This research project is funded by NASA Earth System Science Fellowship Program

1

• Deliverables
• New scientific knowledge expected from the
  proposed project include
• An objective means to identify and monitor KRAs.
• Explicit documentation of temporal trends in KRAs
  within highly variable environments.
• An initial assessment of climate and land use
  change impacts on Kenyan rangelands.
• A basis to develop predictive models for
  rangeland management.
• information useful in quantitative and
  qualitative characterization of rangelands
  especially its inherent heterogeneity as
  sustained by key resource areas.
• Model-based decision support systems to
  strengthen early warning systems (Agatsiva and
  Oroda 2002) would be another product with an
  application in ecological forecasting.

Methods
Introduction
A Conceptual diagram illustrating these
hypotheses (a) NDVI variation within a year is
smaller for KRAs than surrounding rangelands
(b) KRA NDVI interannual variability is less
than that for broader rangeland areas (c) KRA
NDVI interannual variability is greater in drier
areas (d) NDVI differences between KRAs and
broader rangeland areas are greater for dry areas
than for wet areas.

• Primary production in Kenyan rangelands
• is relatively low,
• varies substantially from place to place
• is strongly limited by precipitation.
• Precipitation and production is highly variable
• (CV 33 to 60) (Ellis and Galvin 1994).
• Nested within these rangelands, are areas that
  have regular, reliable precipitation patterns and
  thus more consistent plant production from year
  to year.
• These key resource areas (KRAs) are like forage
  oases used by livestock and wildlife during the
  dry season.
• High plant production potential means that KRAs
  are often converted to cropland, excluding
  utilization by livestock or wildlife (Lane and
  Scoones 1993).
• Declines in KRA area coupled with a series of
  severe droughts (Ellis and Galvin 1994) has
  resulted in grazing-induced degradation in
  rangelands (Scoones 1991).
• Land use change may be reducing availability of
  KRAs for use by pastoralists and wildlife but
  KRAs have only been superficially identified and
  remain poorly characterized.

• Questions
• Can KRAs be distinguished from the broader
  rangeland area by examining differences in annual
  phenology?
• How much does phenological variation from year to
  year within a KRA differ from that of the broader
  rangeland area?
• How does this differ in drier versus wetter
  areas?
• How has KRA area changed over time with changes
  in land use?
• How have changes in climate impacted KRA
  phenology?

NB Standing biomass is correlated to NDVI which
is detectable by remote sensing (Tucker and
Sellers 1986, Pickup et al. 1994). Therefore, we
use NDVI as a surrogate of plant production.
Future plans
Identify KRAs, characterize seasonal phenology
(versus that of broader rangeland areas), and
compare phenologies of KRAs and broader rangeland
areas across an aridity gradient using
ground-based and MODIS data (questions 1, 2, and
3). Data intercomparisons to identify the best
sources of data for addressing questions about
changes over time. Evaluate how changes in
climate, climate variability, and land use change
have impacted KRA accessibility and utility
(questions 4 and 5). Expand the analysis to
other dry rangelands in Africa and elsewhere.
Work with rangeland managers to incorporate
this information into a decision support system.
Apply these ideas about key resources to other
ecological questions other than grazing, such as
fuel wood sources.
Background
Satellite data from MODIS will address questions
about current KRA phenology and distribution.
Coarser GIMMS, SPOT-VEG, and LAC data will be
used to assess how KRAs have changed in response
to climate and land use.

• KRA vs. broader rangeland area phenology
• NDVI variability within a year will be lower in
• KRAs than in the broader rangeland areas.

• Mean annual precipitation and KRA interannual
  variability
• In very arid areas, KRAs will have more
  variability than KRAs in less arid areas.
• Differences in NDVI between KRAs and broader
  rangeland areas will be greater in drier than in
  wetter areas.

Remotely sensed vegetation indices are proposed
to investigate KRA phenology in Kenyan rangeland
districts arrayed across an aridity
gradient. Examining how phenological patterns
have changed (1981-present) will aid in
evaluation of how changes in land use practices
and climate affect access to and utility of KRAs.
Objective methods that may be used to assess or
monitor these crucial areas will be developed.

• Interannual variation in KRA phenology
• NDVI interannual variability in areas identified
  as KRAs will be less than that for broader
  rangeland areas.

References
Agatsiva, J., and A. Oroda. 2002. Remote sensing
and GIS in the development of a decision support
system for sustainable management of the
drylands of Eastern Africa A case of the Kenyan
drylands. Int. Arch. Photogramm. Rem. Sens.
Spatial Infor. Sci. XXXIV42-49 Ellis, J., and
K. A. Galvin. 1994. Climate Patterns and Land-Use
Practices in the Dry Zones of Africa. Bioscience
44340-349. Pickup, G., G. N. Bastin, and V. H.
Chewings. 1994. Remote-Sensing-Based Condition
Assessment for Non-equilibrium Rangelands Under
Large-Scale Commercial Grazing. Ecol. Appl. 4497
517. Scoones, I. 1991. Wetlands in Drylands
Key Resources for Agricultural and Pastoral
Production in Africa. Ambio 20366-371 Tucker,
C. J., and P. J. Sellers. 1986. Satellite Remote
Sensing of Primary Production. Int. J. Rem. Sens.
71395 - 1416

• Impacts of land use change on KRAs
• Characterizing KRAs across diverse rangeland
  conditions and over time will enable the
  assessment of effects of land use change.
• Comparison of NDVI trends in regions with land
  use changes versus those without changes will
  provide a basis to evaluate impacts of land use
  change.

Without KRAs, most of Kenyan rangelands would be
unable to support pastoralism or wildlife, much
less coexistence of the two. Large scale,
comprehensive and repeatable characterization of
KRAs will boost understanding and planning for
sustainable use of the rangelands.

• Effects of changes in climate within a KRA
• Assess changes in NDVI phenology within KRAs over
  time (1981-present) compare KRA and broader
  rangeland area phenology in El Niño and normal
  years.
• Quantifying the magnitude of change and response
  to climate variability is critical for sustaining
  rangeland utility in the face of increasing
  climatic fluctuations.

This research project is funded by NASA Earth
System Science Fellowship Program
Animal fatalities increase during droughts
Characteristics of the various districts that
will comprise the study areas (Data from Kenya
Soil Survey, 1982)