is relatively simple.
Step 2.3. Complete stratification:divide the area into monitoringunitsThis step is often broken into separate parts, basedon the number of stratification criteria. In thefollowing example, three criteria were used.Remember that a single type of monitoring unitmay include many individual units scatteredacross a landscape.
Step 2.3(a) Divide the area into soil-landscapeunits (NRCS ecological sites or functionallysimilar units such as the unit used in the USFSTerrestrial Ecosystem Survey). Landscape units areareas that are relatively homogeneous with respectto slope, aspect and parent material (material fromwhich the soil was formed). As a result, theygenerally have similar soil series, or similar soilcomponents. Where soil series or soil componentsin a landscape unit are functionally similar, they
are included in the same soil-landscape unit.Functionally similar soils have relativelyequivalent potentials to produce a particular typeand amount of vegetation under the same climate.
Soil-landscape units generally correspond toNRCS “ecological sites” (previously referred to as“range sites”). These are also similar to the unitsused in the USFS Terrestrial Ecosystem Surveysystem and to soil-landscape-based landclassification systems developed in New Zealand,Australia and other countries, although some ofthese systems also use current vegetation (see Step2.3b). The grouping of functionally similar soilsinto ecological sites has already been completed inmost areas of the United States, although thespecific criteria used to create unique ecologicalsites varies somewhat among different states.
Soil-landscape units repeat across thelandscape (Fig. 2.2). For example, multiple areason south-facing 10-15% slopes, with 30-50 cm(12-20 in) of soil over granitic bedrock, would beclassified as the same soil-landscape unit.
Step 2.3(b) Subdivide the soil-landscape unitsinto soil-landscape-vegetation units (ifnecessary). Vegetation is generally correlated withlandscape position and soil type, but historicdifferences in land use can lead to thedevelopment of different plant communities onthe same soil-landscape unit (Fig. 2.3; see alsoCh. 24). Vegetation subdivisions are normallybased on the current dominant plant species thatdefine the community. They can also be based onthe presence of critical species, such as exotic orinvasive plants, or by habitat type for a particularanimal. Keep in mind that while soil-landscapeunits are relatively persistent and use-independent, soil-landscape-vegetation units canand do change rapidly.
Step 2.3(c) Subdivide the soil-landscape-vegetation units into monitoring units based onmanagement (soil-landscape-vegetation-management units). A monitoring unit is thelargest contiguous area with the same soil type andplant community that is expected to respondsimilarly to management changes. Pasture borders,distance from water, prescribed fire, woodyvegetation removal and recreational use can be
17
Land stratification
Figure 2.3. Example of the subdivision of landscapeunits (box in Fig. 2.2) into landscape-vegetation units.Here one of the Hills landscape units was subdividedinto landscape-vegetation units.
Figure 2.2. Example of landscape unit stratification.This type of stratification can only be done with aerialphotos. Subdivision into soil-landscape units was notpossible due to lack of soil survey information. Theuse of Soil Survey Maps can make this processeasier and more accurate.
Figure 2.4. Example of the subdivision of landscape-vegetation units into different types of monitoringunits (1-4) based on management. In this case, oneof the Hills-Pinyon-Juniper Savanna units wassubdivided based on the presence or absence ofprescribed fire; and the Hills-Blue grama Grasslandunit was subdivided based on whether or notwoodcutting is planned.
used to delineate monitoring units. Similarmonitoring units (same type) often repeat acrossthe landscape (Figs. 2.1 and 2.4). Figure 2.4 showsfour types of monitoring units.
Step 2.4. Record each type ofmonitoring unit in the MonitoringProgram Design Forms I and II(Chs. 1 and 4)Each type of monitoring unit is recorded onlyonce, even if it repeats across the landscape. Leaveextra rows on Monitoring Program Design Form IIbelow monitoring units in which you expect toinclude more than one monitoring plot.
Hills Monitoring Units1 = Pinyon-Juniper Savanna
No prescribed fire2 = Pinyon-Juniper Savanna
Prescribed fire3 = Blue grama Grassland
No woody removal4 = Blue grama Grassland
Woody removal
18
Chapter 3
Step 3: Assess current statusChecklist
3.1. Select assessment system ............................................................................. _________3.2. Verify that personnel have relevant qualifications ..................................... _________3.3. Complete assessments ................................................................................. _________3.4. Identify and record drivers, threats and opportunities .............................. _________3.5. Refine long-term management and monitoring objectives ....................... _________3.6. Develop/modify management strategy ....................................................... _________
W here possible, the status of each area ofeach monitoring unit (or at least eachtype of monitoring unit) should be
evaluated and recorded in the MonitoringProgram Design Form I (Ch. 1). This evaluationhelps determine the relative usefulness ofestablishing transects in each monitoring unitbased on the objectives identified in Step 1.
Assessments can be qualitative or quantitative.Assessments can use current status, apparenttrend, or trend based on existing monitoring data.All assessments require some kind of reference.Where trend is used, the reference is the status atsome previous time. The reference for the currentstatus is generally the site potential, which isdefined based on soil and climate (e.g., in NRCSEcological Site Descriptions as discussed in Ch. 2).
Step 3.1. Select assessmentsystemThere are a number of protocols currentlyavailable for assessing rangelands. We haveincluded brief descriptions of two we consideruseful: Interpreting Indicators of Rangeland Health(IIRH) for uplands (Pellant et al. 2005; see alsoPyke et al. 2002) and Process for Assessing ProperFunctioning Condition (PFC) for riparian areas(Prichard et al. 1998a, b). These protocols wereselected because they emphasize the capacity ofthe system to function relative to its potential. Inother words, they reflect the current status of thesame fundamental ecosystem attributes that thismonitoring protocol is designed to address. Theyare both at present (2004) widely applied bygovernmental and non-governmental
organizations in the United States. IIRH has beentranslated into Spanish and applied in Mexico.
Both of these protocols, like all qualitativesystems, should be applied by a team of trainedpersonnel with a working knowledge of the localecosystem. Links to PDF (portable documentformat: documents in a format easily downloaded,viewed and printed from the World Wide Web)files of these protocols and training informationare available on the Internet (http://usda-ars.nmsu.edu).
Upland areas. Interpreting Indicators of RangelandHealth (Pellant et al. 2005) (Fig. 3.1). Thispublication describes a process for using 17qualitative indicators to generate assessments ofthe same three attributes addressed by thismonitoring manual: soil and site stability,hydrologic function and biotic integrity. Astandard or reference is established for eachecological site (type of soil-landscape unit).Reference information for each of the 17indicators is summarized in a “Reference Sheet.”Each indicator is placed into one of five categoriesbased on its relative departure from its referencestatus (none to slight, slight to moderate, etc…).Specific combinations of the 17 indicators are thenused to evaluate each of the three attributes.
