Transcript Document

The Pan-STARRS
Moving Object Processing System
(& Science)
Robert Jedicke
(for the Pan-STARRS collaboration)
Institute for Astronomy
University of Hawaii
2004 September 29
IMPACT
IMPACT
IMPACT
The Pan-STARRS
Moving Object Processing System
(& Science)
Robert Jedicke
(for the Pan-STARRS collaboration)
Institute for Astronomy
University of Hawaii
2004 September 16
The Pan-STARRS
Moving Object Processing System
(& Science)
(& Science)
Robert Jedicke
(for the Pan-STARRS collaboration)
Institute for Astronomy
University of Hawaii
2004 September 16
Bigger Further Slower Dumber
Bigger Further Slower Dumber
DEFINITIONS
icier
COMETS
ASTEROIDS
dirtier
DEFINITIONS
Near Earth Objects (NEO)
NEO ZONE
Perihelion < 1.3AU
(about 130 million miles)
DEFINITIONS
Potentially Hazardous Objects (PHO)
PHO ZONE
MOID < 0.05 AU
(about 5 million miles)
PHO Orbit
Earth Collision
at perihelion
Non-Collision ‘PHO’ Orbit
Not at Earth’s
orbit at perihelion
1995 CR
DEFINITIONS
Death Plunge Objects (DPO)*
* Not an official acronym
Solar System Animation #3
DEFINITIONS
Trojans
Main Belt Objects
Trojans
DEFINITIONS
Centaurs
Trans-Neptunian Objects (TNO)
Comets
Short
Period
Comets
Halley
Family
Comets
Long Period Comets
DEFINITIONS
Oort
Cloud
The Pan-STARRS
Moving Object Processing System
(MOPS)
Selected PanSTARRS’s Top
Level Science Requirements
• MOPS shall create and maintain a data
collection of detections and object parameters
(e.g. orbit elements, absolute magnitudes) for
>90\% of the PHOs that reach R=24 for 12
contiguous days during the course of PanSTARRS operations.
• MOPS shall create and maintain a data
collection (DC) of detections and object
parameters (e.g. orbit elements, absolute
magnitudes) for >90% of the members that
reach R=24  12 contiguous days within each
class of solar system object (Main Belt, Trojan,
Centaur, TNO, Comet, etc, except NEO and
PHO) during the course of Pan-STARRS
operations.
Selected PanSTARRS’s Top
Level Science Requirements
• MOPS shall create and maintain a data
collection of detections and object parameters
(e.g. orbit elements, absolute magnitudes) for
>90\% of the PHOs that reach R=24 for 12
contiguous days during the course of PanSTARRS operations.
• MOPS shall create and maintain a data
collection (DC) of detections and object
parameters (e.g. orbit elements, absolute
magnitudes) for >90% of the members that
reach R=24 12 contiguous days within each
class of solar system object (Main Belt, Trojan,
Centaur, TNO, Comet, etc, except NEO and
PHO) during the course of Pan-STARRS
operations.
Why?
REASON #1
REASON #2
SPACEGUARD GOAL
SPACEGUARD GOAL
NASA NEO SDT
Pan-STARRS & PHOs
• 99% completion of PHOs with D>1km
 90% reduction in residual
global impact risk
• 90% completion of PHOs with D>300m
 50% reduction in
sub-global impact risk
REASON #3
REASON #4
Existing Surveys
Existing Surveys – Step 1:
Discovery & Identification
LINEAR
White Sands, NM)
LONEOS
Flagstaff, AZ
NEAT/JPL
Haleakala, Maui
CSS -South
Australia
NEAT/JPL
Palomar, CA
CSS - North
Mt. Lemmon, AZ
Spacewatch
Kitt Peak, AZ
UHAS
Mauna Kea, HI
• 3-5 images/night
• Linear motion
• Very low falsepositive rate
Existing Surveys – Step 2
Linkage & Orbit Determination
MPC
• Links detections
to known
objects
• Identifies new
objects
• Fits orbits to all
objects with
new detections
• Much more…
Existing Surveys – Step 3
Impact Risk Assessment
• Refine orbits
• Calculate impact
probability
Moving Object Processing System
Pan-STARRS
Telescopes
&
Survey
Image
Processing
Pipeline
MOPS
Impact
Probability
• Fully integrated
• Detection, attribution, linking,
orbit identification
• Orbit fitting
• Parallel synthetic data analysis
 Real-time efficiency/bias
Moving Object Processing System
Moving Object Processing System
Moving Object Processing System
• MPC requires that reported detections be real
 forces Pan-STARRS to obtain 3 images/night
 reducing total sky coverage
 reducing total discoveries
• Difficult to control/monitor system efficiency
 introduce synthetic objects into data stream
 determine efficiency in real time
 monitor system performance in real time
PanSTARRS Asteroid Surveying
7
• 10 asteroids within range of PanSTARRS
2
• ~200 / deg @ V=24 @ on ecliptic
7
• 10 detections / month (20X current rates)
Cumulative Observations
160,000,000
140,000,000
120,000,000
100,000,000
80,000,000
60,000,000
PS1 Starts
40,000,000
20,000,000
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1997
1996
1995
0
Observing Cadence
• Every survey mode obtains at least two
images at each location separated by a
Transient Time Interval (15-30 minutes)
 serendipitous positions & colours
• Solar system survey re-visits each
location after 3-6 days
 obtain 3-4 nights/month
 ~12 day arc
Moving Object Processing System
• 2 detections/night
with multi-night linking
 increased sky coverage
 push deeper into noise
 more objects
•synthetic data
 real-time system monitoring
 efficiency determination
 correction for selection effects
Transient Detection (IPP)
4 Telescopes
Stationary
+
Combined
+
Static
Transients
Moving
+
Transient Types
Supernovae/GRB
Difference
Slow Asteroidal Object
Normal Asteroidal Object
Cometary Object
Death Plunge Object
Fast Asteroidal Object
Linking Detections
Day 1
1 Field-of-view
1500 real detections +
1500 false detections
Linking Detections
Day 5
1 Field-of-view
1500 real detections +
1500 false detections
Linking Detections
Day 9
1 Field-of-view
1500 real detections +
1500 false detections
Linking Detections
•Brute force (MPC) approach
 100X Pan-STARRS computing power
• kd-tree (CMU) approach
 ~1/3 Pan-STARRS computer power
Orbit Determination
• Must include
– All major solar system perturbing bodies
– Full error analysis
• Two available solutions
– AstDys (Italy)
– JPL (USA)
Data Storage
• Large by most astronomical standards
• Small in comparison to Pan-STARRS (~1%)
500 TerraBytes
Synthetic Data
• Inject synthetic objects into MOPS
parallel to real data analysis
 monitor system efficiency for correcting
observational selection effects
 monitor system performance to flag unusual
behavior
Synthetic Data
• Synthetic model matches real distributions
 all asteroid and comet types
 realistic orbit and size distribution
 realistic shape, rotation periods, pole
orientations
 + ‘unusual’ orbits e.g. hyperbolic interstellar,
retrograde main belt, distant Earths
MOPS: Known Object Attribution
MOPS: Synthetic Detection
& Noise Generation
MOPS: Orbit Determination
& Attribution Loop
MOPS: Linking New Detections
The Pan-STARRS
Solar System
Survey & Science
Solar System Survey Locations
19:00 HST
00:00 HST
05:00 HST
Evening Sweet Spot
Opposition
Morning Sweet Spot
Pan-STARRS & NEOs/PHOs
• Tens of thousands of NEOs