Reference Sheets for some ecological sites havealready been developed in the United States andMexico. In the U.S., they are included in theupdated NRCS Ecological Site Descriptions.Instructions for developing Reference Sheets wherethey do not already exist are included in the latestversion of IIRH (version 4.0). This method isincluded only to assist in the identification and
19
selection of potential monitoring sites (Ch. 5). Theindicators described should not be used to replace thequantitative monitoring indicators described in thismanual. For additional information on how to applythis method, please refer to the IIRH publication.
Assessment
Riparian areas. Process for Assessing ProperFunctioning Condition (Prichard et al. 1998a, b)(Fig. 3.2). This publication describes a process fordeveloping riparian qualitative assessments. It isalso based on 17 indicators. There are two primarydifferences, though, to the upland areasassessment protocol (IIRH). The first is that,instead of generating a “degree of departure” fromthat expected for the ecological site, the evaluationis designed to rate a stream reach as functional, atrisk or non-functional. The second difference isthat there is no standard reference. The teamcompleting the evaluation must develop a uniquestandard for each area to be evaluated. For thisreason it is essential that a diverse team of trained,knowledgeable and experienced individualscomplete the evaluations for riparian areas.
Figure 3.1. Cover of Interpreting Indicators ofRangeland Health (Pellant et al. 2005).
Step 3.2. Verify that personnelhave relevant qualificationsRelevant evaluator qualifications are listed in eachdocument. It is important to recognize thatexperience and long-term knowledge of theecosystem is often as important as academicqualifications. Academically trained individualswith little field experience will find it difficult toaccurately and consistently apply assessmentprotocols.
Step 3.3. Complete assessmentsPaper and electronic forms are available forcompleting the assessments.
Where? It is more important to completeassessments in areas where the value ofmonitoring and/or a change in management isuncertain. If you already know that an area is in arelatively stable state, it’s usually not worthcompleting an assessment. Be sure to justify allassessments with comments and observations.
Figure 3.2. Cover of Process for Assessing ProperFunctioning Condition (Prichard et al. 1998a, b).
20
Assessment
Both the upland and riparian assessmentsystems are designed to evaluate individuallocations. Record additional notes of off-site effectsand impacts to describe relationships amongmonitoring units. For example, excessive runoff inone monitoring unit may reflect problems in anupslope monitoring unit, or the presence ofinvasive species on one monitoring unit may posea risk to adjacent monitoring units.
Step 3.4. Identify and recorddrivers, threats and opportunitiesA critically important part of the assessmentprocess is identifying drivers, and current andfuture threats and opportunities. Both of theassessment protocols are limited to current statusonly. Areas likely to be threatened by futureactivities, or where future activities present newopportunities, should be considered formonitoring because of their potential for change.
Drivers. Drivers include all factors that cancontribute to changes in the properties andprocesses to be monitored. Typical drivers inrangeland ecosystems are listed in Figure 0.1.Drivers may or may not be threats.
Threats. Threats are drivers that might negativelyimpact the land in the future. Future threats mightinclude increased off-road vehicle activity, invasiveplants that have been identified in the area,cultivation (see Fig. 1.2), overgrazing by wildlife/livestock associated with a change inmanagement, or drought and insect damage. Thelevel of each threat usually varies amongmonitoring units. For example, off-road vehicleactivity is less likely to be a threat on isolatedmesas, and the threat of insect damage isfrequently greater in grass-dominated ecologicalsites. Gully formation is more likely to occur inmonitoring units located downslope of areaswhere an increase in runoff (e.g., associated withroad construction) is anticipated.
Invasive species sometimes pose a high threatin particular soil types. Disturbance can favor theestablishment of invasive species. For example,road graders can disperse African rue (Peganumharmala) rhizomes. Additionally, cheatgrass
(Bromus tectorum) seeds are often dispersed bygrazing animals. Thus it pays to consider allpotential threats and drivers when designing amonitoring program.
Opportunities. New opportunities are often moredifficult to predict than threats, but are at least asimportant to address in a monitoring program.Opportunities might include grants for restorationthat can only be applied to particular areas (e.g.,riparian). A new neighbor or the development of agrass bank in the region might bring newopportunities for cooperative livestockmanagement. Climate change and even short-termweather patterns can be viewed as both threatsand opportunities.
Identifying known or potential futureopportunities for a monitoring unit may influenceyour decision to monitor. Knowledge of suchopportunities can allow flexible management touse them. If monitoring data are collected prior toand following a management change, the effectsof the new management can be quantitativelyevaluated.
3.5. Refine long-termmanagement and monitoringobjectivesNew information can be provided by on-siteassessments and the development of a list ofthreats and opportunities for each monitoringunit. This information can be used to refinemanagement and monitoring objectives. Thesechanges should be recorded in the MonitoringProgram Design Form I (Ch. 1).
3.6. Develop/modifymanagement strategyThe management plan should be finalized (to theextent possible) before beginning site andindicator selection. At the risk of redundancy, werepeat that in order for monitoring to be cost-effective,it must focus on those areas, properties and processesthat are likely to change in response to management(including lack of active management).
21
Chapter 4
Step 4: Select indicators andnumber of measurementsChecklist
4.1. Select monitoring indicators ....................................................................... _________4.2. Define number of monitoring plots ............................................................ _________4.3. Define measurement frequency .................................................................. _________4.4. Estimate time requirements......................................................................... _________
I ndicator selection should be based on theobjectives defined in Step 1 (see Ch. 1). It isimportant to think carefully about what you
need to learn from your monitoring program, andhow precise the data need to be.
Types of indicatorsTwo basic types of monitoring indicators areaddressed in this manual: short-term and long-term. Some (like plant cover) can serve as short-and long-term indicators. The difference betweenshort- and long-term indicators is discussed inQuick Start and in Step 4.1.
In addition to the short- and long-termindicators described in this manual, you may wantto include indicators of potential threats and newopportunities. These are briefly described inChapter 3. Information on threats andopportunities can be used to anticipate futurechanges and adapt monitoring and managementaccordingly.
Reducing monitoring costsThe most effective way to reduce monitoring costsis to minimize the number of measurements.Selecting measurements that generate indicatorsaddressing multiple objectives can minimize costs.For example, the Line-point intercept methoddescribed in Quick Start can be used to generate
Note: Steps 4 and 5 (Chs. 4 and 5) are often completed simultaneously. The number of transects thatcan be monitored often depends on where they are and how many different types of measurements areto be made on each transect. Different types of monitoring units sometimes require differentmeasurements. We suggest reading through Chapter 5 before actually beginning the tasks listed inChapter 4.
ground cover indicators that are important (1) forerosion prediction; (2) for plant cover and speciescomposition; and (3) as an indicator of wildlifehabitat structure. Habitat structure requires theaddition of height measurements to the Line-pointintercept method (Ch. 15).
The measurements described in Quick Start aresufficient to generate all of the indicators requiredfor most monitoring objectives. In many cases,indicators generated from the Quick Startmeasurements can substitute for the more time-consuming measurements described in thefollowing chapters. For example, the Single-ringinfiltrometer (Ch. 8) is a direct measurement ofhow quickly water will soak into the soil(infiltration capacity), but it is very timeconsuming. The Soil stability test (Quick Start) isless time consuming and, together with indicatorscalculated from the Line-point and Gap interceptmeasurements, can generate information relevant tothe infiltration capacity of the soil (see Section III).Another option is to make the more time-consuming measurements (generally Level 4 inTable 4.1) at a few high-priority locations.
Monitoring Intensity (Table 4.1). Where onlyqualitative documentation of change is required,photographs (Level I) are often sufficient. Level IImonitoring intensity (semi-quantitative) is
22
appropriate where only the core indicatorsincluded in Quick Start are required, and wherethe data will always be collected by the sameperson. Level III monitoring intensity is the sameas Level II (i.e., Quick Start methods), except thatthe measurements are more precise and repeatable.
In many cases, only a subset of Level II or IIImeasurements is necessary. For example, wherethe primary concern is a change in woody shrubcover, Line-point intercept (Level III) or step-point(Level II) alone is often sufficient if woody speciescomprise at least five percent of the foliar cover.The Belt transect (Level II or III) is appropriatewhere the only concern is early detection ofundesirable plant establishment, or when thespecies/ functional group you wish to monitor isvery sparse (less than five percent cover).
Level IV measurements are usually included toaddress specific concerns or objectives that cannotbe addressed using the basic measurements.
Step 4.1. Select monitoringindicatorsThe monitoring indicators selected will determinewhich measurements are needed. Selectingmeasurements that generate multiple indicators,or that generate indicators that address multipleobjectives, can often reduce costs.
Table 4.2 lists the measurements described inboth volumes of this manual and briefly describes
.1.4elbaT .ytisnetnignirotinomfosleveL
leveL evitcejbO stnemerusaeM
I nisegnahcegralfonoitatnemucodevitatilauQ.erutcurtsnoitategev
.stniopotohpdradnatstashpargotohP
II nisegnahcfonoitatnemucodevitatitnauq-imeSytilibatsliosdnaerutcurts,noitisopmocnoitategev
.)IIIleveLnahtelbataeperssel(
cisabotsevitanretlaevitatitnauq-imeS.)tratSkciuQnidebircsed(stnemerusaem
III nisegnahcfonoitatnemucodevitatitnauQ.ytilibatsliosdnaerutcurts,noitisopmocnoitategev
evitatitnauqcisabruoffoeromroenO-eniL:tratSkciuQnidebircsedstnemerusaemtsetytilibatslioS,tpecretnipaG,tpecretnitniop
.tcesnarttleBdna
VI ehtnisegnahcfonoitatnemucodevitatitnauQretaw,noitcapmoc,.g.e(seussicificepsfosutats
knabmaertsronoitcudorpevitategev,noitartlifni.)ytilibats
.51-7sretpahCeeS.suoiraV
Indicator selection
the relevant monitoring objectives for each. It alsoincludes some of the indicators that can begenerated from each measurement. UseMonitoring Plot Design Form II (end of Ch. 4) andTable 4.2, together with your objectives (outlinedin Monitoring Program Design Form I, Ch. 1) andthe results from your assessment, to select theappropriate measurements for each monitoring unit.
Short-term. Short-term indicators should reflectshort-term management objectives. Mostmanagement plans require very few short-termindicators. For example, if management calls foreliminating off-road vehicle traffic from an area,the only indicator you need to monitor is vehicletracks (modified Belt transect, Gap intercept orsimply recording the number of tracks per 100paces). For livestock grazing in arid and semi-aridecosystems, residual ground cover (step-pointtransect), together with stocking rate information,is often sufficient. Typical short-term indicatorsare listed on the form at the end of the Quick Startvolume.
Long-term. Long-term indicators should reflectlong-term changes in the landscape caused bychanges in management, climate and so on.Monitoring objectives (Ch. 1), together withassessment results (Ch. 3) and state and transitionmodels (Ch. 24), can be used to help identifyappropriate indicators.
23
Indicator selection.2.4elbaT tnemerusaemdetamitsesedulcniCxidneppA.srotacidnidnastnemerusaemfoweivrevO
nidetsilsrotacidniehtrofstnemeriuqer dlob .rotacidnihcaefonoitinifedarofyrassolGehtro71retpahCeeS.
tnemerusaeM …edulcnI rotacidnI
setubirttA
&lioSetiS
ytilibatS
-ordyHcigol
noitcnuF
citoiBytirgetnI
tniop-eniLtpecretni
tratSkciuQ()51.hCdna
retaw,ksirnoisoreliosrof…seicepsnisegnahc,noitartlifni
ni,.e.i(revocronoitisopmoc)smargorpgnirotinomllaylraen
)%(revocrailoF)%(revoclasaB)%(dnuorgeraB
)%(revocdnuorG).on()etamitsemuminim(ssenhcirseicepS)seicepsyb(stpecretnitnalpdaedfonoitroporP
seicepsro,puorglanoitcnufybrevoC)%(.cte,ciffart,gnizarg,erifottnatsiser
)%(revocrettiLthgiehegailofdnanoitcurtsbolausiV)derusaemthgiehnehw(ytisrevid
XXXX
XXXXX
X
XXX
XXX
XX
lasaBdnatnalPtpecretnipaG
)tratSkciuQ(
citoxednanoisoredniwrof…,)yponac(ksirnoisavnitnalp
ksirnoisoreretawliosrofdna)lasab(noitartlifniretawdna
)%(mc52>spagyponacniecafruslioS)%(mc05>spagyponacniecafruslioS
)%(mc05>spaglasabniecafruslioS)%(mc001>spaglasabniecafruslioS
tf___romc___>spagniecafruslioS
XXXXX
XXXXX
XXXXX
tsetytilibatslioS)tratSkciuQ(
,)htob(ksirnoisoreliosrof…gnilcycrettamcinagro
citoiborcimdna)ecafrusbus()ecafrus(tnempolevedtsurc
)ssalc(ytilibatsecafruS)ssalc(ytilibatsecafrus-buS
6ssalc=seulavecafrusfonoitroporP
XXX
XXX
XXX
tcesnarttleB)tratSkciuQ(
seicepsnisegnahctcetedot...,.g.e(ytisnedrorevocwolhtiw
)sevisavnifonoitcetedylrae
ytisnedtnalPssalcezisybytisnedtnalP
XX
–tsetnoitcapmoCtcapmi
retemortenep)7.hC(
asinoitcapmocliosnehw…melborplaitnetoprotnerruc
tnemercnihtpedrepsekirtsforebmuNseiponactnalp-rednu:secapsretnifooitaR
XX
XX
XX
gnir-elgniSretemortlifni
)8.hC(
yltnerrucsinoitartlifnierehw…liosybdetimilyllaitnetopro
erutcurts
)rh/mm(etarnoitartlifnIseiponactnalp-rednu:secapsretnifooitaR
XX
noitcudorptnalP)9.hC(
gniyrracerovibrehrof…dnasetamitseyticapacwolfygrenemetsysoce