Size-frequency distribution
Orbit distribution
Source fitting
Genetic families?
Pan-STARRS & the Main Belt
• Pan-STARRS will find as many objects in
one lunation as have been identified
since the discovery of Ceres in 1801
Pan-STARRS & the Main Belt
• 10,000,000 MB objects in ten years








Size-frequency distribution
Orbit distribution
New small asteroid families
Asteroid/comet transition objects
Asteroid collisions
Pole Orientations
Rotation Rates
Shapes
Pan-STARRS & Trojan Asteroids
6
1,000,000
5
100,000
4
10,000
Series1
Known
Series2
Pan-STARRS
3
1,000
2
100
1
10
01
1
Jupiter
2
Saturn




3
4
Uranus
Neptune
Trojans of all giant planets
L4 & L5 swarm statistics
Genetic families
SFD through rollover at H~11
Jewitt 2003, ‘Project Pan-STARRS and the Outer Solar System,’ EMP
Pan-STARRS & Comets
• Pan-STARRS will find ~10X as many
comets per year as all existing surveys
• 1,000’s of comets in ten years operation
 Dormant detections at large distance
 Size-frequency distribution
 Orbit distribution
• INTERSTELLAR ! ! !
Pan-STARRS & Comets
• Comet designation problem
• New Proposal
 Comet Jedicke-XXX
 X=(0-9,a-z,A-Z) (base 62)
 allows for ~240,000 comets
P/Jedicke 1996A1
Pan-STARRS & TNOs
• ~20,000 TNOs





Inclination distribution
Size-frequency distribution
Orbit distribution / dynamical structure
More Plutos?
~100 wide binaries
Pan-STARRS & Distant Planets
New Plutos
320AU
New Earths
620AU (50AU)
New Neptunes
1230AU (130AU)
New Jupiters
2140AU (340AU)
Jewitt 2003, ‘Project Pan-STARRS and the Outer Solar System,’ EMP
Pan-STARRS Minor Planet Summary
8
10,000,000
7
1,000,000
6
5
100,000
Series1
Known
4
10,000
Series2
PS 1 Year
Series3
PS 10 Years
3
1,000
2
100
1
10
01
1
2
3
4
5
6
7
8
9
10
PS1 - 2006
PS4 - 2008
Coming soon to an island near you.
Pan-STARRS Problem:
Pan-STARRS plans on using a very wide ‘Solar
System’ G filter but is required to reach R=24.
Assuming that the R-filter transmission is 100% in the
range [R1,R2] and 0% outside that range and that the
G-filter has similar performance in the range [G1,G2]
where G1<R1 and G2>R2, what is the ratio of the
required exposure times in the two filters to reach
R=24 in the AB magnitude system?
Assuming that Vega is a black-body, what is the
answer in the Johnson system?
Make other reasonable assumptions as necessary 