tnalpybnoitcudorpdnanoitcudorplatoT,.g.e(spuorglanoitcnufdnaseiceps
)egarof)etamitsemuminim(ssenhcirseicepS
X
X
X
X
seicepstnalPssenhcir)01.hC(
fosetamitseesicerprof…tniop-eniLees(ssenhcirseiceps
)noitcudorptnalPdnatpecretni
ssenhcirseicepS X
noitategeVerutcurts
)11.hC(
forotacidnidradnatsrof…tniop-eniLees(revoctatibah
)tpecretni
noitcurtsbolausiVytisrevidthgiehegailoF
XX
ytisnedeerT)21.hC(
ylediwootsnoitalupoprof…stcesnarttleBrofdesrepsid
ytisnedtnalPssalcezisybytisnedtnalP
XX
lennahcnairapiRyevrusnoitategev
)31.hC(
noitategevgnitnemucodrof…sknabmaertsgnolaegnahc
)%(revocrailoF,ydoow,.g.e(puorglanoitcnufybrevoC
)%().cte,seicepsgnizilibats-knab
XX
XX
XX
lennahcnairapiReliforpyllugdna
)41.hC(
siygolohpromlennahcerehw…seillugroegnahcotdetcepxe
gnirevocerrogninepeedera
oitarhtped-htdiWelgnaknaB
XX
XX
24
Indicator selection
For example, many land managers in thewestern United States need to identify andmonitor grass-dominated states that are at risk ofchanging to shrub-dominated states, which areassociated with higher erosion rates. State andtransition models define the states and transitionsfor the area of interest. The assessment would helpidentify areas potentially at risk of a change instate. The assessment, as well as the state andtransition model, assist in identifying indicatorsassociated with a state change (e.g., grassmortality, reduced infiltration and/or shrubestablishment). The qualitative indicators includedin the Interpreting Indicators of Rangeland Healthprotocol help focus attention on processes and theassociated properties that should be monitored(Pellant et al. 2005).
Step 4.2. Define number ofmonitoring plotsDefining the number of monitoring plots is abalancing act between what changes need to bedetected (benefits) and the resources available(costs). Use the factors listed below, along withAppendix C, to determine the number of plotsneeded. The number of short-term monitoringplots should be determined separately from thenumber of long-term monitoring plots. Afterdetermining time estimates (Step 4.4), it may benecessary to revisit this step to reduce costs.
Short-term. Use the recommendations listed forlong-term measurements (below and in AppendixC) as a general guide for how many measurementsyou need. As with long-term measurements,monitoring more locations (plots) is generallybetter than increasing the number ofmeasurements at each plot.
Long-term. The number of measurements requireddepends on four factors:
(1) the amount of variability within theecological site (lower variability requiresfewer measurements);
(2) the size of the change you want to detect(larger minimum changes require fewermeasurements for detection);
(3) how sure you want to be that if you say achange has occurred (or has not occurred),
you’ll be right (statistical certainty – lesscertainty requires fewer measurements);
(4) whether you want to detect change at theplot scale (a plot selected to represent thesoil-landscape-vegetation managementunit) or at the landscape scale (ranch orwatershed level). Fewer measurements arerequired to detect change at the plot scalethan at the landscape scale. However, todetect change at the landscape scale, fewermeasurements are required per plotbecause multiple plots are used.
Appendix C describes three options forestimating the number of vegetation transects andsoil measurements you will need. It includes tablesthat allow you to create unique recommendationsbased on each of the four factors listed above.These tables are based on spreadsheets that alloweven more flexibility in monitoring program design.The downloadable (http://usda-ars.nmsu.edu)spreadsheets will allow you to change transectlength and number of points per transect, as wellas minimum detectable change and statisticalparameters.
Table 4.3 lists one set of recommendations fora semiarid grassland monitoring unit, based onOption 2 in Appendix C. Each of the long-termfactors listed above affects measurementrecommendations. For example, referring to theinformation presented in Table 4.3, if we wantedto detect a minimum change of five percent bareground we would need four plots, while for achange of ten percent, only two plots are needed.
Step 4.3. Define measurementfrequencyMeasurement frequency should be matched toexpected rates of change based on minimumdetectable change selected in Step 4.2. If thesmallest change in basal cover you can detect isfive percent and it takes at least five years for thischange to occur, it’s a waste of time to repeatmeasurements more frequently.
25
Indicator selection
Step 4.4. Estimate timerequirementsUse Monitoring Program Design Form II toestimate total time requirements. Timerequirements can vary by a factor of four or more,depending on vegetation structure, speciesidentification requirements, weather, and observerexperience and condition. Some people prefer towork by themselves, while others prefer a datarecorder. Expect to double total time requirements
.3.4elbaT gnirotinomdnalssargdira-imesanihtiwegnahctcetedotderiuqerstolpforebmuNCxidneppAni71.C-51.CselbaTnodesaberewsetamitseesehT.)elacsepacsdnal(tinu
dexim,)liosydnas(dnalssargsuorefinolotsrof)etamitseelddim(naidemehtgnisu,)2noitpO(m05eerhtrof)dedarged(dnalssargsuorefinolotsdnadnalssargsuorefinolots/suotamozihr
.tolprepstcesnart
tnemerusaeM)rotacidni(
muminiMelbatceteD
*egnahC
stolP muminiMelbatceteD
*egnahC
stolP
tpecretnitniop-eniL05:dnuorgerab(
)tcesnartrepstniop
%5 4,6,24=naideM
%01 2,2,22=naideM
tpecretnitniop-eniLstniop05:revocrailof(
)tcesnartrep
%5 6,7,26=naideM
%01 2,2,22=naideM
tpecretnipaGyponaC)]tf7.1~[mc05>spag(
%5 11,6,88=naideM
%01 3,2,22=naideM
tsetytilibatslioS6:ytilibatsecafrus(repstnemerusaem
)tolp
tinu1 3,2,63=naideM
stinu2 2,2,22=naideM
.noitanalpxerofCxidneppAees;5.0=ohr;8.0=rewop;2.0=p*
for the first year to allow for plot establishmentand characterization. Double them again if it is thefirst time a person has established plots andcompleted these measurements.
If the time requirements seem too high, don’tgive up! Carefully review the assumptions youhave made about the indicators needed andstatistical precision required. Review yourobjectives. Many indicators are interesting, butoften just a few are essential.
26
IImr
oF
ngise
Dmar
gor
Pg
niroti
no
M.
tsiltnempiuqe
ehtdnatsil
sihtes
U.tolphcae
rofdennalp
stcesnartforeb
mundna
stolpgnirotino
mfotsiL.
mrofsihtfo
kcabeht
no)5.h
C(airetirc
noitcejertsildna
hcaorppanoitceles
etisebircse
D.stnemeriuqertne
mpiuqedna
emit
nalpot
6retpah
Cni
)tfr
om(____:
htg
neltces
nartdra
dnat
Seno
elcricllec
hcaeni
tol
pre
pstces
nartf
ore
bm
un
droce
R
gnir
otin
oM
ema
Nti
nU
morf(gnirotino
Mmargor
Pngise
D)I
mroF
tol
Pe
man(
).o
nr
o
SP
G____:e
no
Z/m
utaD
ro
MT
U?
gn
oL/ta
L
)s(tcesnar
Tssa
pm
ocs
noitceri
d)s
htu
miza(
oto
hP
stni
oP
/sotohp:tcesnart
______
-eni
Lt
nio
pt
pecretni
/stp:tcesnart_______
tpecret
nipa
G:pag
yponac.nim
tfro
mc______:paglasab.ni
mtf
romc______
ytilibatsli
oS
/stnemerusae
m:tcesnart
______
tleB
tcesnart
:htdiw
_____ (tf
rom
) enoelcric
reht
O-er
usaem
:tne
m
reht
O-er
usaem
:tne
m
setani
dro
oC
yp
ona
Clasa
Becafr
uS
-b
uS
ecafrus
:lato
T
tcesnart
ro
tne
merusae
m/emi
T1 :
emitlat
oT
2 :
1.yad
tsrifeht
noe
mithcu
msa
semit
ruofot
owt
deenelpoep
tsoM
.ecneirepxefo
syadeerht
tsaelta
retfaderiuqer
emit
nodesab
fietarucca
tsoM
2".tcesnart
rotne
merusaem/e
miT"
yb"lato
T"ylpitlu
M
:)5.h
Cees(
airetircn
oitcejert
olP
27
Advantages• Can be representative of all areas (if sufficient
number of plots included).• Easy to apply.• Statistically valid.• Is clearly “unbiased.”
Disadvantages• Not very cost-effective.• Rarely includes locations in sensitive areas or
areas of special concern because they usuallyrepresent a relatively small proportion of thetotal land area.
• Not sensitive enough to monitor degradationand recovery except where changes areoccurring throughout all parts of themonitoring unit.
Chapter 5
Step 5: Select monitoring plotlocationsChecklist
Step 5.1. Choose and apply site selection approach .......................................... _________Step 5.2. Select “rejection criteria” and use to eliminate unsuitable locations . _________
Step 5.1. Choose and apply siteselection approachThere are three approaches to selecting monitoringplot locations: (a) random, (b) stratified randomand (c) subjective. Each approach has advantagesand disadvantages. The one you select depends onyour monitoring objectives, knowledge of the areato be monitored and the number of plots you canafford to monitor. In most cases, we recommendthe stratified random approach for developingcost-effective, statistically valid monitoringprograms.
Regardless of the site selection approach youchoose, use Monitoring Program Design Form II(Ch. 4) to record information for each plotselected. Describe the approach used to select thetransects, and any rejection criteria, on the form.
Step 5.1(a) Random Plot Selection. Plots can berandomly selected using any map or aerialphotograph. Simply create a fine-scale grid andplace it on top of the map or photo. This can beeasily done by placing one ruler on the bottom ofthe map with the “0” end in the lower left cornerand a second ruler perpendicular to it along theleft edge, again with the “0” end in the lower leftcorner. Randomly select two distances on eachruler (e.g., 6.1 and 10.7 in Fig. 5.1) and find thepoint where the two lines intersect. Repeat untilyou have selected all plot locations. Make sureeach plot is at least 200 m from the closestneighboring plot.
If a DOQQ or other orthorectified image isavailable, the same process can be applied using agrid of UTMs instead of the ruler. These coordinatescan then be entered directly into a GPS unit.
Figure 5.1. Random selection of monitoring locationsusing rulers and an aerial photo. The numbers 10.7and 6.1 were randomly selected (see text).
cm
10.7 cm
0000000000000000 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1211
109
87
65
43
21
0
6.1 cm --.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.->
-.-.
-.-.
-.-.
-.-.
-.-.
-.>
28
Step 5.1(b) Stratified Random Plot Selection.Stratified random sampling is identical to randomsampling except the number of plots in each typeof monitoring unit is predetermined. Plot locationwithin each type of monitoring unit is randomlyselected. This allows monitoring to focus on areaswith a high degradation risk or recoveryopportunity.
For example, in Figure 5.1, if the primaryobjective is to monitor for degradation risk andthe primary degradation process is tree invasion,then a higher proportion of the plots should belocated in the blue grama grassland, even thoughmost of the monitoring unit is pinyon-junipersavanna.
Calculating indicator averages is slightly morecomplicated with stratified random than withrandom. For stratified random, calculate theaverage value for each type of monitoring unit.Then multiply each average value by theproportional area covered for the correspondingtype of unit (e.g., 0.3 for a type of unit covering30% of the total area monitored). The average forthe total area monitored is the sum of all theproducts (monitoring unit x proportion of area).The example in Table 5.1 shows that it’s easierthan it sounds.
Advantages• Can be representative of all areas (if sufficient
number of plots included, plots are located in alltypes of monitoring units, and the total area ofeach type of monitoring unit is known).
• Statistically valid.• Cost-effective.• More sensitive to areas with a high probability
of change (degradation and/or recovery).
Disadvantages• Requires pre-stratification (this should already
be done as part of the design process. SeeCh. 2).
Step 5.1(c) Subjective Plot Selection. Subjectiveplot selection includes all approaches in which theperson designing the monitoring program decideswhere to locate the plots without using a gridsystem. This nonrandom approach has been usedto select a majority of existing monitoring plots.Most historic USFS and BLM monitoring transectswere selected subjectively by experienced rangeconservationists using the “key area” conceptdiscussed below.
Subjective site selection can result in muchmore sensitive and representative monitoringprograms. However, this is only possible wherequalified personnel with a good understanding oflocal soil and vegetation patterns and processesdesign such monitoring programs.
Advantages• Sensitive to local patterns and land use.• Does not necessarily require access to maps
and photographs.• Inexpensive.
Disadvantages• High potential for bias.• Difficult to extrapolate.
Key areasA key area is a tract of land that is assumed to berepresentative of much larger areas and is likely toreflect the effects of management changes onthese larger areas. Key areas are often used insubjective plot selection.
Site selection
.1.5elbaT arofdnuorgerabegarevagnitaluclaCstinugnirotinomfosepyteerhthtiwdehsretaw
noitcelestolpmodnardeifitartsnodesab(.)hcaorppa
tinUepyT
noitroporPaerAlatoT
egarevAeraB
dnuorG
noitroporPegarevAx
eraBdnuorG
A 8.0 %02 %61
B 1.0 %05 %5
C 1.0 %04 %4
.gvA %52
29
Key areas have been used in the design ofmany monitoring programs throughout the world.Key areas, like any subjective approach, can beextremely effective when applied by qualifiedpersonnel.
Where used for monitoring livestock grazingeffects, key areas are usually placed in an area thatreflects typical livestock use. They are not locatednear watering points, mineral supplements, fences,trails or isolated areas of a pasture that areinfrequently visited. The recommended distancefrom water varies with topography, vegetation andspecies or class of livestock.
Step 5.2. Select “rejectioncriteria” and use to eliminateunsuitable locationsList the rejection criteria in the space at thebottom of the Monitoring Program Design Form II(Ch. 4). Thoroughly describe the reasoning used toselect these criteria. This is important because thecriteria are used to help define how themonitoring data will be extrapolated and becausewhat seems intuitive to us today may not seemintuitive to other individuals, or even to ourselves,many years later.
Rejection criteria can be based on almostanything. Many monitoring programs excludeareas that are thought to be anomalous becausethey receive unusually high or low levels ofdisturbance. Examples of rejection criteria include:(1) plots must be located a minimum of 100 yardsfrom a road or watering point (to avoidunrepresentative high disturbance areas); (2) noplots on rock outcrops or slopes greater than
50 percent (these areas are unlikely to bedisturbed).
Specific locations may also be anomalousbecause of landscape position. For example, areasthat receive unusual amounts of runoff or haveunusually dense stands of trees in a savanna maybe rejected because they are not representative oflarger areas.
Large areas that are not expected to changebecause they have crossed a threshold are alsooften omitted from monitoring programs. Thestate and transition model and indicators used tojustify omission of these areas should be listed.
Rejection criteria should be carefully selectedto ensure areas that should be monitored are notomitted. Also, the most unusual areas are oftenthose that change the most quickly and may serveas early-warning indicators of degradation orrecovery in other parts of the landscape. Ratherthan excluding these anomalous areas, we suggestthat a stratified random site selection approach beused where possible. This allows apparentlyanomalous areas to be clearly identified as part ofthe monitoring program and potentially includedin a future expansion of the monitoring program.Where there are areas of less interest (e.g., thepost-threshold areas), monitoring may be limitedto a few photo points.
It is highly recommended that a list ofrejection criteria be developed prior to selectingand visiting monitoring locations. Deciding toreject areas after visiting them because they “don’tlook right” introduces bias. See Chapter 17 foradditional guidance on the use of soil andlandscape features to improve monitoring datainterpretation.
Site selection
30
Chapter 6
Checklist6.1. Establish and permanently mark plots and transects ............................... _________6.2. Describe monitoring plots and record GPS locations, including
coordinate system, datum and zone .......................................................... _________6.3. Record long-term data ................................................................................ _________6.4. Error check and copy the data and keep copies in different locations ..... _________
A fter you have gone through the previousfive steps, this one should seem easy. It’simportant to carefully mark and describe
each monitoring plot for two reasons: so you canfind it again and so you can compare your dataagainst data collected on plots with similar soils,topography and climate — all of the things thatdetermine site potential. Use the equipmentchecklist for pre-field planning.
Step 6.1. Establish andpermanently mark plots andtransectsBy now you should have already selected wherethe plots are to be located (Ch. 5). Be sure to verifythat the site is suitable by checking it against the“rejection criteria” you list on the back of theMonitoring Program Design Form II (Ch. 4).
Step 6.1(a) Upland spoke design plots (Fig. 6.1).Place a permanent stake into the ground at thecenter of the monitoring plot. This stake will alsoserve as the photo point (Quick Start).
Using a randomly selected azimuth (compassdirection: 1° to 360°), extend a tape in the azimuthdirection to a distance of 5 m (15 ft) further thanthe length of the transect. Install a stake at the 5 mmark. This will serve as the 0 m end of yourtransect, because the transect begins 5 m from thecenter point (Fig. 6.1). Mark the far end of thetransect with a stake.
Repeat transect establishment at regularintervals in a circle around the plot. The intervaldepends on the number of transects. For mostapplications, there will be three transects, with120° between each.
Figure 6.1. (a) Three spoke design plots locatedwithin an upland monitoring unit. The starting point ofeach transect is 5 m from the plot’s center. (b) Singletransect design maximizes spatial distribution acrossthe landscape.
Step 6.1(b) Single transect upland plots. Anchorand mark the 0 m end of the transect. Using arandomly selected azimuth (compass direction:1° to 360°), extend a tape in that direction thelength of the transect. Mark the far end of thetransect with a stake.
Step 6.1(c) Single transect riparian plots (Fig. 6.2).Anchor and mark the 0 m end of the transect.Ensure the 0 m end is placed such that the transectwill cross the riparian channel perpendicular tothe channel, and the 0 m end is 5 m beyond theriparian zone. Extend the tape perpendicular tothe riparian channel. Mark the far end of thetransect with a stake.
Step 6: Establish monitoring plots
(a)
(b)
31
Plot establishment
Border of riparian channel (water flow area)
Border of riparian zone (defined by vegetation)
Riparian channel veg survey
Riparian channel profile
Standard transect anchored outside riparian zone
Figure 6.2. Single-transect plots crossing a streamwithin a riparian monitoring unit. Note that transectsare anchored outside the riparian zone. See Chapters13 and 14 for the Riparian channel vegetation surveyand the Riparian channel profile measurementsassociated with riparian transects. Single-transectplots can be used in upland areas.
Step 6.2. Describe monitoringplots and record GPS locations,including coordinate system,datum and zoneAt a minimum, fill out the Required section of theMonitoring Plot Description Form (found at theend of this chapter) when you establish eachmonitoring plot. This will ensure the samephysical location is always monitored.
The Recommended section providesinformation regarding the plot’s potential tosupport a given plant community, and enables youto verify the plot is on the mapped soil andecological site. The data in this section allow youto determine how similar this plot is to other plotswithin the same ecological site. The informationgathered here can help identify potential offsiteinfluences. Data gathered in this section also assistin determining applicability of extrapolating plotdata to the landscape level.
The Optional section addresses plotdisturbances and management history. Thisinformation is valuable for data interpretation. Itcan help identify potential causes of trends andassist with important land management decisions.
The Recommended and Optional sections ofthe data form are important for data interpretation.It is best to fill out this information when youestablish the plot, as you need to be on the plot tocomplete these sections. However, if you havetime constraints, these sections can be completedduring a second plot visit.
Required Section. Record the site, managementunit and/or pasture name where the monitoringplot is located. A site, management unit or pastureis a distinct geographic unit typically with thesame landowner and a relatively homogeneousmanagement/disturbance regime. There are oftenmultiple plots located within the same site,management unit or pasture.
Complete the remaining portions of theRequired section to ensure a permanent record ofplot and transect locations. Record the locations(e.g., GPS coordinates) on the Monitoring ProgramDesign Form II and the Monitoring Plot DescriptionForm. If you are using GPS coordinates, make sure torecord the coordinate system, datum and zone,and whether the data are in English or metricunits. Record geographic locations of the plot’scenter and the beginning and end of each transect.Document each transect’s compass azimuth andrecord the declination used, if any.
Recommended Section. Record the average long-term precipitation under “Avg. Precip,” indicatingwhether the units are English or metric. Determinethe soil series by comparing soil observations(recorded in the table) with a soil survey. Do notrely on soil maps alone, although they can be usedas guides. Soils can be extremely variable withinbroadly defined map units.
Dig a small pit at the plot’s center and at theend of each transect. The pit should be deepenough to detect soil horizons that significantlyaffect plant growth. An auger or soil probe can beused instead of a pit if you are already familiarwith the soils in the area. Any soil information is
32
Plot establishment
helpful. Don’t worry if you are not sure how todistinguish soil horizons.
If you suspect there may be more than onetype of soil within your plot boundaries, be sure tolocate at least one of your mini pits within eachsoil type. It is very important to record the soildepth. Record the depth to any horizon that islikely to restrict water movement or root growth.If no restricting horizon is encountered, record themaximum depth of the pit and write “norestricting horizon.”
Record the upper and lower boundary of eachmajor horizon under soil depth (e.g., “Btkhorizon” or “horizon with clay and carbonateaccumulation” from 23 to 40 cm) (refer to USDA-NRCS 1999). Record information about eachhorizon in a separate row on the data form. Referto The Field Book for Describing and Sampling Soils,Version 2.0 (Schoeneberger et al. 2002), foradditional guidance.
If available, the soil survey will also provideinformation on soil parent material. For areas inthe United States that have been mapped by theNRCS, the ecological site can be determined bylooking up the soil series online (http://esis.sc.egov.usda.gov) or by consulting the FieldOffice Technical Guide in your local NRCS officeor NRCS website. If ecological sites have not beendeveloped, this space can be used to record anyother land classification information (e.g., USFSTerrestrial Ecosystem Unit) that may assist withdata interpretation.
Record the slope using a clinometer or otherinstrument. Record the slope’s aspect, lookingdownslope, in compass degrees (e.g., 108°) orcardinal direction (SE). Record the slope shape asconvex ( ), concave ( ) or linear (straight, notcurved) (Lal 2003).
Record the landscape unit on which your plot islocated. If the plot is located on a hill/mountain,select the appropriate hillslope profile component(see list on data form). If the plot is located on aterrace, indicate whether it is on the riser (fairlyshort, steep, linear slope that forms the sideslope ofthe terrace) or the tread (a broad, relatively levelplanar portion forming the top of the terrace thatcan extend laterally for great distances). Refer to anyof the following for assistance: Encyclopedia of SoilScience (Lal 2003), the Field Book for Describing and
Sampling Soils (Schoeneberger et al. 2002), Landformsof the Basin and Range Province (Peterson 1981),Geomorphology of Soil Landscapes (Wysocki andZanner 2003) and National Soil Survey Handbook,Part 629, Glossary of Landform and GeologicTerms, online at http://soils.usda.gov/technical/handbook/detailedtoc.html#629 (USDA-NRCS 2003).
Optional Section. Record recent weather patternsfor the previous 12 months and the year prior.Record any disturbances and managementinformation that might impact the plot. Describeall known or observed wildlife and livestock use,including utilization, seasonality, intensity andresidual cover. Describe previous managementhistory dating as far back as possible. Describeoffsite influences, including, but not limited to(a) unusually high runoff or erosion from upslopeareas, (b) management practices, (c) presence ofinvasives in the area, and (d) roads. Document anyother pertinent information under “Othercomments.” Draw a depiction of the plot andrecord all potential drivers and other influences.
Step 6.3. Record long-term dataThere are four options:
(1) Enter data onto paper data formsphotocopied from each chapter.
(2) Enter data onto paper Microsoft® Excelspreadsheet forms printed from the websitelisted below.
(3) Enter directly onto downloaded Excelspreadsheet forms using a laptop or tabletcomputer.
(4) Enter directly into a database (see websitefor future download at http://usda-ars.nmsu.edu).
Step 6.4. Error check and copythe data and keep copies indifferent locationsAll data should be checked for errors, using thesteps below. This process can and often shouldtake as long as data entry.
Step 1. Before leaving the plot, review all forms forcompleteness and legibility. If you used a digital
33
Plot establishment
camera, review all of the photos to be sure thatplot information is visible. If the data were entereddirectly into a computer, open and inspect eachfile and make a backup copy on a separate device.
Step 2. Immediately upon returning from the field,copy the data and store in at least two differentbuildings. This is especially important if you areusing electronic data entry. Unless you have atleast one backup copy on media that you are surewill be readable in the future (remember 51/2 indisks, punch cards and tape drives?), then werecommend printing copies of your data forarchiving. Be sure that someone else knows wherethe extra copies are.
Step 3. Review all data for obvious errors. Forexample, check to see that each column on theLine-point intercept form includes only thosecodes assigned to that column. Gap interceptsshould not overlap. Soil stability values should bebetween 1 and 6. See Table 6.1 for an example of aCompaction test (Ch. 7) dataset that includespotential outliers (extreme values) at positions 3and 6 that may or may not be errors. If we knowthat this site had shallow soil or large rocks ornear-surface bedrock, we could justify deleting thetwo bolded data points by assuming that thepenetrometer struck a rock. We could also deletethem if we were fairly certain that the datarecorder had lost count. However, it is alsopossible that both measurements were made on a
game or livestock trail. In this case, they should beretained. When in doubt, retain the data and makea note.
Step 4. If the data were originally recorded onpaper forms, re-check each value after computerentry. One way to do this is for one person to readthe values aloud from the data form while anotherchecks the values in the spreadsheet or database.Your data are now ready for indicator calculation(Ch. 16 and individual methods chapters) andinterpretation (Ch. 17).
“Outliers”Values that are clearly outside the range of expected variation may be omitted if it is clear that
these values were due to measurement or data recording errors. Extreme care should be taken toensure that these values are in fact mistakes before omitting them.
A famous ecologist once said, “There are data that are outliers (meaning that they lie outside thenormal range of variation) and then there are “out-and-out liars.” Be sure that you only exclude the“out-and-out liars.” The other outliers may be extremely important in defining the current status ofthe system.
.1.6elbaT htiwatadtsetnoitcapmoCfoelpmaxEebtonyamroyamtaht)epytdlobni(sreiltuoowt
".srailtuo-dna-tuo"
nosoPeniL
sekirtSevitalumuC
mc5 mc01 mc51 mc02
1 5 11 81 82
2 3 9 81 03
3 4 01 02 47
4 4 11 22 03
5 6 11 91 72
6 5 01 42 89
7 4 21 22 33
8 3 9 71 62
9 4 21 91 72
01 5 31 02 92
34
tsilkcehctnempiuqE . roftpecxe,edamerastnemerusaememithcaederiuqereratsilsihtnidedulcnismetillAyratnemelppusrofsnmulocddA.tsil"tnempiuqEnoitpircseDdnatnemhsilbatsEtolP"ehtniylnodnuofsmetiesoht
.tnemhsilbatsetolpgnirotinomnoliatederomrof6.hCeeS.tnempiuqelanoitiddarofswordnasdohtem
tnempiuqEnoitpircseDdnatnemhsilbatsEtolP ?evaHerawtfosyrtneatadhtiwretupmocdleifROslicnep,smrofatad,draobpilC
retemonilC
ssapmoC
stolpgnirotinomfopamro,deretnestniopyawhtiw,tinuSPG
rabernignidnuoprofremmaH
snoitanibmocetagdnasyeK
)yrassecenfi(deifitonrenwodnaL
tuoyal"ekops"rof,3yllaedi,1tsaelta–)htgneltcesnartehtnahtregnolm5(epatgnirusaeM
)htpedliosrof(erusaemepatlateM
sdnetcesnartgnikramrofsekatsrehtoro)tf3(rabeR
levohS
slootnoitaziretcarahclioS
)srohcnaepat(6-2:sekatsleetS
setamitseerutxetliosrofliosgninetsiomrofretaW
.)noitacifitnediliosrofdesusetanobraclioserehwdedeenylno(ecnecsevrefferof)dicacirolhcordyh(lCHM1ytefasllawollofdna)teehSataDytefaSslairetaM(SDSMniatbo,desufI.snrubesuacnaclCH:noituaC
.senilediug
)noitcellocatadllaylraenrofdedeen(tnempiuqEcisaBssapmoC
stolpgnirotinomfopamro,deretnestniopyawhtiwtinuSPG
snoitanibmocetagdnasyeK
)yrassecenfi(deifitonrenwodnaL
tuoyal"ekops"rof3yllaedi,1tsaelta-)htgneltcesnart(epatgnirusaeM
)srohcnaepat(6-2:sekatsleetS
dohteM/tnemerusaeMhcaErofderiuqeRtnempiuqElanoitiddA
tnempiuqEotohPstniop
-eniLtniop
tpecretni
paGtpecretni
lioSytilibats
tset
tleBtcesnart
:rehtO :rehtO :rehtO
gnol)tf5(m5.1:elopCVP X
aremaC X
otohProdraobklahc,draobetihWdracDItniop
X
nepgnikramkcihT X
ROslicnep,smrofatad,draobpilCatadhtiwretupmocdleif/potmlaP
erawtfosyrtneX X X X X
retnioprehtorogalfniP[mm1<pit( 1/ 52 )retemaid]ni
X
rehtoroekatsgninnip,kcitsreteMdorrokcitsffits
X
tikytilibatslioS X
rep)zo23(retil1:retawdezinoieD)selpmas81(tset
X
hctawpotS X
tsaelta(epipCVP 1/2 tcesnarttleBfo)htdiw
X
35
)2fo1egaP(mroFnoitpircseDtolPgnirotinoM
deriuqeR
_____:htgneLeniL_____________:etaD______________:erutsaP/tinUtnemeganaM/etiS )?tfrom(
_______?tfrom_______:enoZ___________:mutaD________________:metsySetanidrooCSPG
eniL/tolP htumizA
)dnem0,.g.e(__________:noitacoL )dnem05,.g.e(_________:noitacoL noitavelE
edutitaL/gnihtroN edutignoL/gnitsaE edutitaL/gnihtroN edutignoL/gnitsaE ?tfrom
tolpotsnoitceriD _________________________________________________________________________:
_________________________________________________________________________________________
dednemmoceR
___________:picerP.gvA______________:ytnuoC____:etatS ______________________:etiS.locE
_________________________:)etisnoyfirev(tinUpaM/seireSlioS __________________:lairetaMtneraP
______:tcepsA%_____:epolS
:epahSepolSlacitreV raeniL____evacnoC____xevnoC____
:epahSepolSlatnoziroH raeniL____evacnoC____xevnoC____
tipiniMnoitacol
lioShtped
)ni/mc( erutxeT
kcoRstnemgarf
)%(-sevreffE
ecnec roloCerutcurtS
edarGerutcurtS
epahS
ecnetsisnoChtgnertS
)tsiomroyrd(
:tinUepacsdnaL
nialptalF___**ecarreT___nisabdoolF/nialpdoolF___tnomdeiPnaF___*sniatnuoM/slliH___
:tnenopmoCepolslliH* ____________:rehtO__epolseoT__epolstooF__epolskcaB__redluohS__timmuS__
tnenopmoCecarreT** _____________:rehtO___daerT___resiR___:
36
yrotsiHtnemeganaMdnasecnabrutsiD)2fo2egaPmroFnoitpircseDtolPgnirotinoM(
lanoitpO
:rehtaewtneceR :shtnom21tsaP ________:thguorD _________:lamroN __________:teW
:shtnom21suoiverP ________:thguorD _________:lamroN __________:teW
:)ylppatahtllakcehc(secnabrutsiDtneceR
erifdliW___ retawmorfnoitisopedlioS___
stnedorllamS___ dniwmorfnoitisopedlioS___
)stnedorton(slammamregraL___ seitilitudnuorgrednU___
retaW___ senilnoissimsnartdaehrevO___
dniW___ ___________________________________:)ebircsed(rehtO___
__________________________________________________________:esukcotsevildnaefildliwebircseD
________________________________________________________________________________________
________________________________________________________________________________________
retaw,gniwolp,gnidees,lortnocburhs,erifdebircserp,nalpgnizarg,.e.i(yrotsihtnemeganamebircseD
___________________________________________________________________________________:)stinu
________________________________________________________________________________________
______________________________________________________:tolpsihtnosecneulfnietisffoebircseD
________________________________________________________________________________________
________________________________________________________________________________________
________________________________________________________________________:stnemmocrehtO
________________________________________________________________________________________
________________________________________________________________________________________
.secneulfnirehtodnasrevirdetis-nodnaetis-ffollaebircseD.epacsdnalgnidnuorrusdnatolpwarD
![Customers Churn Prediction using Artificial Neural ... · devising of churn policy [16] is depicted in Fig. 2. Fig. 2. The Six Steps for Customer Churn Prediction. Churn prediction](https://static.fdocuments.us/doc/165x107/5e71a26fb4acff71e10cc1fe/customers-churn-prediction-using-artificial-neural-devising-of-churn-policy.jpg)
